Results of the 24-bit challenge
Ethan Winer
A week ago I invited folks here to compare a series of audio
files, to see how audible are the differences between dithering
from 24 bits to 16 versus truncating. I also included files that
were truncated to (approximately) 13, 11, and 9 bits, which
correspond to recording at peak levels of -20, -30, and -40.
There are five files altogether, as described on my web page at
www.ethanwiner.com/bitstest.html. Here are results:
I received 23 replies altogether, and only one person identified
all five files correctly. Not to minimize the one person who got
them all right, but since the two worst files were fairly
obvious, one person in 23 getting the order of the remaining
three files correct likely falls within the boundaries of normal
probability. Then again, based on his comments ("#1 was very
clean throughout ... the guitar string buzz on #4 was more
grainy") shows that perhaps he really did hear the changes in
quality!
The short answer is that most folks were not able to distinguish
any of the files. This comment is typical: "To tell the truth I
couldn't hear any difference." In fact, one person thought the
dithered file (the best one) was the worst, and a few picked the
worst one (9 bits) as the best! So it seems that grit - whether
caused by analog tape or digital artifacts - can make a
recording sound "better" even if it is in fact worse in terms of
pure accuracy. Note that the totals below do not add up to 23
because most people got some right but missed others.
Number of people who admitted they heard no difference in any of
the files: 4
Number of people who thought the best file sounded worst: 1
Number of people who thought the worst file sounded best: 4
Number of people who correctly identified the two worst (11- and
9-bit) files: 9
Everyone who sent me their guesses has been sent the list of
files in order of degradation. I am not posting the order here,
so that others who want to try the test can still do so in the
future. After listening to the files, email me which you think
is which and I'll send the results by return email.
Failings of this test: Several people pointed out that the real
advantage of 24-bit recording is when an entire project is
recorded and mixed at 24 bits, because the increased resolution
helps minimize errors that accumulate from gain changes and
plug-in processing. At some point I hope to test that too. And
though nobody mentioned this, I'll add that another possible
advantage of 24-bit recording is the smoother decay of
instrument and reverb tails. Neither of these were tested here,
so all that this test can claim to show is how important 24 bits
and dithering are when used with full-level audio. However, this
experiment also shows that recording at an average level of -20
to avoid distorting - another claimed advantage of using 24
bits - is probably not as harmful to audio fidelity as many
would think.
Thanks to all who participated!
--Ethan
George Perfect
> Folks,
>
> A week ago I invited folks here to compare a series of audio
> files, to see how audible are the differences between dithering
> from 24 bits to 16 versus truncating. I also included files that
> were truncated to (approximately) 13, 11, and 9 bits, which
> correspond to recording at peak levels of -20, -30, and -40.
> There are five files altogether, as described on my web page at
> www.ethanwiner.com/bitstest.html. Here are results:
>
> I received 23 replies altogether, and only one person identified
> all five files correctly.
Ethan - did you ask what these people used to listen to the files?
Unless you know they were all using identical (very good) monitoring systems
your survey is as likely to measure the quality of people's playback systems
as the population's ability to distinguish between recordings at different
bit depths.
Just a thought.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
Andy
too, but haven't gotten around to doing the actual tests that will help me
decide. A lot of people just *assume* that 12 or 13 bits sounds terrible...
when on most program material, I bet it doesn't. There are some things that
are more sensitive to this, but I would imagine acoustic guitar would be one
of the more sensitive.
I think the comment about the monitoring is valifd though... most people
don't have great monitors hooked up to their computers.
Andy
Ethan Winer <nos...@nospam.com> wrote in message
news:RT4m8.321951$pN4.21...@bin8.nnrp.aus1.giganews.com...
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
Arny Krueger
news:3c98e106$0$8507$ed9e...@reading.news.pipex.net...
> Ethan - did you ask what these people used to listen to the files?
The tests were obviously very real-world - people listened with what
they had.
> Unless you know they were all using identical (very good)
monitoring systems
> your survey is as likely to measure the quality of people's
playback systems
> as the population's ability to distinguish between recordings at
different
> bit depths.
Requiring that the monitoring systems have to be "identical" (George,
I can't believe that use used that word, but you are "Perfect" aren't
you?) is totally unreasonable. It would be a reasonable requirement
if there was only one monitoring system in the world that could
possibly do the deed. Of course that's practically unreasonable too,
because if there was only one monitoring system in the world that
could do the deed, it would be one more can possibly exist at the
current state of the art.
In fact, the most significant variable in listening tests like this,
given that most people who'd take a test like this have at least
reasonably good monitoring systems; is the listening room,
particularly its acoustics and its background noise level. To do bit
resolution tests "right" you need a recording made in a very quiet
room, and you need to listen in a very quiet room. This idea is not
unique to me, it's been published for years by an international
standards body. Please see ITU Recommendation BS 1116-1. You can find
out more about this document at
http://www.itu.int/itudoc/itu-r/rec/bs/1116-1.html . The approximate
cost of this 26 page document in MS Word or PDF format is $12.00 US
as of 10/1/2000.
> Just a thought.
Just a red herring from someone who really doesn't understand the
issues that underlie tests like this, and based on his recent
behavior, seemingly doesn't want to learn.
Arny Krueger
"Andy" <ayc8N...@cornell.edu> wrote in message
news:3c98f...@corp.newsgroups.com...
> Thanks, that's very interesting... I've been thinking about going
to 24-bit
> too, but haven't gotten around to doing the actual tests that will
help me
> decide. A lot of people just *assume* that 12 or 13 bits sounds
terrible...
> when on most program material, I bet it doesn't.
Not if properly dithered. You can listen to musical samples at a wide
variety of bit depths by downloading files from
http://www.cdabx.com/technical/bits44/index.htm .
> There are some things that
> are more sensitive to this, but I would imagine acoustic guitar
would be one
> of the more sensitive.
Acoustic guitar is good, particularly if you record a string that has
been plucked and is naturally dying out in a very quiet room, and
miced close.
> I think the comment about the monitoring is valid though... most
people
> don't have great monitors hooked up to their computers.
IMO, they should have at least competent monitors, and they should
take positive steps to deal with the usual computer fan noises.
George Perfect
Your actions really are very childish indeed. I grew up with my surname and,
believe me, there isn't a way you can quote it or use it that would cause me
anything more than sympathy for your immaturity.
I deliberately stopped responding to you in the original 24 bit thread so
that everyone didn't have to hear your constant pointless arguing. So tell
us all what point you are trying to make here? (no - on second thoughts,
DON'T!) I have my doubts that it's anything to do with audio and everything
to do with bolstering your insecurities.
I think everyone (with one notable exception) who reads this group and read
my message is adult enough to understand the clear and simple point I
raised.
I am happy to let the other folks here form their opinions of my views - and
my behaviour.
You might like to remember that they also form a view of yours.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:p2km8.3648$4i2.67...@newssvr15.news.prodigy.com...
Ethan Winer
> Unless you know they were all using identical (very good)
monitoring systems <
I don't see how that is related at all. If a bunch of people
who are at least audio amateurs, if not professionals, have a
hard time distinguishing between 24 bits and 13 bits, I think
that says a lot all by itself. And who do we record music for
anyway? Mostly people whose playback systems are a LOT worse
than what the folks in these newsgroups and forums use.
--Ethan
George Perfect
>
> I don't see how that is related at all. If a bunch of people
> who are at least audio amateurs, if not professionals, have a
> hard time distinguishing between 24 bits and 13 bits, I think
> that says a lot all by itself. And who do we record music for
> anyway? Mostly people whose playback systems are a LOT worse
> than what the folks in these newsgroups and forums use.
>
Then my apologies - I misunderstood what you were trying to measure. Or
maybe I didn't.
The point I made was simple - don't assume that everyone participating in
your test has a listening environment capable of exhibiting the differences.
But if your test was a survey of what people *and* their playback systems
can distinguish then you have a valid result.
psychy
> "Ethan Winer" wrote
> >
> > I don't see how that is related at all. If a bunch of people
> > who are at least audio amateurs, if not professionals, have a
> > hard time distinguishing between 24 bits and 13 bits, I think
> > that says a lot all by itself. And who do we record music for
> > anyway? Mostly people whose playback systems are a LOT worse
> > than what the folks in these newsgroups and forums use.
> >
>
> Then my apologies - I misunderstood what you were trying to measure. Or
> maybe I didn't.
>
> The point I made was simple - don't assume that everyone participating in
> your test has a listening environment capable of exhibiting the
differences.
>
> But if your test was a survey of what people *and* their playback systems
> can distinguish then you have a valid result.
>
personally i'd aggree that the relevance is limited. what you've tested is
the resolution of the audio playback systems rather than the differences
between the rates.
it's a bit like viewing the mona lisa and a scribbled cartoon through a
green screen monitor in 300*200 res. and 16 colours. probably fairly
similar, but only because of the lack of "headroom" for the complexity.
the real validity is that it doesn't *matter* what you dither to for bog
standard playback, which is worth knowing. but what you started with is
going to become relevant after alot of processing - and don't we use high
sampling rates for this very reason?
~psychy..
Arny Krueger
> Arny - I see you have chosen to pollute another thread.
I look at it as an attempt at cleaning-up pollution that was already
there.
> Your actions really are very childish indeed.
As if this name-calling you just did is adult behavior.
> I grew up with my surname and,
> believe me, there isn't a way you can quote it or use it that would
cause me
> anything more than sympathy for your immaturity.
George, you inability to take well-intentioned and well-supported
correction created this situation. You think you are "Perfect", but
the facts show that's in name only.
> I deliberately stopped responding to you in the original 24 bit
thread so
> that everyone didn't have to hear your constant pointless arguing.
It is true that it's pointless to argue with you George since you put
yourself above all else. AES papers, various technical papers on the
web, proven expertise with digital, they all seem to mean nothing to
you.
>So tell us all what point you are trying to make here?
Simple - George, you make lots of mistakes when you endeavor to
criticize what you don't understand.
>(no - on second thoughts,
> DON'T!) I have my doubts that it's anything to do with audio and
everything
> to do with bolstering your insecurities.
George, if I was insecure would I proceed as boldly as I have?
> I think everyone (with one notable exception) who reads this group
and read
> my message is adult enough to understand the clear and simple point
I
> raised.
It was a simple point and it was simply wrong.
> I am happy to let the other folks here form their opinions of my
views - and
> my behavior.
George, you're the one who started personally attacking me last week.
Bottom line, you can dish it out, but you can't take it!
> You might like to remember that they also form a view of yours.
George, I'll stand by what I've said. I've tried to deal with you an
adult level and you corrupted the discussion with gratuitous personal
attacks last week. I'm not of a mind to let you continue your
baseless attacks on people who are trying to shed some light on this
matter like Ethan.
Dug
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:p2km8.3648$4i2.67...@newssvr15.news.prodigy.com...
I think George's point was that the test doesn't really conclude wether or not
people are capable of telling the difference as there are too many variables
(mainly the different playback hardware, such as monitors) I'd have thought a
someone like yourself would agree with that.
Arny, I can only assume that you're trying to fill some sort of void in your life.
Good luck to you,
Dug
George Perfect
>
> personally i'd aggree that the relevance is limited. what you've tested is
> the resolution of the audio playback systems rather than the differences
> between the rates.
> it's a bit like viewing the mona lisa and a scribbled cartoon through a
> green screen monitor in 300*200 res. and 16 colours. probably fairly
> similar, but only because of the lack of "headroom" for the complexity.
>
The analogy that sprang to my mind is it's like having an eye exam where the
optometrist asks you to look at the chart through a net curtain - you are
measuring only what you can see under those circumstances - not what you
*could* see given a clear view of the chart.
Suso Ramallo
I don't know wath is your problem with mr. Krueger, but I agree with Winer
and Krueger, is absolutely true.
I was tested in a studio enviroment (NC 20) the same takes (Original 24/96)
of a piano and vocals, recorded with a DPA mics and Earthwork
pre, and RME ADI 8 DS AD/DA (Genelec monitors). I dithered and resampling to
16/44 and resampling to 24/48. I started the test with a lot of experienced
engineers, and nobody had clear recognition. The next experience with this
take, was with a musicians (pianist included) and the same result, nobody
had a clear recognition.
Conclusion: 16/44 is enougth for listening, the same conclusion of Winer
test.
16 bits offers a 96 dB of dinamic
Bye
"George Perfect" <xgeo...@byline.co.uk> escribió en el mensaje
news:3c99fa33$0$8514$ed9e...@reading.news.pipex.net...
Arny Krueger
"Dug" <du...@dugf.co.uk> wrote in message
news:a7d977$cnk$1...@news8.svr.pol.co.uk...
> I think George's point was that the test doesn't really conclude
wether or not
> people are capable of telling the difference as there are too many
variables
> (mainly the different playback hardware, such as monitors) I'd have
thought a
> someone like yourself would agree with that.
I think that could have been George's point if he expressed that
thought.
> Arny, I can only assume that you're trying to fill some sort of
void in your life.
Dug, I can only assume that you are trying to fill some kind of void
in your life by jumping into a discussion that used to have nothing
to do with you. You don't have enough conflict in your life?
Dug
> Dug, I can only assume that you are trying to fill some kind of void
> in your life by jumping into a discussion that used to have nothing
> to do with you. You don't have enough conflict in your life?
Yes, the whole was boredom and you amuse me.
You could hardly call a usenet thread private, I have as much right to post as
anyone.
This is irrelevent though, you're diverting from the point.
This test shows that the average user with the average setup, cannot accurately
tell the difference,
which would have been my hypothesis had I made one.
What the test doesn't show, is wether or not the average user is capable of
hearing the difference.
Dug
(Krueger's 38 Stratagems For Winning Arguments, strategy #39 = contradict)
George Perfect
misunderstandings and no accusations of quoting you incompletely or out of
context. And I'll make it clear that this is the very last time I am going
to waste my time on you.
The facts (evidenced by your behaviour here and in other groups ) reveal you
as an attention seeker. A spammer, a troll and argument-stirrer of the first
order. One whose sanity, honesty and integrity has been questioned time
after time. Someone whose voluminous trail through the Usenet groups is
littered with example after example of respected experts discarding your
theories and notions. More than that, someone who makes up and changes the
rules of engagement as he goes along.
You are very fond of throwing out challenges to people to "prove" their
points and very adroit at ignoring such challenges when put to you or
changing the subject to avoid answering questions.
So, in a moment, in the simplest and most direct way I know how, I am going
to challenge you to substantiate the statements you make in your most recent
message. I will be fascinated to see how you respond.
But first I'd like to talk about why you offend me.
I see you still like to address me as "Perfect". Apart from your obvious
intent to score a cheap point using the name I inherited at birth (which as
I said reflects on you, not me) this behaviour is rude. If you like, I will
start addessing you as "Krueger". I don't think you'd like it "Krueger".
Because you see, "Krueger", it is aggressive and deliberately offensive. Get
the point "Krueger"?
Such behaviour is infantile. If you want to have the adult debate you say
you seek then first learn how to behave in polite adult company. Allow me to
remind everyone that you recently addressed another member of this group as
"Kraut" and made several racially offensive remarks in your attempt to
establish your superiority. I am not going to repeat your remarks here -
anyone who didn't see them will find the message in the archive.
Let's look at one of your pronouncements in this latest message:
> It is true that it's pointless to argue with you George since you put
> yourself above all else. AES papers, various technical papers on the
> web, proven expertise with digital, they all seem to mean nothing to
> you.
Er ... which of us puts himself above all else? Let's see ...
Your "evidence" against me is no more than your supposition that I did not
read the papers to which you referred. Prove it! As I quoted some of them
back at you I doubt you'll have an easy time. I rely on the record in the
archives.
And I challenge you to provide one shred of actual evidence that I put
myself above anyone else - even you.
But let's not talk about me - I know you find that boring. Let's talk about
you.
Specifically how you see your standing in the recording industry (never mind
this unassuming backwater).
Here's what you wrote about George Massenburg a few hours ago:
"They are like George Massenberg who has used the podium of the AES to
make fun of people and concepts they don't believe in, no matter how
solid and widely accepted those concepts and practices are. Now that
was professional of him, wasn't it?"
And here's a few of your thoughts on Bob Katz:
"I'm very sure that Katz doesn't do DBTs as a rule. He drops words
like "blind test" from time to time, but where the rubber hits the
road, nobody sees him as a serious player in the sensitive, reliable
listening test situation."
I assume your use of the word "nobody" excludes just about the entire U.S.
professional recording industry who regularly pay Mr Katz for the very
quality you claim he does not possess.
You have quite strong feelings about Mr Katz. Here's another snippet:
"He and his clique obviously and vociferously hate my guts and
everything my friends and I stand for. They are not the universally
worshipped audio gods that some make him out to be."
Tell us - is this just your opinion or is it a provable fact? Personally, I
doubt that Bob Katz would deem you worthy of enough attention to hate your
guts but that's a side issue. I also doubt that Bob Katz is (or has any
desire to be) a "universally worshipped audio god". I'll let others discover
for themselves why Mr Katz might just have very good reasons to think
unkindly of you.
These remarks are far from atypical of you. I invite anyone who can be
bothered to visit http://groups.google.com and search for your name to read
for themselves some of your other views on the industry leaders and
organisations whose work you nevertheless are happy to quote when it suits
you.
Wake up Arny. George Massenburg and Bob Katz *EARNED* the respect the
recording community affords them the hard way - by consistently building a
body of work that is the envy of their peers.
When you can get 1% of the number of people who respect either of these two
men to take you seriously then you can start to crow. Until then, all you
are doing is showing your ignorance and your arrogance.
I ask again - which of us puts himself above others?
The challenge was recently put to you in this group - if you are so darned
good, publish papers yourself. Subject yourself to peer review. Persuade
those you respect (you do respect *somebody*, dont you?) to take you
seriously. You obviously don't respect me so why waste so much of your
valuable time picking fights when you have all that measuring to do?
Now, let's see some proof. Leaving aside your more questionable claims -
such as your claim to be "well-intentioned" - let's have you substantiate
these specific statements you made:
1. You wrote: "As if this name-calling you just did is adult behavior."
Please quote the name I called you in the message I "just did".
2. You wrote "Simple - George, you make lots of mistakes when you endeavor
to criticize what you don't understand." Please review my posts to this
group and provide evidence of "lots of mistakes" - let's leave out the
recent digital audio thread because you and I are just never going to see
eye to eye on that one. Aw, what the heck - I'm in a generous mood, you can
have that one as your starter for ten. I'm a frequent poster here so you
have plenty of material to find "lots" of mistakes. Let's see them. Expose
me.
3. You wrote: "It was a simple point and it was simply wrong." Provide a
proof that the majority of people in this group will accept, that what I
said was wrong.
4. You wrote: "I'm not of a mind to let you continue your baseless attacks
on people who are trying to shed some light on this matter like Ethan."
Please quote the paragraph, comment, phrase or word where I make any kind of
"attack" on Ethan, his views, his exercise or anything about him.
5. As your point was general (you allege "people"), please quote *any*
evidence of a "baseless attack" I have made on anyone.
I do agree you are "not of a mind".
So, in wrapping up Arny, I hereby add my voice to the challenge. Get
yourself published in a respected journal. Earn the respect of your peers.
Earn some Grammys. Record some music for heaven's sake and let us all hear
how well your DBT-based approach translates into musical excellence.
Do these things and *then* come back here with your high and mighty attitude
and unswerving dogma and there's a chance I might take you seriously.
Until then, I shall continue to view you as a pompous, bad-mouthed, ignorant
and attention seeking prat.
Now *THAT* is me calling you names.
Let's see how many disagree with me.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:tppm8.10215$GK5.185...@newssvr16.news.prodigy.com...
George Perfect
My problem with Mr Krueger is nothing to do with whether he, I or you can
hear differences between 16 bit and 24 bit recordings.
For a better understanding of our differences, please refer to my other
recent post in this thread.
Bye
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Suso Ramallo" <jram...@worldonline.es> wrote in message
news:a7dabm$iec$1...@talia.mad.ttd.net...
Arny Krueger
> Arny - your entire message is quoted below again so there can be no
> misunderstandings and no accusations of quoting you incompletely or
out of
> context. And I'll make it clear that this is the very last time I
am going
> to waste my time on you.
The basic problem George is that you peed your pants in public, and
now you want to blame me for it.
> The facts (evidenced by your behavior here and in other groups )
reveal you
> as an attention seeker.
Why don't you try some introspection, George?
>A spammer, a troll and argument-stirrer of the first order.
Name-calling anybody?
>One whose sanity, honesty and integrity has been questioned time
after time.
That's one of the last resorts of the true scoundrel - claim that
only people who are insane would oppose your views.
> Someone whose voluminous trail through the Usenet groups is
> littered with example after example of respected experts discarding
your
> theories and notions.
Simply not true. There are people who are truly experts, and with
them I have no problems. There are people who like to put on airs and
say things that aren't true. With them, there might be some
difficulties.
> More than that, someone who makes up and changes the
> rules of engagement as he goes along.
Just a second George, first you claim that I'm consistent, and then
you claim that I'm not.
> You are very fond of throwing out challenges to people to "prove"
their
> points and very adroit at ignoring such challenges when put to you
or
> changing the subject to avoid answering questions.
The shoe fits George, wear it!
> So, in a moment, in the simplest and most direct way I know how, I
am going
> to challenge you to substantiate the statements you make in your
most recent
> message. I will be fascinated to see how you respond.
My response is that if you've got some questions you want me to
answer, then why don't YOU ask them, George?
> But first I'd like to talk about why you offend me.
I offend you because I don't bow to your false and mistaken claims,
George.
> I see you still like to address me as "Perfect". Apart from your
obvious
> intent to score a cheap point using the name I inherited at birth
(which as
> I said reflects on you, not me) this behavior is rude. If you like,
I will
> start addressing you as "Krueger". I don't think you'd like it
"Krueger".
> Because you see, "Krueger", it is aggressive and deliberately
offensive. Get
> the point "Krueger"?
Call me what you will, Mr. Perfect. Krueger surely does not offend
me.
> Such behavior is infantile.
That's one of the last resorts of the true scoundrel - claim that
only immature people oppose their views.
>If you want to have the adult debate you say
> you seek then first learn how to behave in polite adult company.
Good advice George. When you start acting on it in your own life,
I'll be very happy.
>Allow me to
> remind everyone that you recently addressed another member of this
group as
> "Kraut" and made several racially offensive remarks in your attempt
to
> establish your superiority.
Allow me to point out that that the person in question used that as
his "handle". George, are you desperate here or what?
> I am not going to repeat your remarks here -
> anyone who didn't see them will find the message in the archive.
The remarks were the kind of a remark that is entirely proper for one
"kraut" named "Krueger" to address to another "kraut" who is AFAIK
nameless.
> Let's look at one of your pronouncements in this latest message:
> > It is true that it's pointless to argue with you George since you
put
> > yourself above all else. AES papers, various technical papers on
the
> > web, proven expertise with digital, they all seem to mean nothing
to
> > you.
> Err ... which of us puts himself above all else? Let's see ...
> Your "evidence" against me is no more than your supposition that I
did not
> read the papers to which you referred.
Not at all. My point is that said AES papers supported my claims, and
George, you don't seem to be in agreement with them.
>Prove it! As I quoted some of them
> back at you I doubt you'll have an easy time. I rely on the record
in the
> archives.
George, since you are asking for proof of something that I never
claimed, I decline to provide proof.
> And I challenge you to provide one shred of actual evidence that I
put
> myself above anyone else - even you.
I'd say offhand George that your personal attacks from just this one
post of yours are completely sufficient. You've claimed that I'm
immature and that I'm insane. May I be so presumptuous as to suggest
that you don't think of yourself in those terms, George? I rest my
case!
> But let's not talk about me - I know you find that boring. Let's
talk about
> you.
> Specifically how you see your standing in the recording industry
(never mind
> this unassuming backwater).
I'm just a guy who has been curious about certain recording hardware
and has endeavored to satisfy that curiosity.
> Here's what you wrote about George Massenburg a few hours ago:
> "They are like George Massenberg who has used the podium of the AES
to
> make fun of people and concepts they don't believe in, no matter
how
> solid and widely accepted those concepts and practices are. Now
that
> was professional of him, wasn't it?"
> And here's a few of your thoughts on Bob Katz:
> "I'm very sure that Katz doesn't do DBTs as a rule. He drops words
> like "blind test" from time to time, but where the rubber hits the
> road, nobody sees him as a serious player in the sensitive,
reliable
> listening test situation."
> I assume your use of the word "nobody" excludes just about the
entire U.S.
> professional recording industry who regularly pay Mr. Katz for the
very
> quality you claim he does not possess.
George, where did I say that I'm a recording engineer who thinks he
is in competition with George Massenberg or B9d Katz?
> You have quite strong feelings about Mr. Katz. Here's another
snippet:
> "He and his clique obviously and vociferously hate my guts and
> everything my friends and I stand for. They are not the
universally
> worshipped audio gods that some make him out to be."
> Tell us - is this just your opinion or is it a provable fact?
I know for sure that one can find a lot of people who are recording
engineers and have strong exceptions to the antics of both men.
>Personally, I
> doubt that Bob Katz would deem you worthy of enough attention to
hate your
> guts but that's a side issue.
I think you ought to do a little more searching in Google Groups, Mr.
Perfect.
>I also doubt that Bob Katz is (or has any desire to be) a
"universally worshipped audio god".
A little hyperbole, perhaps?
> I'll let others discover
> for themselves why Mr. Katz might just have very good reasons to
think
> unkindly of you.
Sure, because I've been publicly critical of his writings about
jitter and dither. Because I think that he says a lot of things that
he wouldn't if he were well-informed by appropriate listening tests.
> These remarks are far from atypical of you. I invite anyone who can
be
> bothered to visit http://groups.google.com and search for your name
to read
> for themselves some of your other views on the industry leaders and
> organizations whose work you nevertheless are happy to quote when
it suits
> you.
When someone says something that fits, I quote it.
> Wake up Arny. George Massenburg and Bob Katz *EARNED* the respect
the
> recording community affords them the hard way - by consistently
building a
> body of work that is the envy of their peers.
The word "consistently" here would appear to be controversial.
> When you can get 1% of the number of people who respect either of
these two
> men to take you seriously then you can start to crow. Until then,
all you
> are doing is showing your ignorance and your arrogance.
That presumes that I want to be in a popularity contest Mr. Perfect.
I obviously don't.
> I ask again - which of us puts himself above others?
Mr. Perfect, there are such things as true facts, things that are
scientifically provable. People have misunderstandings and people
make mistakes. When the facts and the understandings don't meet, then
there is controversy.
> The challenge was recently put to you in this group - if you are so
darned
> good, publish papers yourself. Subject yourself to peer review.
Persuade
> those you respect (you do respect *somebody*, don't you?) to take
you
> seriously. You obviously don't respect me so why waste so much of
your
> valuable time picking fights when you have all that measuring to
do?
The problem I tried to address with you Mr. Perfect were your
misapprehensions about digital. You turned it into a personal fight
and now you are turning it into a vendetta.
> Now, let's see some proof. Leaving aside your more questionable
claims -
> such as your claim to be "well-intentioned" - let's have you
substantiate
> these specific statements you made:
> 1. You wrote: "As if this name-calling you just did is adult
behavior."
> Please quote the name I called you in the message I "just did".
This message Mr. Perfect, where you called me "insane" and
"infantile"
> 2. You wrote "Simple - George, you make lots of mistakes when you
endeavor
> to criticize what you don't understand." Please review my posts to
this
> group and provide evidence of "lots of mistakes" - let's leave out
the
> recent digital audio thread because you and I are just never going
to see
> eye to eye on that one. Aw, what the heck - I'm in a generous mood,
you can
> have that one as your starter for ten. I'm a frequent poster here
so you
> have plenty of material to find "lots" of mistakes. Let's see them.
Expose
> me.
I've already addressed those items in context. I see no need to
catalog and consolidate them. Please do your own homework, Mr.
Perfect.
> 3. You wrote: "It was a simple point and it was simply wrong."
Provide a
> proof that the majority of people in this group will accept, that
what I
> said was wrong.
I've already addressed those items in context. I see no need to
catalog and consolidate them. Please do your own homework, Mr.
Perfect.
> 4. You wrote: "I'm not of a mind to let you continue your baseless
attacks
> on people who are trying to shed some light on this matter like
Ethan."
> Please quote the paragraph, comment, phrase or word where I make
any kind of
> "attack" on Ethan, his views, his exercise or anything about him.
I've already addressed those items in context. I see no need to
catalog and consolidate them. Please do your own homework, Mr.
Perfect.
> 5. As your point was general (you allege "people"), please quote
*any*
> evidence of a "baseless attack" I have made on anyone.
I've already addressed those items in context. I see no need to
catalog and consolidate them. Please do your own homework, Mr.
Perfect.
> I do agree you are "not of a mind".
That would agree with your former claims that I am "insane", Mr.
Perfect.
> So, in wrapping up Arny, I hereby add my voice to the challenge.
Get
> yourself published in a respected journal. Earn the respect of your
peers.
> Earn some Grammys. Record some music for heaven's sake and let us
all hear
> how well your DBT-based approach translates into musical
excellence.
I have no interest in winning Grammys or even making recordings that
would relate to such a thing.
> Do these things and *then* come back here with your high and mighty
attitude
> and unswerving dogma and there's a chance I might take you
seriously.
George, you can do what you will to me, but you can't do as well as
you could if you continue to contradict the science that I've been
reciting to you.
> Until then, I shall continue to view you as a pompous, bad-mouthed,
ignorant
> and attention seeking prat.
BTW Mr. Perfect, these accusations and instances of name-calling
would relate to your question (1), above, would they not?
> Now *THAT* is me calling you names.
As were the former instances involving insanity and immaturity.
> Let's see how many disagree with me.
I think your own words have damned you sufficiently, Mr. Perfect.
Good job! ;-)
Dug
I sawe the error of my ways a while back, It's been too hard to resist arguing
with you Arn (as you intended), but I'll no lo longer rise to it.
Enough arguing, can we just agree that some of us are not willing to accept Arny's
bible-like recitations in blind faith, and that whatever the truth is, it is
discovered in a more constructive way.
You've had your five minutes of notoriety Mr. Krueger, kindly put your spoon away
and record something useful.
Dug
Suso Ramallo
Hello,
> I or you can
> hear differences between 16 bit and 24 bit recordings.
I can't, I tried (Explained in my other post).
> For a better understanding of our differences, please refer to my other
> recent post in this thread.
Sorry, but my news server don't have your old posts.
Bye
George Perfect
I asked very specific questions and once again you have refused to answer
them or attempted to side step the issues while throwing more confusion into
the pot.
I note that John Atkinson is still waiting for you to respond to his very
precisely worded questions - some four months after you made unfounded
allegations concerning Stereophile's jitter testing procedures, and Glenn
Zelniker still awaits retraction of your defamatory remarks - in which you
inferred he had exchanged a favourable review of one of his company's
products in return for providing untrue evidence that overturned your theory
that Stereophile's jitter test was faulty due to lack of dither (and where
have we heard that argument lately?)
Can I remind you Arny? You were proven wrong. Not on one minor point but in
everything you had said and alleged.
And not by one man's opinion but by an entire group of designers and
engineers working for competing organisations.
It was put to you that you should retract and apologise. To your shame you
did not. Instead, you high-tailed it out of there and brought your daft
notions and unpleasant practices here.
Returning to present events. I specifically quoted your entire message so
that your words could be seen in context and there could be no confusion.
I challenge you again to substantiate the clear statements you made.
1. You accused me of calling you names in the message to which you replied.
What names?
2. You accuse me of making "lots of mistakes". Specify them.
3. You state that the point I made concerning the results of Ethan's test
was "wrong". Prove it.
4. You accuse me of mounting a "baseless attack" on Ethan and on "people"
unnamed. Provide the evidence. Show us the words I used. Live by your own
standards for once - I don't accept you have already covered this in
context. Because you have not. It's a simple question. Answer it.
5. You have alleged in several posts to this group that I did not read the
AES papers whose references you gave. Prove it.
These are *your* allegations - made without prompting. I am challenging you
to do as you demand of others. Substantiate your claims. Each and every one
of them.
These are not difficult questions to answer. The message in which I
supposedly called you names is just a few lines in length. It's not much to
ask you to quote from it. My earlier reply to Ethan was even shorter - it
can take no time at all to quote the passage containing my "baseless
attack". It took me just a short while to find hundreds of Usenet messages
damning you to hell and back. You have set yourself the far simpler
challenge of finding merely "lots" of mistakes I am supposed to have made.
So come on Arny - you claim to be a man of truth and science.
Put up. Or shut up.
As to your recent racial slurs, the proof of the pudding is in the eating.
The person to whom you addressed your remarks was deeply offended. This is a
matter of public record. You have not apologised.
Turning to George Massenburg and Bob Katz, I asked several specific
questions which you have chosen to ignore. To the earlier insults you hurled
at them, you now add this:
> I know for sure that one can find a lot of people who are recording
> engineers and have strong exceptions to the antics of both men.
I challenge you again. Prove it. Name names. Cite references. Let us hear
from these "people who are recording engineers".
I tell you this much. For every name you can come up with I will provide ten
to turn out in favour of Mr Massenburg and Mr Katz..
Of course, your inate superiority is so well established that you don't need
the respect of others. Just as well really.
> That presumes that I want to be in a popularity contest Mr. Perfect.
> I obviously don't.
I have news for you. You don't qualify to *enter* a popularity contest.
Face it Arny, you are almost alone in your views and opinions. You have run
your notions and theories past recording professionals, magazine editors,
equipment designers, engineers and just about everybody else.
The consensus? You are nuts and your theories amount to so much hot air.
I stand on several years worth of archived public records in making this
statement.
Much of the advice you hand out with such authority is just lain wrong -
like your suggestion that someone should "phase invert" a guitar to have it
appear out of both left and right speakers. Ignoring any notion of mono
compatibility, of course. Your response when your error was pointed out to
you? You told the eminent recording engineer and studio owner he was wrong!
Ha!
In fact, if this were a painting group and Leonardo Da Vinci were posting
here you'd be the one telling him that he was wasting his time using such a
fine brush when painting the Mona Lisa because your double-blind viewing
tests proved that an audience of your friends sat behind a curtain couldn't
make out the girl's face anyway!
Like any fanatic, you have convinced yourself that you are right and the
rest of the world doesn't understand you. You would get my sympathy if you
weren't such a nasty character and your activities didn't do so much damage
to the groups you barge into.
Turning back to the topic that started all this, you *STILL* haven't
addressed the original points I made and the questions I asked other than to
repeat over and again that you are right and I am wrong.
I hate to spoil the party by stating the blindingly obvious that I do not
lack solidly based support for the statements I made. Both the practical
and the theoretical aspect of those statements have been set out, examined,
turned over and shown to hold up.
All you can do is recycle over and again the work of others whom, as soon as
it suits you, you will in turn insult. Research papers are just that - they
document research and findings. If well executed they can present the best
understanding we might have at the time they are written.
But you have to place them in context. You have claimed that the papers you
cite are "proof" of your statements while claiming that those cited by
others that counter your arguments are incorrect. You cannot have it both
ways.
Push has come to shove. Time to put up the ante my friend. We needn't bring
in any outside experts here - all the evidence we need is on the table in
front of us.
So, answer my questions. They aren't hard and you made the statements to
which they relate. So prove it. Substantiate your statements and your
allegations against me.
If you can't, I'll rely on the group to decide on the true worth of your
opinions.
Arjan P
>
> Just a red herring from someone who really doesn't understand the
> issues that underlie tests like this, and based on his recent
> behavior, seemingly doesn't want to learn.
I knew it right from the header: Here we go again... Another thread to kill
right away! Mr. Krueger, I'll ask it once again: please leave this formerly
almost flameless group, and let it get back to that state...
Luck, Arjan
-----To reply via email remove "at." from address.-----
---[This is a spam evasion manouvre, not to bug ya.]---
Where I come from: http://www.bootlegs.demon.nl/studio/
Arny Krueger
> Oh no Arny - you really do not get off that easily.
Well, what I did is get you to make a liar out of yourself yet again,
George. Do you remember just writing:
"And I'll make it clear that this is the very last time I am going to
waste my time on you."
;-)
> I asked very specific questions and once again you have refused to
answer
> them or attempted to side step the issues while throwing more
confusion into
> the pot.
No George you have not asked me sufficiently specific questions.
You've tried to force me to do your homework for you. If you've got a
specific question ask it. If you want a research service, find one
and hire it.
> I note that John Atkinson is still waiting for you to respond to
his very
> precisely worded questions - some four months after you made
unfounded
> allegations concerning Stereophile's jitter testing procedures, and
Glenn
> Zelniker still awaits retraction of your defamatory remarks - in
which you
> inferred he had exchanged a favourable review of one of his
company's
> products in return for providing untrue evidence that overturned
your theory
> that Stereophile's jitter test was faulty due to lack of dither
(and where
> have we heard that argument lately?)
George, you are really desperate for attention aren't you? Why are
you poking your nose into a dispute I have with someone else in
another newsgroup? Besides, I've answered Mr. Atkinson's questions a
number of times. He just doesn't like my answers.
> Can I remind you Arny? You were proven wrong. Not on one minor
point but in
> everything you had said and alleged.
That's a false claim. Mr. Atkinson uses selective quoting and twists
of meanings to try to make issues where they don't exist.
> And not by one man's opinion but by an entire group of designers
and
> engineers working for competing organisations.
George, you are really desperate for attention aren't you? Why are
you poking your nose into a dispute I have with someone else in
another newsgroup? You've obviously despaired of prevailing in your
controversy with me, if you are dredging up this irrelevant stuff.
> It was put to you that you should retract and apologise. To your
shame you
> did not. Instead, you high-tailed it out of there and brought your
daft
> notions and unpleasant practices here.
The fact of the matter is that I haven't high-tailed it out of
anyplace. In fact, Mr. Atkinson is current evading some questions of
mine.
> Returning to present events. I specifically quoted your entire
message so
> that your words could be seen in context and there could be no
confusion.
So?
> I challenge you again to substantiate the clear statements you
made.
Asked and answered.
> 1. You accused me of calling you names in the message to which you
replied.
> What names?
Asked and answered.
> 2. You accuse me of making "lots of mistakes". Specify them.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> 3. You state that the point I made concerning the results of
Ethan's test
> was "wrong". Prove it.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> 4. You accuse me of mounting a "baseless attack" on Ethan and on
"people"
> unnamed. Provide the evidence. Show us the words I used. Live by
your own
> standards for once - I don't accept you have already covered this
in
> context. Because you have not. It's a simple question. Answer it.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> 5. You have alleged in several posts to this group that I did not
read the
> AES papers whose references you gave. Prove it.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> These are *your* allegations - made without prompting. I am
challenging you
> to do as you demand of others. Substantiate your claims. Each and
every one
> of them.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> These are not difficult questions to answer. The message in which I
> supposedly called you names is just a few lines in length. It's not
much to
> ask you to quote from it. My earlier reply to Ethan was even
shorter - it
> can take no time at all to quote the passage containing my
"baseless
> attack". It took me just a short while to find hundreds of Usenet
messages
> damning you to hell and back. You have set yourself the far simpler
> challenge of finding merely "lots" of mistakes I am supposed to
have made.
> So come on Arny - you claim to be a man of truth and science.
> Put up. Or shut up.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> As to your recent racial slurs, the proof of the pudding is in the
eating.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> The person to whom you addressed your remarks was deeply offended.
This is a
> matter of public record. You have not apologised.
He calls himself "Kraut". Where's the beef?
> Turning to George Massenburg and Bob Katz, I asked several specific
> questions which you have chosen to ignore. To the earlier insults
you hurled
> at them, you now add this:
> > I know for sure that one can find a lot of people who are
recording
> > engineers and have strong exceptions to the antics of both men.
> I challenge you again. Prove it. Name names. Cite references. Let
us hear
> from these "people who are recording engineers".
I feel no need to bring the names of a number of people who are
uninvolved with our dispute into this matter.
> I tell you this much. For every name you can come up with I will
provide ten
> to turn out in favour of Mr. Massenburg and Mr. Katz..
I don't recall that I was trying to start a popularity contest.
> Of course, your innate superiority is so well established that you
don't need
> the respect of others. Just as well really.
> > That presumes that I want to be in a popularity contest Mr.
Perfect.
> > I obviously don't.
> I have news for you. You don't qualify to *enter* a popularity
contest.
Not news, I repeat, I don't want to be in a popularity contest.
> Face it Arny, you are almost alone in your views and opinions.
Not hardly. A number have posted that they agree with my claims in
our original dispute about how digital audio works.
>You have run
> your notions and theories past recording professionals, magazine
editors,
> equipment designers, engineers and just about everybody else.
These aren't my notions. You ought to read some good books about
digital - say Ken Pohlman's book.
> The consensus? You are nuts and your theories amount to so much hot
air.
Nope, they are motherhood and apple pie.
> I stand on several years worth of archived public records in making
this
> statement.
Not news, I repeat, I don't want to be in a popularity contest.
> Much of the advice you hand out with such authority is just lain
wrong -
> like your suggestion that someone should "phase invert" a guitar to
have it
> appear out of both left and right speakers. Ignoring any notion of
mono
> compatibility, of course. Your response when your error was pointed
out to
> you? You told the eminent recording engineer and studio owner he
was wrong!
I run into recordings that do this all the time. Some people can be
dogmatic about mono compatibility and other people can break
so-called such rules whenever they want to.
> Ha!
> In fact, if this were a painting group and Leonardo Da Vinci were
posting
> here you'd be the one telling him that he was wasting his time
using such a
> fine brush when painting the Mona Lisa because your double-blind
viewing
> tests proved that an audience of your friends sat behind a curtain
couldn't
> make out the girl's face anyway!
George, I'm not responsible for your ludicrous fantasies.
> Like any fanatic, you have convinced yourself that you are right
and the
> rest of the world doesn't understand you. You would get my sympathy
if you
> weren't such a nasty character and your activities didn't do so
much damage
> to the groups you barge into.
Much of the rest of the world understands me quite well and I
understand them. It's just people like you Mr. "Perfect" that don't
know when to quit.
> Turning back to the topic that started all this, you *STILL*
haven't
> addressed the original points I made and the questions I asked
other than to
> repeat over and again that you are right and I am wrong.
That's fine with me, Mr. Perfect. I tried.
> I hate to spoil the party by stating the blindingly obvious that I
do not
> lack solidly based support for the statements I made.
You surely do, Mr. Perfect.
> Both the practical
> and the theoretical aspect of those statements have been set out,
examined,
> turned over and shown to hold up.
They surely don't, Mr. Perfect.
> All you can do is recycle over and again the work of others whom,
as soon as
> it suits you, you will in turn insult. Research papers are just
that - they
> document research and findings. If well executed they can present
the best
> understanding we might have at the time they are written.
Several of the author's I quoted, like Vanderkooy and Lipshitz are
personal friends not to mention well-known authorities.
> But you have to place them in context. You have claimed that the
papers you
> cite are "proof" of your statements while claiming that those cited
by
> others that counter your arguments are incorrect. You cannot have
it both
> ways.
What AES papers have you produced Mr. Perfect that contradict my
claims?
> Push has come to shove. Time to put up the ante my friend. We
needn't bring
> in any outside experts here - all the evidence we need is on the
table in
> front of us.
Mr. Perfect, I've presented good prima-facae evidence. You've
declined to believe it. No skin off my nose.
> So, answer my questions. They aren't hard and you made the
statements to
> which they relate. So prove it. Substantiate your statements and
your
> allegations against me.
I've already addressed those items in context. I see no need to
catalog and consolidate them.
> If you can't, I'll rely on the group to decide on the true worth of
your
> opinions.
I repeat, I don't want to be in a popularity contest.
George Perfect
I will merely note that having been asked (politely) for the second time to
answer several straightforward and simple questions and to substantiate the
allegations and claims that you made you have refused to do so.
The dispute you have with John Atkinson is not irrelevant. It clearly shows
the pattern of your behaviour.
As to Bob Katz - a man whose hearing is generally accepted to be exemplary -
your "problem" with him is that he can hear things you can't. That he
refuses to participate in one of your self-publicising "double blind
listening tests" does not justify the public vilification to which you
continually subject him.
You have an odd view of "truth" and a very low standard of "proof" when it
suits you.
For my part (and I see I am not alone here) I know all I want or need to
know of your views.
You have now been asked several times to leave this group.
I, for one, cannot wait for you to do so.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:Wmvm8.10339$0q4.188...@newssvr16.news.prodigy.com...
John Atkinson
> "George Perfect" <xgeo...@byline.co.uk> wrote in message
> news:3c9a67e8$0$8511$ed9e...@reading.news.pipex.net...
> > his very precisely worded questions - some four months after you
> > made unfounded allegations concerning Stereophile's jitter testing
>
> I've answered Mr. Atkinson's questions a number of times. He just
> doesn't like my answers.
my repeated and specific questions concerning the unsupported
allegations he made about supposed unethical and dishonorable behavior
on my part was to tell me to stop my "bleating" as people were tired of
it. :-(
Concerning this newsgroup's name, I am currently recording a choir in
Minnesota in _6_ tracks of 24-bit. Over 100GB of data so far, and
there's one more session to go!
John Atkinson
Editor, Stereophile
88.2kHz super fidelity.
Van
I've been a songwriter a couple decades. I've been updating my technology.
I've got product that was created 13-15 years ago and am now exploring my newer
recording and music equipment.
I've been talking to some guys who use ProStudio on MAC G4 for their DAWs and
it's clean and rich sound. They insist I should run out and buy a similar
setup, but I'm not so sure.
My site (http://www.dedserius.com/) has mp3s of Wasted Tears and every song on
it. That CD was created using:
Yamaha 120S 4-track tape recorder,
Yamaha FB-01 8-voice poly synth module (stop laughing),
'91 Yamaha Electric Guitar,
'90 ibanez UE-303 B multi-effects pedal,
L.A. Metal pedal >:),
Peavey KB-100 amps (used for reverb on all vocal tracks),
Yamaha RX-17 drum machine,
audio-technica Pro 3H mic,
commoder 128 with Dr. T's Music Sequencing software. >:)!
Along with the above equipment, I have acquired the following gear:
Many i386 PC's and 1 DEC Alpha # All but one run Linux/UNIX; No Macs, yet!
Cakewalk on Win2k,
Ovation "Applause" round-body accoustic guitar with pickup,
SoundBlaster 128-bit Platinum card,
Korg 01/rw 32-voice poly synth module with drum kits,
Fostex FD-4 digital multi-track recorder,
Peavey C8P master controller.
My immediate goal is to re-master 6 tracks from Wasted Tears to a quality decent
enough to sell at gigs as a CD. I've begun mastering one of them and here's
what I've done so far:
1. MIDI'ed the drums and bass on the Korg;
2. Panned MIDI channel 10 on the Korg to use the Korg's drum kit pan assignments
and tracked 1 and 2 of the Fostex FD-4 to the Korg's #1 and #2 1/4" outputs, as
the drums;
3. Panned MIDI channel 1 on the Korg to #3 1/4" output and into FD-4 track 3, as
the bass;
4. Panned MIDI channel 2 on the Korg to #4 1/4" output and into FD-4 track 4,
for the keyboard parts;
5. Attached the UE-303 to the Aux send and return and turned on the flanger and
compressor and set the pot in the aux section on track 3 of the FD-4 to about
midway toward Aux1, which appears to apply about half the compressor effect on
the UE-303 as I play it back.
After getting tracks 1-4 down on the FD-4, I'll mix the drums/bass/keys down to
2 rehearsal tracks and put down vocals and guitars on the freed up tracks.
TIA for your insights/advisements on my (potential) misgivings. Any
recommendations on which songs should be prioritized would be extremely helpful,
too. >:)
Best Regards,
Van
--
=================================================================
Linux rocks!!! http://www.dedserius.com/
=================================================================
George Perfect
I've sent a short reply to your email to me - I'd rather answer some of your
questions here in group.
It looks as if you are re-recording your older songs - strictly speaking
'mastering' is a process that comes after all the recording and mixing is
finished and might apply here only if you were just trying to polish up the
original recordings.
I listened to Wasted Tears and Violet Blues. They both sound good for songs
that were recorded on a cassette multi-tracker - you've obviously got the
basics well in hand.
I like your songs and they deserve the better presentation you are trying to
give them. Perhaps it would help if I set out how I would approach the task
(with one eye firmly fixed on keeping the budget under control!)
Let's first look at what I think the songs need:
1. Better instrumentation - the drum and keyboard sounds are very
'machine-like' - better quality sounds and a looser rhythm would both go a
long way to letting these songs breathe.
2. Better vocal quality. I have an idea of the vocal sound you are trying
for. Nothing that a decent mic and a decent vocal channel (or pre-amp /
compressor combination) wouldn't solve
3. Better effects - the reverb and other effects you are using betray their
origins.
If I were you, I would beg, borrow or swap some of my existing equipment for
a better computer as the starting point to constructing a better DAW. You
don't need anything too fancy - you'd find a second-hand PII/300MHz enough
for what I'm going to suggest - try for one with a decent motherboard (ASUS
is my favourite) with at least UDMA 66 (pref. UDMA100) disk controllers.
Read through the motherboard newsgroups to get a quick feel for which boards
had problems with the early UDMA 100 interfaces etc.
Ask around (especially with your background) and you'll be sure to find
someone throwing a machine with spec similar to this out or selling it off
for $100 or less.
Your SoundBlaster is not ideal but it will do for now - and it has the
advantage of its soundfont capability which you might find very useful.
Computer in place, you now have the room to explore more of Cakewalk's
functionality.
Better keyboard sounds are available through soft instruments (DXi and VSTi
formats) - visit http://www.kvr-vst.com/index.php for a good overview of
what is available. While you are there, take a look at the samplers and drum
machines/samplers as well.
You could also usefully look at the huge range of soundfonts available. Jrem
and Knud are two regulars here who make far more use of soundfonts than I
ever have so I'll let them chip in with advice.
I believe any mix of soft instruments / samplers / soundfonts will give you
a greater range of higher quality sounds for your songs.
Turning to vocals, a decent mic is the starting point. The baseline is
surely (sorry about the pun) the Shure SM58. This is a dynamic mic and needs
no phantom power. It is also easy to use (good pattern, built-in pop shields
etc.). You will never regret buying an SM58.
Whichever mic you choose, you will need a preamp to raise its output to line
level. Do not be tempted to use the mic input of the SoundBlaster - it is
the most awful device known to man. You could use the mic stage of your
Yamaha cassette unit or the FD4 but better quality will definitely come if
you use a purpose designed preamp.
You will also need a good compressor on your voice while recording.
Listening to Violet Blues the distortion on voice at several points is very
obvious. Even if you get the levels under control, a good compressor will
alter (hopefully improve) the quality of your vocal recordings and get you
closer to the sound I think you are trying to achieve.
If you were using a better quality soundcard, one approach with a digital
system would be to track the vocals dry (ie, no compression) then compress
later in the mix. But the noise floor of the SoundBlaster is so high that I
wouldn't recommend this approach as you will bring the noise up even further
as the compressor's make-up gain is used.
BTW - don't over-compress during recording. It's an easy thing to do but
compression (like reverb) can't be undone later. You want only enough
compression to keep the levels in hand - you can add more compression later
to get the sound you want.
An often recommended small preamp is the Presonus Blue Tube. Couple this
with an RNC compressor and you have what many here would consider more than
adequate.
Alternatives you could consider are the new Focusrite TrakMaster channel
(preamp, compressor and EQ combined) or one of the smaller Joe Meek units -
though I'd advise you not to buy a Meek unit without listening to it on your
voice first - they have a distinct 'colour' all their own which you may love
or hate.
Any of these preamps/channels opens up a wider choice of mics. Something
like a MXL2001P or MXL V67G might suit your voice well. These condensor mics
need phantom power (see the FAQ at http://go.to/recordingfaq) to operate.
They would also do a more than adequate job of recording (say) acoustic
guitar - something an SM58 might struggle with.
With this little lot in place you should be able to:
a) record keyboards / drums in MIDI with a good range of quality sounds -
application of a "groove" or (heven forbid!) actually playing the drum
sampler 'live' will give a more human feel to your songs.
b) multi-track and layer sounds without having to bounce as you do at
present - you might, for example, find it useful to double-track your vocals
on some songs.
c) apply better effects - even Cakewalk's bundled plug-ins should sound
better than those you use at present. As time and funds allow you can easily
add better reverbs/delays/flangers. The quality of some of the free plug-ins
is impressive in any case.
If your version of Cakewalk is old, you might also consider n-Track software
from FA Soft (www.fasoft.com) which is used with great affection by many
here.
I couldn't suggest you give even a passing glance at ProTools. It is a
closed, proprietary system that requires very expensive hardware to run. It
would not provide you with anything you can't do at least as well and far
cheaper.
Hopefully this gives you at least some food for thought. I hope I have set
out a road map you might be able to travel over time.
Again, welcome to am4t - let the adventure begin!
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Van" <vanb...@dedserius.com> wrote in message
news:3C9ADF12...@dedserius.com...
nickspicks.com
I would agree about the play back gear though.
if the point of the test is to compare everyones ears, then identical
playback gear is the correct way to go about an experiment like this.
All things must be equal for all participants. Otherwise, its' a "neat"
test, but bad science. Too many uncontrolled variables.
my $.02
Arny Krueger
"Ethan Winer" <nos...@nospam.com> wrote in message
news:Lfnm8.335408$pN4.22...@bin8.nnrp.aus1.giganews.com...
> > Unless you know they were all using identical (very good)
> monitoring systems <
> I don't see how that is related at all.
Ethan, you are running into a classic situation that has been
repeated dozens of times in the past.
If your test had a positive outcome, nobody would be crabbing about
other people's systems.
When people claim that they've heard dramatic improvements with their
new 24 bit card, nobody is going to crab about whether or not their
system was good enough to possibly hear the difference. In fact,
that's the nature of the brag: "My ears and system are so good that I
can hear the difference 24 bits makes."
The other thing is that some people are going to hear dramatic
improvements with their new 24 bit card. This can easily happen if
their previous equipment was very inadequate or defective. For
example, I have a CD player whose analog output sounds "thin" because
some capacitors had became defective. I measured it when it was new
and it was just fine. When I swapped it out and put in a new DVD
player with 24 bit converters, there is a dramatic improvement in
sound quality via the analog outputs. (Of course I had been using the
defective player's digital outputs for several years!) I'm sure that
I can hear a loss of 3 dB at 100 Hz and more than that at lower
frequencies in a blind test. But, what does this prove except that
the old equipment had become defective?
That points out another problem with user testimonials based on
casual listening. Nobody knows for sure what was really happening at
the time of the test. In contrast, tests based on distributing sound
files at least hold the hardware environment constant.
> If a bunch of people
> who are at least audio amateurs, if not professionals, have a
> hard time distinguishing between 24 bits and 13 bits, I think
> that says a lot all by itself.
Of course it does. But, you've got a lot of personal agendas running
in the background.
For example, people who believe that present-day digital equipment
lacks the sound quality that they've grown accustomed to have often
crystallized their misapprehensions about digital around the idea
that 16 bits isn't enough. After all, the high end audio magazines
repeated this every which way but loose for over a decade. Their
belief that 16 bit digital lacks the resolution of analog is
vindicated if suddenly 24 bit ushers in a new millennium. They can
also get off their tired analog horses without admitting that they've
been wrong about 16 bit.
> And who do we record music for anyway? Mostly people whose
playback systems are a LOT worse
> than what the folks in these newsgroups and forums use.
I think that we shouldn't excuse the use of inadequate monitoring
systems on the inadequacies of other people's systems. I always favor
targeting the best possible system, and that's what I do with the
work that I do.
Dug
"nickspicks.com" <nos-pa...@nickspicks.com> wrote in message
news:2HFm8.17413$26.11...@typhoon.maine.rr.com...
Ditto.
Arny Krueger
"John Atkinson" <stereophi...@compuserve.com> wrote in message
news:9b0b3a02.02032...@posting.google.com...
> "Arny Krueger" <ar...@hotpop.com> wrote in message
news:<Wmvm8.10339$0q4.188...@newssvr16.news.prodigy.com>...
> > "George Perfect" <xgeo...@byline.co.uk> wrote in message
> > news:3c9a67e8$0$8511$ed9e...@reading.news.pipex.net...
> > > I note that John Atkinson is still waiting for you to respond
to
> > > his very precisely worded questions - some four months after
you
> > > made unfounded allegations concerning Stereophile's jitter
testing
> > > procedures...
> >
> > I've answered Mr. Atkinson's questions a number of times. He just
> > doesn't like my answers.
>
> Actually, this is incorrect. The only answer Mr. Krueger made to
> my repeated and specific questions concerning the unsupported
> allegations he made about supposed unethical and dishonorable
behavior
> on my part was to tell me to stop my "bleating" as people were
tired of
> it. :-(
Mr. Atkinson, you seem to be very much in the need of a few answers,
answers that you think I have. Your story has touched my heart, and
I've decided to extend you some undeserved courtesy.
I'll answer 5 questions of yours, subject to the following
conditions:
(1) The questions have to be questions from John Atkinson, and not
arrived at with the past or present assistance of anybody else.
(2) The questions have to be questions that I can reasonably be
expected to answer without travel, research, consultation other
people or with various archives or references. For example, they
can't relate to odd specifics of audio gear that I don't have
personal experience with. The list of relevant audio gear can be
found at www.pcavtech.com .
(3) The questions have to be on-topic in the context of this forum,
which is audio recording and playback.
(4) The questions have to be closely related to digital audio
technology. The questions can be questions that you've asked me
before.
(5) Any questions that don't fit these rules will simply be ignored,
but will count against the allotment of 5 questions.
Dug
You list-maker you. ;)
Ethan Winer
> The point I made was simple - don't assume that everyone
participating in your test has a listening environment capable
of exhibiting the differences. <
Here's the real issue: If only a very small percentage of people
can hear the improvement, and they need to raise the level by 40
dB. as a reverb tail is dying out, and they have to use a
particular brand of loudspeaker or whatever, then how important
is it really? I mix music for people to listen to on whatever
system they have.
My real objection to the claims for various high-end "tweaks"
like 24 bits, 96 KHz., expensive cables (sorry,
"interconnects"), low-jitter outboard (and of course expensive)
word clock generators, are:
1. This is a consumerist issue, where a lot of newbies and
not-so-newbies who don't know any better are sold a bill of
goods by audio dealers who tell them, "The reason your mixes
suck is because your Mackie pres are so lame. Son, spend a
couple of grand on this here fine product and you'll hear an
enormous improvement." Where enormous is the keyword. As if the
newbie's problem is inadequate mike pres or cables or word clock
(a personal fave), as opposed to the REAL problem which is lack
of experience and (probably) a terrible sounding room.
2. This nonsense gets passed around faster than a rumor at an
old-age home. It amazes me that MOST people spout these "facts"
as if they were actually true. I follow a lot of groups every
day, and I have seen many people make claims that they obviously
are just parroting.
3. Just as galling as the dealers who profit from
misunderstanding, the biggest spreaders of fiction are the audio
magazines. When I wrote my "Audio Myths" article for Audio
Media magazine a few years ago, I pitched it to Electronic
Musician first. Scott Wilkinson at EM told me the editors
looked at my ten points, and "disagreed" with every one of them!
How do you disagree with scientific fact?!
4. I see a DIRECT relationship between a belief in audio "magic"
and a belief in astrology, tarot cards, homeopathy, and an
endless list of other nonsense. The is perhaps the most
important issue of all: Most people don't know anything about
scientific method, and they don't want to learn. "My mind is
made up, don't confuse me with the facts." Or, in this case,
with what is practical.
Do 24 bits give better resolution than 16? Absolutely. Is it
possible to design an external word clock box that has lower
jitter than the stock crystal oscillator in a decent pro audio
sound card. Of course, given the budget. Do some speaker cables
have less capacitance than others? Sure. But do any of these
things matter? More to the point, are they worth paying extra
for?
--Ethan
Ethan Winer
> If your test had a positive outcome, nobody would be crabbing
about other people's systems. <
No fooling!
> That points out another problem with user testimonials based
on casual listening. Nobody knows for sure what was really
happening at the time of the test. <
For sure, and most tests are so "casual" as to be useless. Like
comparing microphones using different performances. I saw a
mike shootout recently where each mike was tested with a
different performance on an acoustic guitar. But you can change
the high-end a LOT by just strumming a little closer to or
farther from the bridge! Or picking harder or softer.
> But, you've got a lot of personal agendas running in the
background. <
Yes, but I'm bothered most by all the personal attacks and name
calling.
Can't we all just discuss the issues? :->)
--Ethan
Ethan Winer
> if the point of the test is to compare everyones ears, then
identical playback gear is the correct way to go <
Who said the test was to compare ears? I would love to have a
pair of $25,000 speakers, but I don't have them, none of my
friends have them, and you probably don't either. So what
speakers do we standardize on? Mackies? Events? JBL? Then
people will complain that while the speakers are now all the
same, the rooms are different. And the power amps are
different. And the speaker wires are different.
I think my test, if nothing else, shows that most people who are
"into" audio and have probably at least a decent playback
system, can't tell much degradation even when the quality is
reduced to 13 bits. Did you take the test and send me your
results?
--Ethan
George Perfect
were trying to measure.
But I'm not at all sure you are so clear yourself.
Proving that an FM radio station, rebroadcast and heard over AM doesn't let
you hear the cymbals says nothing about the quality of FM broadcasting.
Similarly, when seeking to determine the lower levels of audible resolution,
you won't get any meaningful results if the playback systems are themselves
incapable of reproducing the detail you seek to measure.
If you had made it clear that you wanted to establish that most people's
playback systems or environments don't allow people to hear high definition
audio I wouldn't have bothered commenting in the first place.
From the rest of your message, it seems you have a personal axe to grind in
this area. But you'll search long and hard to find evidence of me telling
people to spend money on high priced anything unless they derive value from
it. As it happens, I do have a high quality music reproduction system. I
bought it only after listening to several alternatives for each component in
situ at length and chose those that I considered best. I would advise anyone
else to do the same.
Value for money is not something where you, I or anyone else can make
pronouncements. One man's pleasure is another man's waste of funds.
> I think my test, if nothing else, shows that most people who are
> "into" audio and have probably at least a decent playback
> system, can't tell much degradation even when the quality is
> reduced to 13 bits. Did you take the test and send me your
> results?
The words "most", "probably", "much" don't speak to me of science as I
understand it. The science of marketing men and popular jourmalism, perhaps.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Ethan Winer" <nos...@nospam.com> wrote in message
news:cdIm8.348934$pN4.23...@bin8.nnrp.aus1.giganews.com...
Arny Krueger
> Nick,
>
> > if the point of the test is to compare everyone's ears, then
> identical playback gear is the correct way to go <
> Who said the test was to compare ears?
Right. People need to look at situations like this somewhat in the
abstract.
> I would love to have a
> pair of $25,000 speakers, but I don't have them, none of my
> friends have them, and you probably don't either. So what
> speakers do we standardize on? Mackies? Events? JBL?
Mackie HR 824s are good, JBL LSR28Ps are good. Never heard Events.
> Then people will complain that while the speakers are now all the
> same, the rooms are different.
Which of course they all are. And since room noise levels are a major
factor in evaluations like this, its a big uncontrolled issue.
>And the power amps are different. And the speaker wires are
different.
Those we don't have to worry about so much.
> I think my test, if nothing else, shows that most people who are
> "into" audio and have probably at least a decent playback
> system, can't tell much degradation even when the quality is
> reduced to 13 bits.
13 bits is a very early standard, that was proposed I believe by the
BBC for distributing classical music.
The CD standard was originally 14 bits. They gratuitously added 2
more bits, mostly because they could do so quite easily.
Arny Krueger
> Ethan - calm down, please. I already apologised if I misunderstood
what you
> were trying to measure.
>
> But I'm not at all sure you are so clear yourself.
>
> Proving that an FM radio station, rebroadcast and heard over AM
doesn't let
> you hear the cymbals says nothing about the quality of FM
broadcasting.
George, if you are saying that people did this test with monitoring
setups that were of AM-radio quality, I think that's unlikely.
Otherwise, you've just used a grossly misleading example.
> Similarly, when seeking to determine the lower levels of audible
resolution,
> you won't get any meaningful results if the playback systems are
themselves
> incapable of reproducing the detail you seek to measure.
That's such a tiny part of the whole picture that it's not even a
half-truth.
The fact of the matter is that when I am seeking to determine
audibility of the lower levels of distortion, then I know that the
results will be limited by the weakest link in the chain, whatever it
is. So I proceed in an engineering-like fashion tabulating what I
know about every portion of the system, from recording room to
playback room and including the listener's ears. I find that the
major practical limits to resolution for a 16/44 system are usually
the rooms.
So what to do? I work on my recording techniques and I come up with
some of the widest dynamic range recordings that are generally
available to the public, and post them at
http://www.pcabx-pro.com/technical/sample_rates/index.htm for people
to download and listen to, to determine the practical limits of
resolution that they can observe.
> If you had made it clear that you wanted to establish that most
people's
> playback systems or environments don't allow people to hear high
definition
> audio I wouldn't have bothered commenting in the first place.
The fact of the matter is that with just Ethan's.wav files we lack su
fficient knowledge to determine what the problem is.
> From the rest of your message, it seems you have a personal axe to
grind in
> this area. But you'll search long and hard to find evidence of me
telling
> people to spend money on high priced anything unless they derive
value from
> it. As it happens, I do have a high quality music reproduction
system. I
> bought it only after listening to several alternatives for each
component in
> situ at length and chose those that I considered best. I would
advise anyone
> else to do the same.
I've done the same thing. But the limit to its resolution is still
the residual noise levels of the room(s).
> Value for money is not something where you, I or anyone else can
make
> pronouncements. One man's pleasure is another man's waste of funds.
But, there are ways to determine the observable resolution of most
people's music systems. It's generally very difficult to determine if
the resolution of the system is limited by the hardware or the
listener unless you get back results that are at similar levels to
what the listener can reasonably be expected to do.
> > I think my test, if nothing else, shows that most people who are
> > "into" audio and have probably at least a decent playback
> > system, can't tell much degradation even when the quality is
> > reduced to 13 bits. Did you take the test and send me your
> > results?
> The words "most", "probably", "much" don't speak to me of science
as I
> understand it. The science of marketing men and popular journalism,
perhaps.
Notice the hypercritical, evasive answer.
Suso Ramallo
> I've done the same thing. But the limit to its resolution is still
> the residual noise levels of the room(s).
I complety agree, mask the low level info.
Typical room noise in a citty is 40 dBA SPL, you need amplify the
quantization noise above 40 dBA SPL for listen this noise, the top of the 96
dB of dimanic (dinamic that offers 16 bit), you need over 130 dB SPL (A very
strong speakers) for hear all the dinamic. This is a unreal escenario.
Bye
Ethan Winer
> Mackie HR 824s are good, JBL LSR28Ps are good. Never heard
Events. <
I don't like the Events (at least the particular model I heard)
as much as the Mackie HR824s because the high end sounded hyped.
I love the Mackies which all my friends and clients have, but my
own home studio has a pair of JBL 4430s (15-inch + horn) biamped
by a pair of Crown PowerBase amps.
--Ethan
Ethan Winer
> calm down, please.
I am very calm. :->) Unlike others here and in other groups I
frequent, I have no personal baggage attached to any of this.
My point is to share and learn, and I (try to) remain open to
new info.
> Proving that an FM radio station, rebroadcast and heard over
AM doesn't let you hear the cymbals says nothing about the
quality of FM broadcasting. <
Yes, but most project studios have at least decent speakers, if
not very good ones. You can't really compare modern near fields
to AM radio! Though I do get your point even if I think the
example is a bit extreme. So to anybody that took my test and
listened on Radio Shack or worse speakers: you need to get
better speakers and try it again.
> Similarly, when seeking to determine the lower levels of
audible resolution, you won't get any meaningful results if the
playback systems are themselves incapable of reproducing the
detail you seek to measure. <
Yes, yes, and once again yes. But as I have said repeatedly, if
you require a super fancy system to hear a tiny difference, then
the difference is not as important as so many salespeople want
you to believe. Using 16 bits instead of 24 is NOT the reason
most amateur mixes sound crappy, and if I have an axe to grind,
that is the main point I want to make.
> As it happens, I do have a high quality music reproduction
system. <
Me too.
> Value for money is not something where you, I or anyone else
can make pronouncements. One man's pleasure is another man's
waste of funds. <
Yes, but Consumer Reports is a very popular magazine because a
lot of people do care about this issue.
--Ethan
Arny Krueger
FWIW, I have NHT Pro A10s in the studio.
Charlie Escher
><snip>
>Here's the real issue: If only a very small percentage of people
>can hear the improvement, and they need to raise the level by 40
>dB. as a reverb tail is dying out, and they have to use a
>particular brand of loudspeaker or whatever, then how important
>is it really? I mix music for people to listen to on whatever
>system they have.
>
>My real objection to the claims for various high-end "tweaks"
>like 24 bits, 96 KHz., expensive cables (sorry,
>"interconnects"), low-jitter outboard (and of course expensive)
>word clock generators, are:
>
>1. This is a consumerist issue, where a lot of newbies and
>not-so-newbies who don't know any better are sold a bill of
>goods by audio dealers who tell them, "The reason your mixes
>suck is because your Mackie pres are so lame. Son, spend a
>couple of grand on this here fine product and you'll hear an
>enormous improvement." Where enormous is the keyword. As if the
>n
>ewbie's problem is inadequate mike pres or cables or word clock
>(a personal fave), as opposed to the REAL problem which is lack
>of experience and (probably) a terrible sounding room.
>
So do you really think you can change the situation? I used to get asked
to do sales in many of the stores I did repairs for, but I was generally
too honest for my managers' tastes. The one thing that sales taught me
is that you sell warm, fuzzy feelings, not anything else. And that's OK,
IMHO. As I said before in this thread, if it feels good, it IS good. In
another post you mentioned Consumer Reports. Ever see their take on a
Porsche, for instance? They miss the whole point of the thing, perhaps
saying that a Monte Carlo has a softer ride, or some such irrelvant
observation...doh!
>2. This nonsense gets passed around faster than a rumor at an
>old-age home. It amazes me that MOST people spout these "facts"
>as if they were actually true. I follow a lot of groups every
>day, and I have seen many people make claims that they obviously
>are just parroting.
>
This is true, and is an unavoidable artifact of USENET. You can readily
turn it to your advantage though, you know {VBG}.
>3. Just as galling as the dealers who profit from
>misunderstanding, the biggest spreaders of fiction are the audio
>magazines. When I wrote my "Audio Myths" article for Audio
>Media magazine a few years ago, I pitched it to Electronic
>Musician first. Scott Wilkinson at EM told me the editors
>looked at my ten points, and "disagreed" with every one of them!
>How do you disagree with scientific fact?!
>
People vote with their wallets. I don't read any audio magazines any
more, but I wish you great luck should you decide to start your own
eventually.
>4. I see a DIRECT relationship between a belief in audio "magic"
>and a belief in astrology, tarot cards, homeopathy, and an
>endless list of other nonsense. The is perhaps the most
>important issue of all: Most people don't know anything about
>scientific method, and they don't want to learn. "My mind is
>made up, don't confuse me with the facts." Or, in this case,
>with what is practical.
>
Homeopathy has a long, documented history of working for lots of people,
including myself. If it's just placebo effect (plenty of research says
otherwise though), fine. I don't have any qualms about paying for
something that makes me feel better. The history of "scientific" Western
medicine is rife with problems in any case, as I'm sure you would agree.
Sorry, but I've seen too many hard-to-explain bits of serendipity create
musical "magic" to put a lot of faith in anyone else's definition of
what is practical. But then, people pay me good money to work my Voodoo
in the audio domain, so I guess my take is suspect at best, eh?
Ever see a sound guy pretend to tweak a few fader settings, and the band
or club owner respond positively? Is the sound guy a charlatan, an
artist, a pragmatist, or what ? To me, that's show business, and a
beautiful thing in many ways. If the band actually starts playing
better, was it worth paying for that wave of the hand?
>Do 24 bits give better resolution than 16? Absolutely. Is it
>possible to design an external word clock box that has lower
>jitter than the stock crystal oscillator in a decent pro audio
>sound card. Of course, given the budget. Do some speaker cables
>have less capacitance than others? Sure. But do any of these
>things matter? More to the point, are they worth paying extra
>for?
>
To lots of people, yes on all counts, I'd say. FWIW, you haven't yet
convinced me that you aren't one of those folks who are trying to
justify their choice of economy gear as being "just as good as XXX
expensive widget". We all see and hear through different filters, which
is fine. So, what's it gonna do for you if you can convince me that I
misspent my bucks on the gear I prefer, for whatever reason?
BTW, I commend you for your well reasoned, civil tone. I used to walk in
your shoes, more or less. I still work a day job that requires rigorous
scientific methodology, and I'd like to think that I understand its
place and value. I'm sure that I'm speaking for a number of hobby
recordists here who just have other priorities than spending hours on
double blind tests and the like. I hope you get some measure of
satisfaction in advancing your agenda, whatever it may be.
sincerely,
Charlie Escher
John Atkinson
> Mr. Atkinson, you seem to be very much in the need of a few answers,
> answers that you think I have. Your story has touched my heart, and
> I've decided to extend you some undeserved courtesy.
the incorrect statements you made about his behavior. Then
priovide the answers to the questions I asked you throughout
December.
> I'll answer 5 questions of yours, subject to the following
> conditions:
threw out unsupported and presumably untrue comments
about things other people were supposed to have said or
done. "You made the claim, you provide the proof," as you
demand of others.
John Atkinson
Editor, Stereophile
Arny Krueger
"John Atkinson" <stereophi...@compuserve.com> wrote in message
news:9b0b3a02.02032...@posting.google.com...
You've got 4 questions left, Mr. Atkinson having forfeited your first
one on the grounds that it requires research to answer.
George Perfect
>
>
> You've got 4 questions left, Mr. Atkinson having forfeited your first
> one on the grounds that it requires research to answer.
>
You agree you owe me some answers then.
Ethan Winer
> So do you really think you can change the situation?
I'm old and wise enough to know I can't change it completely,
but I can at least try to help a little.
> you mentioned Consumer Reports. Ever see their take on a
Porsche, for instance? They miss the whole point of the thing <
My point wasn't that CR is a perfect magazine, but that a lot of
people care about consumerist issues.
> Homeopathy has a long, documented history of working for lots
of people, including myself. If it's just placebo effect (plenty
of research says otherwise though), fine. <
I can't let that pass. There is no legitimate reseach that says
Homeopathy is anything more than placebo. If you know of any
legtimiate research that says otherwise, please tell me where I
can find it.
> So, what's it gonna do for you if you can convince me that I
misspent my bucks on the gear I prefer, for whatever reason? <
I have no interest in convincing you that you wasted money. The
people I'm trying to reach are the folks who are unhappy with
their recordings, and believe wrongly that they need to spend a
lot of money - which they may not have - on overpriced boutique
preamps, A/D convertors, or cables etc.
> I still work a day job that requires rigorous scientific
methodology, and I'd like to think that I understand its place
and value. <
Then how can you possibly believe in homeopathy?! It is a
proven fraud, with absolutely no scientific basis.
--Ethan
Ethan Winer
> FWIW, I have NHT Pro A10s in the studio.
I've never heard their speakers, but aren't they the outfit that
sells an EQ with a 40 KHz. band? :->)
--Ethan
Arny Krueger
Not that I know of!
Here's their web site with product line:
http://www.nhtpro.com/prohome.asp
Arny Krueger
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:Jm0n8.11346$hF3.2042802235@newssvr16.news.prodigy.com...
> >
> >
> > You've got 4 questions left, Mr. Atkinson having forfeited your
first
> > one on the grounds that it requires research to answer.
> You agree you owe me some answers then.
No, I'm just trying to help him out.
Notice that I asked him a number of questions yesterday that he's not
exactly falling over himself to answer. It's always this way with
him. Lately, he generally doesn't respond to about 2/3 of my replies.
He just deletes them and pretends they never happened.
Hogge
> sells an EQ with a 40 KHz. band? :->)
>
> --Ethan
That's Nightpro with their "air band" eq.
/Hogge
Jny Vee
<vanb...@dedserius.com> wrote:
> I've been talking to some guys who use ProStudio on MAC G4 for their DAWs and
> it's clean and rich sound. They insist I should run out and buy a similar
> setup, but I'm not so sure.
'clean rich sound' starts and ends primarily (I'd go so far as to say
exclusively) with the performer and instrument and arrangement. It has
very little to do with any non-broken DAW that works at 44/16 or
better. The only digital system with any sort of track record of 'rich'
sonics is PARIS and that's it's own world of trouble.
> My immediate goal is to re-master 6 tracks from Wasted Tears to a quality
> decent
> enough to sell at gigs as a CD.
> I've begun mastering one of them and here's
> what I've done so far:
what you describe below has nothing to do with 'mastering' but with
completely reworking the production.
> 1. MIDI'ed the drums and bass on the Korg;
(snipped)
--
< Help Keep The Net Emoticon Free! >
kcc
.
> George,
>
> > Unless you know they were all using identical (very
good)
> monitoring systems <
>
people
> who are at least audio amateurs, if not professionals,
have a
> hard time distinguishing between 24 bits and 13 bits, I
think
for
> anyway? Mostly people whose playback systems are a LOT
worse
> than what the folks in these newsgroups and forums use.
>
>
Speaking as one of the people who only identified the 9 and
11 bit versions correctly, I could hear a difference in the
other files, but couldn't relate it to a level of quality,
and I'm still trying to make to a connection to the
subjective aspects of quality. (Or maybe I'm just trying to
make myself feel better about missing 3 of 5.) Anyway, the
truncated file sounded brighter and richer in harmonic
content. Coming from a cassette background I associated
this with better quality and assumed it was the dithered
file.
In hind sight, I'm wondering if the act of truncating could
have a bias toward creating even or odd harmonics or
something that could make it sound "better". I assume we
all EQ our mixes to make them sound "better" despite the
fact that the action reduces bit depth in the same way as
changing volume. Therefore maybe "quality" of sound and bit
depth should not be linked in an absolute way.
So what does "quality" mean? No one heard the guitar live
so there's no telling how accurately any of the samples are
at reproducing the original. If the test is to see if I can
hear 3 bits of sampling noise over the background noise (in
my room or the recording room), then I failed on 3 out of 5
since the differences I heard were not noise. I would say
the other effects are subjective and it's probably not
appropriate to use terms like quality or better.
Similarly, the "quality" of the system used to listen to the
test probably isn't as important as the noise that it adds
to the sound. A crappy system that is quiet, but happens to
be biased toward bringing out the sampling error would make
passing the test easier than a quality system. I agree that
to just test people's ears it should be the same system,
crappy or quality. But as others have pointed out, it
should be the same room, from the same position, at the same
temperature, same humidity, etc, etc.
At the end of the day the truncated file sounded best to me,
but I'm still going to continue to record at 24 bits because
it will give me more options in how I can screw it up in the
mix down.
P.S. What is the theoretical difference in bit depth between
the dithered file and the truncated file. If the truncated
file is truely truncate rather than rounded, it should be
only 15 bits or maybe 15.5. The dithered file is only 16
bits no matter how it got to be that way, but I guess since
dithering affects a series of samples rather than just the
one, it could have a dynamic range of more than 16 bits.
After all, a 2 bit system can produce a square wave with
infinite dynamic range if the frequency evenly divides into
the sampling rate. Am I correct in assuming the dithered
file falls well short of 24 bits? 17?
Ken
Arny Krueger
"kcc" <kcco...@attbi.com.NOSPAM> wrote in message
news:JMan8.102603$702.21863@sccrnsc02...
> Speaking as one of the people who only identified the 9 and
> 11 bit versions correctly, I could hear a difference in the
> other files, but couldn't relate it to a level of quality,
> and I'm still trying to make to a connection to the
> subjective aspects of quality. (Or maybe I'm just trying to
> make myself feel better about missing 3 of 5.) Anyway, the
> truncated file sounded brighter and richer in harmonic
> content. Coming from a cassette background I associated
> this with better quality and assumed it was the dithered
> file.
A natural mistake.
> In hind sight, I'm wondering if the act of truncating could
> have a bias toward creating even or odd harmonics or
> something that could make it sound "better".
Truncation causes a rich mixture of spurious responses including both
harmonic and intermodulation distortion. Some of the distortion
products are aharmonic.
>I assume we
> all EQ our mixes to make them sound "better" despite the
> fact that the action reduces bit depth in the same way as
> changing volume.
If the recording is properly dithered, the reduction of bit depth
manifests itself only as decreased signal-to-noise ratio.
>Therefore maybe "quality" of sound and bit
> depth should not be linked in an absolute way.
Well, as a decrease in signal-to-noise ratio, similar to what happens
with analog.
> So what does "quality" mean?
That is a very deep and subjective question!
> No one heard the guitar live
> so there's no telling how accurately any of the samples are
> at reproducing the original.
Most serious blind testing is done in such a way that there is an
absolute reference of sorts. For example, if you make a 24 bit
recording, and want to know whether cutting it down to 16 bits
damages the sound quality, you use the 24 bit recording as your
reference, and compare the 16 bit recording to it, to see if you can
hear a difference.
You can find a large number of listening tests that are set up this
way at http://www.cdabx.com/technical/bits44/index.htm and
http://www.pcabx-pro.com/technical/sample_rates/index.htm .
> If the test is to see if I can
> hear 3 bits of sampling noise over the background noise (in
> my room or the recording room), then I failed on 3 out of 5
> since the differences I heard were not noise.
I don't know if it is relevant whether or not the most important
thing is whether we hear sonic degradation as being a certain thing,
or whether we just hear it as a difference.
> I would say
> the other effects are subjective and it's probably not
> appropriate to use terms like quality or better.
That is how differences sound when they are very small. We still hear
them, but it is difficult or impossible to properly describe them.
> Similarly, the "quality" of the system used to listen to the
> test probably isn't as important as the noise that it adds
> to the sound.
In many situations, this is the case.
>A crappy system that is quiet, but happens to
> be biased toward bringing out the sampling error would make
> passing the test easier than a quality system.
Experience shows that sonically flawed systems can be better than
good systems for hearing certain specific flaws, but that the good
system does a better job across the board - when many different kinds
of failures are present (one or more at a time).
> I agree that
> to just test people's ears it should be the same system,
> crappy or quality.
Ideally that would be the case. However, if you want to test ears,
you really just need a system that is better than the ears. There can
be several systems that are different, but sufficiently better than
the ears. The results obtained with any of them should suffice.
> But as others have pointed out, it
> should be the same room, from the same position, at the same
> temperature, same humidity, etc, etc.
That would be ideal. However, in the real world, it never happens. If
you study statistical acoustics, you find that the room is constantly
changing.
> At the end of the day the truncated file sounded best to me,
> but I'm still going to continue to record at 24 bits because
> it will give me more options in how I can screw it up in the
> mix down.
I follow that policy, too.
> P.S. What is the theoretical difference in bit depth between
> the dithered file and the truncated file.
There doesn't have to be any difference in bit depth at all. You can
have a file that is 11 bit dithered and 11 bit undithered, and the
bit depth is obviously the same. However, the two files will sound
different.
> If the truncated file is truly truncate rather than rounded, it
should be
> only 15 bits or maybe 15.5.
It would be 15 bits.
> The dithered file is only 16
> bits no matter how it got to be that way, but I guess since
> dithering affects a series of samples rather than just the
> one, it could have a dynamic range of more than 16 bits.
A 16 bit file can never have more resolution or dynamic range than 16
bits. However, signals smaller than one bit high can be reliably
coded with a series of 16 bit numbers if dither is used.
> After all, a 2 bit system can produce a square wave with
> infinite dynamic range if the frequency evenly divides into
> the sampling rate.
No, because dynamic range is not based on comparing zero to the
largest value that can be described. Dynamic range is based on the
ratio between the largest value that can be described and the
smallest non-zero value that can be described.
> Am I correct in assuming the dithered
> file falls well short of 24 bits? 17?
A 16 bit dithered file still only have the dynamic range and
resolution of 16 bits, dither or not.
George Perfect
"Arny Krueger" wrote
>
> If the recording is properly dithered, the reduction of bit depth
> manifests itself only as decreased signal-to-noise ratio.
>
Here's one we can all test for ourselves with ease.
Take a 16 or 24 bit recording that has some fine detail in it.
Now without changing the sampling rate convert it to 8 bits (4 bits if you
can - Arny sets no limits in his pronouncement) - don't forget the dither
now. I'll leave the choice of dither type to you - try all you have
available if you want.
Listen to the original and listen to the reduced bit depth version. If Arny
is right all you will hear is a little more noise. I suggest you'll hear
something different. Try it.
You can also test Arny's statement this way. If you have a reasonable
graphics card in your computer (one that handles 16 bit or 24 bit colour
depth) load a 24 bit encoded ("milions of colours") photograph which
exhibits some finally graded colour variations - a picture containing sky
and clouds will often do. Now without changing the image size (ie, sample
rate) convert the picture to 16 colour depth (4 bits).
Compare the original and 8 bit version of the pictures side by side. See
those bands in the graduated areas that weren't in the original? They're not
caused by noise but by lack of resolution.
This process is IDENTICAL to what happens when you reduce the bit rate of an
audio signal. Dither has nothing to do with it. In fact, if you use
Photoshop or Paintshop Pro you can choose different dither algorithms as you
perform the bit reduction. Feel free to try them - if you succeed in finding
one that matches the specific image you used you may find that the edges
between the bands become blurred. But ... if your monitor (or printer) is
good enough that picture is never going to look like the original with just
a little noise added.
Zoom in on the image and you will note that each pixel can only take one of
256 colours. If you used dither (or the program added dither automatically)
you will see that adjacent pixels that take the same colour or were maybe
only one or two colour steps removed in the original now take some random
value - this is the noise that has been added.
Zoom back out and compare several areas of the original image with the same
area in the bit-reduced copy and you will see a different colour.
Occasionally the difference will be pronounced.
I leave it to you to imagine what the process does to an audio signal.
If there was anything in Arny's 'free lunch' theory we'd all be listening to
4 bit encoded CDs.
nickspicks.com
> Who said the test was to compare ears?
it was a listening test, no?
the point of the test, as you state is to prove that most people can't
*hear* the difference.
I don't think that you can tell me that playback gear is inconsequential in
a trial like this.
Its just not a great test w/everyone listening on different gear. I'm not
saying we should all standardize on gear for this test (though that would
satisfy my argeument), I'm saying that the test isnt' valad due to this
factor...which is HUGE in a listening test.
The way to do this would be to have one listening room, and many people.
I'm not saying it's feasable, I'm just saying that this would produce more
accurate results. With this, you would need gear capable of showing off the
virtues of 24bit audio and yes, that is better than average gear. You can't
fault the source if the playback gear is not up to the task.
Understand, I'm not for or against either side of this.
I would love to have a
> pair of $25,000 speakers, but I don't have them, none of my
> friends have them, and you probably don't either. So what
> speakers do we standardize on? Mackies? Events? JBL? Then
> people will complain that while the speakers are now all the
> same, the rooms are different. And the power amps are
> different. And the speaker wires are different.
my point exactly.
You need 100% controll over all of these things in order to successfuly show
valid results.
>
> I think my test, if nothing else, shows that most people who are
> "into" audio and have probably at least a decent playback
> system, can't tell much degradation even when the quality is
> reduced to 13 bits.
it's an assumption that most people who took this test were listening with a
decent playback system.
Did you take the test and send me your
> results?
nope.
Chris G.
wars?
Personally I think both Arny and George can be nice guys most of the time,
but that they both can get a little full of themselves at times as
well...that's probably why they argue so much.... Too much alike those two
yes? Hmmm? (My Yoda imitation. LOL! ) :)
My opinion was that the original test in this thread was interesting. Not
perfect scientific proof of anything due to the obvious flaws in the
experimental design, but nonetheless still useful "real world" information.
Hell I still love a lot of 8 bit recorded music done on the old Commodore
Amiga computers. My brother did some music using some mod-trackers and a
state of the art (at the time) 8 bit sampler for the Amiga and they actually
sound much warmer and more pleasant then his later 16 bit recordings.
Science be damned, it sounded good.
Anyhow.....I don't care much for bits and dither and all that... I believe
in simply doing what my ears tell me sounds good even if it goes against
scientific audio facts. Science is not absolute truth. It's an attempt at
getting closer to truth using rational scientific methodology by controlling
for variables as much as possible.
But there are almost always the possibilities of external variables
affecting the design so to take any science dealing with audio as
absolutelely unshakable biblical truth is silly I think. If science tells
me one thing, but my ears tell me something very different, I go with my
ears.
However with that said, science is good. It's just that people get so
wrapped up in their data and their results that they fail to see the forest
through the trees and just get their egoes all wrapped up around their
precious experiments. That's one reason why I decided not to study
experimental psychology (where I was actually thinking of studying audio
perception) and went into cultural anthropology instead(the red-headed
stepchild of the social sciences). It's soooo much more open ended and
enlightened overall.
Anyhoo....my advice to Arny and George....just let it go........let it
go...run free!!!
Chris G.
aka-Miles Teg<GD>
--
"Getting eaten by a crocodile is JUST like falling asleep... in a blender!"
- Homer Simpson
"Arjan P" <ne...@bootlegs.at.demon.nl> wrote in message
news:3C9A6DAC...@bootlegs.at.demon.nl...
> Arny Krueger wrote:
> >
> > Just a red herring from someone who really doesn't understand the
> > issues that underlie tests like this, and based on his recent
> > behavior, seemingly doesn't want to learn.
>
> I knew it right from the header: Here we go again... Another thread to
kill
> right away! Mr. Krueger, I'll ask it once again: please leave this
formerly
> almost flameless group, and let it get back to that state...
>
> Luck, Arjan
>
> -----To reply via email remove "at." from address.-----
> ---[This is a spam evasion manouvre, not to bug ya.]---
> Where I come from: http://www.bootlegs.demon.nl/studio/
Chris G.
mind. Haven't you ever seen Bill Moyers interview with Joseph Cambell on
"The Power of Myth"? Great stuff.
Myth is good. Ok..I'm going back to my tarrot cards now....the cards show
that I'll soon be flamed. LOL!
Chris G.
--
"Getting eaten by a crocodile is JUST like falling asleep... in a blender!"
- Homer Simpson
"Ethan Winer" <nos...@nospam.com> wrote in message
news:cdIm8.348934$pN4.23...@bin8.nnrp.aus1.giganews.com...
Chris G.
Chris G.
--
"Getting eaten by a crocodile is JUST like falling asleep... in a blender!"
- Homer Simpson
"Suso Ramallo" <jram...@worldonline.es> wrote in message
news:a7fvu8$kcv$1...@talia.mad.ttd.net...
Chris G.
and just totally ignore what I'm telling them about my own damn body even
when they don't have a clue what's going on with me (or pretend that they do
regardless). Those types piss me off to no end especially when I'm paying
good money to them.
Having lots of knowledge does not = the ability to use that knowledge in a
meaningful way. I also have had wonderful success with a variety of
herbal remedies for various ailments. I swear by my Echinacea for example.
Science be damned it shortens or sometimes halts colds and flu in its tracks
when I take it.
At any rate, to me music IS magic. Making music is a profoundly spiritual
experience to me and I think many audio engineers lose track of that. Part
of creating a good mix is becoming enveloped in that "magic" in order to
make changes to the audio that will heighten that magic even if it means
making the recording a bit duller or more distorted. Recording is an art
and by its very nature is highly subjective. Science is only there to help
us out in making subjective judgements on audio and to help us improve our
"art" by giving us the best tools possible for interpreting what we hear in
our heads if that means degrading the audio.
I haven't heard of any audio engineers who mix using scientific methodology
and who don't regard their own ears as the ultimate measurment device.
So don't sweat the petty stuff and don't pet the sweaty stuff.
ROCK ON!!! (sorry I just felt like saying that).
Chris G.
--
"Getting eaten by a crocodile is JUST like falling asleep... in a blender!"
- Homer Simpson
"Charlie Escher" <charl...@gorge.net> wrote in message
news:3C9C18BB...@gorge.net...
> Ethan Winer wrote:
snip---
> > >4. I see a DIRECT relationship between a belief in audio "magic"
> >and a belief in astrology, tarot cards, homeopathy, and an
> >endless list of other nonsense. The is perhaps the most
> >important issue of all: Most people don't know anything about
> >scientific method, and they don't want to learn. "My mind is
> >made up, don't confuse me with the facts." Or, in this case,
> >with what is practical.
> >
> Homeopathy has a long, documented history of working for lots of people,
> including myself. If it's just placebo effect (plenty of research says
> otherwise though), fine. I don't have any qualms about paying for
> something that makes me feel better. The history of "scientific" Western
> medicine is rife with problems in any case, as I'm sure you would agree.
>
> Sorry, but I've seen too many hard-to-explain bits of serendipity create
> musical "magic" to put a lot of faith in anyone else's definition of
> what is practical. But then, people pay me good money to work my Voodoo
> in the audio domain, so I guess my take is suspect at best, eh?
>
> Ever see a sound guy pretend to tweak a few fader settings, and the band
> or club owner respond positively? Is the sound guy a charlatan, an
> artist, a pragmatist, or what ? To me, that's show business, and a
> beautiful thing in many ways. If the band actually starts playing
> better, was it worth paying for that wave of the hand?
-snip
> sincerely,
>
> Charlie Escher
>
>
Ethan Winer
> In hind sight, I'm wondering if the act of truncating could
have a bias toward creating even or odd harmonics or something
that could make it sound "better". <
Maybe, but only if we can hear subtle changes happening 80-90
dB. below the program in the presense of the program. However,
you make an excellent point: That which sounds better is not
necessarily that which is more accurate. I explained that in
the Audio Myths article I wrote for Audio Media Magazine (on the
Articles page at my web site). Here's a paragraph from the
Conclusion:
"When all else is equal, people will generally pick the brighter
(or just louder) version as sounding better, unless of course
the sound already was too loud or bright. People will sometimes
report a difference even in an "A/A" test, where nothing at all
has changed! And just because something sounds "better," it is
not necessarily higher fidelity. Goosing the treble and bass or
adding a little compression often makes a track sound better,
but that doesn't mean the result is more faithful to the
original source material."
> A crappy system that is quiet, but happens to be biased toward
bringing out the sampling error <
I'm not sure what properties those would be, that "bring out"
such errors.
> P.S. What is the theoretical difference in bit depth between
the dithered file and the truncated file. If the truncated file
is truely truncate rather than rounded, it should be only 15
bits or maybe 15.5. The dithered file is only 16 bits no matter
how it got to be that way <
There's no such thing as 15.5 bits - either the 16th bit was
used or it wasn't. You are correct that all the files are 16
bits. But since that's how audio is distributed, and the
proponents of 24-bit recording say 24 is better anyway because
the improvement shows up when you dither, it's not unreasonable
to distribute the test files that way too.
--Ethan
Ethan Winer
> That's Nightpro with their "air band" eq.
Thanks for the clarification. I wrongly assumed they were the
same outfit.
--Ethan
Ethan Winer
> I don't think that you can tell me that playback gear is
inconsequential in a trial like this. <
Of course not, but if you need very expensive speakers to hear
the difference, then it's probably not as important as so many
believe. I am not saying that 24-bit recording is never
worthwhile - heck, I'm trying hard not to take any position
except "show me." But I do believe that 24-bit recording is not
necessary to get good results. And not using 24 bits is
certainly not why mixes suck. If someone is on a budget I would
much rather see them buy a decent pair of speakers and a modest
sound card, than a cheaper pair of speakers and a 24-bit sound
card.
--Ethan
nickspicks.com
I dont think you need the "best money can buy" playback gear to hear the
difference, but you certainly need some low-end pro gear or high end
consumer stuff.
say, a lucid 2496 D/A and a pair of good monitors w/sub. I'm sure that
would do it.
it's a subjectvie test on my end, but I believe I can hear the difference
between 24 and 16bit on my very modest playback gear consisting of either an
Edirol UA-5 or Audiophile 2496 into a10 year old sony amp and some B&W 30x
two way towers.
that's just me though...
> Nick,
>
> > I don't think that you can tell me that playback gear is
> inconsequential in a trial like this. <
>
> Of course not, but if you need very expensive speakers to hear
> the difference, then it's probably not as important as so many
> believe. I am not saying that 24-bit recording is never
> worthwhile - heck, I'm trying hard not to take any position
> except "show me."
that's a good stance.
But I do believe that 24-bit recording is not
> necessary to get good results.
I fully agree with this.
And not using 24 bits is
> certainly not why mixes suck. If someone is on a budget I would
> much rather see them buy a decent pair of speakers and a modest
> sound card, than a cheaper pair of speakers and a 24-bit sound
> card.
mixing or editing with a higher world length is certainly advantageous
though, no arguement there from anyone I'd think.
Van
> better. The only digital system with any sort of track record of 'rich'
> sonics is PARIS and that's it's own world of trouble.
>
So, is it not possible to get "rich" sonics from a digital multi-track? DAW
(such as Cakewalk Pro Audio; ProStudio)? Could this (if that's the case) be
remedied by recording digitally, then running the final mix through an analog
4-track prior to mastering.
> what you describe below has nothing to do with 'mastering' but with
> completely reworking the production.
>
> > My immediate goal is to re-master 6 tracks from Wasted Tears to a quality
> > decent
>
Newbie as charged...
Van
--
=================================================================
Linux rocks!!! http://www.dedserius.com/
=================================================================
Arny Krueger
"Van" <vanb...@dedserius.com> wrote in message
news:3C9E0323...@dedserius.com...
> Jny Vee wrote:
> > The only digital system with any sort of track record of 'rich'
> > sonics is PARIS and that's it's own world of trouble.
More digital misapprehensions...
> So, is it not possible to get "rich" sonics from a digital
multi-track?
If anybody seriously believes this...
kcc
.
>
> "kcc" <kcco...@attbi.com.NOSPAM> wrote in message
> news:JMan8.102603$702.21863@sccrnsc02...
>
> > infinite dynamic range if the frequency evenly divides
into
> > the sampling rate.
>
> No, because dynamic range is not based on comparing zero
to the
> largest value that can be described. Dynamic range is
based on the
> ratio between the largest value that can be described and
the
> smallest non-zero value that can be described.
>
infinite S/N ratio? If the wave were created digitally so
that there no room noise or A/D issues, the noise level
could be zero. We, or at least I, tend to think of dynamic
range and S/N as interchangeable. Is my flaw in the
definition of "noise"? I still think of noise in terms of
tape hiss or that stuff they pump into my office that
suppose to be white noise, but misses by a mile. Throwing
out my extreme and unrealistic example, as soon a noise is
above zero, it has to be at least as loud as the floor of
the dynamic range, but can it be higher? Does "noise" have
to be sufficiently random to mask all quieter sounds and
thereby create the floor? E.g. are pops and clicks on an LP
noise? Or back to the example, if a sample reproduces the
frequency characteristic of the source with 100% accuracy,
but at the wrong volume (assuming both are a constant
volume), is that "noise"?
Thanks
Ethan Winer
> it's a subjectvie test on my end, but I believe I can hear the
difference between 24 and 16 bit on my very modest playback gear
<
That's why I asked if you had downloaded the files and sent me
your estimation of which is which. Until you actually test it
for yourself - and it doesn't have to be my test using my
files - you're only guessing.
--Ethan
kcc
> Ken,
> > A crappy system that is quiet, but happens to be biased
toward
> bringing out the sampling error <
>
> I'm not sure what properties those would be, that "bring
out"
> such errors.
The noise in most digital recordings is too subtle for me to
describe the sonic characteristics, even when I can hear
something, so let me make a (perhaps flawed) analogy with
tape. If you were trying detect the difference in hiss
between two cassette tapes, having a higher quality tape
deck, pre-amp, amp and cables should make the test easier.
But lower quality speakers with no low end and a spike at
the frequency of the hiss may be better for the test than
high quality speakers. I guess I don't really know what I
meant by "sampling error". This gets back to my question
about what noise is from my follow-up to Arny.
>
> > P.S. What is the theoretical difference in bit depth
between
> the dithered file and the truncated file. If the
truncated file
> is truely truncate rather than rounded, it should be only
15
> bits or maybe 15.5. The dithered file is only 16 bits no
matter
> how it got to be that way <
>
> There's no such thing as 15.5 bits - either the 16th bit
was
> used or it wasn't. You are correct that all the files are
16
> bits. But since that's how audio is distributed, and the
> proponents of 24-bit recording say 24 is better anyway
because
> the improvement shows up when you dither, it's not
unreasonable
> to distribute the test files that way too.
>
> --Ethan
13 bit test is ultimately measuring a 3 bit change, not an
11 bit change. But now I'm curious if you took the 24 bit
file and did -/+60db (or whatever amount was appropriate) to
get it down to 9 bits, then dithered it from 24 (with padded
zeros) down to 16 bits, how would that sound? I think of
dithering as smart rounding based on the values of
neighboring samples. If that's all it is, only the 16th bit
would change, but I'm guessing it's more. At the extreme,
if a sample were negative when all of the surrounding
samples were positive, would dithering flip the sign (i.e.
the first bit)? If it does affect more than 1 bit, would
dithering from 16 bit to 16 bit be a valid option?
I was thinking of 15.5 bits as a unit of measurement as
opposed to the literal meaning, but even that view may be
wrong. If you did a (theoretical) comparison of
your -/+30db test to a -/+ 31db test, is there a difference
in the dynamic range or does it truly have to round to the
nearest bit?
P.S. Thanks again for the test, and thanks to everyone
that's contributing to the thread. I'm learning a lot, even
when it degenerates to less than 9 bits. Mostly, I learned
I really need a quieter computer.
Ken
Ben Bradley
C'mon, Arny, we're all waiting for a good 'starved plate' Direct-X
plug-in to warm up our .wav tracks ... ;)
-----
http://listen.to/benbradley
Arny Krueger
"kcc" <kcco...@attbi.com.NOSPAM> wrote in message
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:YTin8.11933$A93.213...@newssvr16.news.prodigy.com..
> > "kcc" <kcco...@attbi.com.NOSPAM> wrote in message
> > news:JMan8.102603$702.21863@sccrnsc02...
> > > After all, a 2 bit system can produce a square wave with
> > > infinite dynamic range if the frequency evenly divides into
> > > the sampling rate.
> > No, because dynamic range is not based on comparing zero to the
> > largest value that can be described. Dynamic range is based on
the
> > ratio between the largest value that can be described and the
> > smallest non-zero value that can be described.
> Ok, I'm wrong. But what if I had said it would have an infinite
S/N ratio?
Still no go.
>If the wave were created digitally so
> that there no room noise or A/D issues, the noise level
> could be zero.
No, because there is still quantization noise, even if the wave is
created digitally.
> We, or at least I, tend to think of dynamic
> range and S/N as interchangeable.
They tend to be closely related.
> Is my flaw in the definition of "noise"?
Quantization noise needs to be considered.
>I still think of noise in terms of
> tape hiss or that stuff they pump into my office that
> suppose to be white noise, but misses by a mile.
The noise they pump into your office probably starts out as pink
noise, but of course it gets spectrally shaped by the speakers and
the rooms.
> Throwing
> out my extreme and unrealistic example, as soon a noise is
> above zero, it has to be at least as loud as the floor of
> the dynamic range, but can it be higher?
Depends what one calls "noise".
> Does "noise" have
> to be sufficiently random to mask all quieter sounds and
> thereby create the floor?
The word "noise" is used quite broadly, and can be taken to mean many
things. For sure, a noise does not have to mask all quieter sounds to
be a major contributor to the noise floor.
> E.g. are pops and clicks on an LP noise?
By some definitions of noise, yes.
What is noise?
It can be anything that is not signal. In this case it includes
distortion.
It can be anything that is not correlated with the signal. In this
case it does not include distortion.
>Or back to the example, if a sample reproduces the
> frequency characteristic of the source with 100% accuracy,
> but at the wrong volume (assuming both are a constant
> volume), is that "noise"?
That would be "distortion". But distortion can be considered to be
part of "noise".
There's a reason why I post lots of FFT plots at www.pcavtech.com.
They clarify what is considered "signal", "noise" and "distortion".
nickspicks.com
Arny Krueger
> Among a number of questionable statements ...
Ah here we go again.
;-)
> "Arny Krueger" wrote
> > If the recording is properly dithered, the reduction of bit depth
> > manifests itself only as decreased signal-to-noise ratio.
> Here's one we can all test for ourselves with ease.
> Take a 16 or 24 bit recording that has some fine detail in it.
> Now without changing the sampling rate convert it to 8 bits (4 bits
if you
> can - Arny sets no limits in his pronouncement) - don't forget the
dither
> now. I'll leave the choice of dither type to you - try all you have
> available if you want.
Been there, done that. You end up with one hell of a noisy recording.
There's no way you can reduce good clean audio to 4 or 8 bits and
hold the sample rate the same, and not degrade its sound.
> Listen to the original and listen to the reduced bit depth version.
If Arny
> is right all you will hear is a little more noise. I suggest you'll
hear
> something different. Try it.
It's a ludicrous experiment because 8 bits is very noisy. However, I
provide samples that degrade a 24 bit recording to 16, 15, 14, 13,
12... bits so that people can evaluate this issue for themselves. I
provide samples that use dither and don't use dither. I even provide
a sample that uses dither, but not enough. Here's the place to
listen: http://www.cdabx.com/technical/bits44/index.htm .
> You can also test Arny's statement this way. If you have a
reasonable
> graphics card in your computer (one that handles 16 bit or 24 bit
colour
> depth) load a 24 bit encoded ("millions of colours") photograph
which
> exhibits some finally graded colour variations - a picture
containing sky
> and clouds will often do. Now without changing the image size
(i.e., sample
> rate) convert the picture to 16 colour depth (4 bits).
Yet another ludicrous experiment.
Basically, George is arguing that you can surely taste the difference
caused by adding a tiny grain of salt to a barrel of water because
you can taste the difference caused by adding a pound of salt to a
teacup.
> Compare the original and 8 bit version of the pictures side by
side. See
> those bands in the graduated areas that weren't in the original?
They're not
> caused by noise but by lack of resolution.
That would be solarizing. There is such a thing as reducing color
depth or gray scale resolution with dither, and the result is that
you don't get the banding or solarization effect.
> This process is IDENTICAL to what happens when you reduce the bit
rate of an
> audio signal.
If you reduce the color or gray scale resolution in a similar
fashion, the outcome will be analogous to what happens with audio
signals.
Since George mentioned Paintshop Pro, let's talk about Paintshop Pro.
I have version 4.12 which I use for basic web art. I use Corel Draw
and Photopaint for more creative work. In Paintshop Pro you can
reduce color depth using the Colors, Decrease Color Depth Command bar
option. When you do this you get a dialog box that has an option
called "Error Diffusion". This is roughly like dithered
requantization, and eliminates the color banding that George claims
will always happen. Try converting a high resolution color
photograph both ways!
> Dither has nothing to do with it.
Dither has everything to do with it.
>In fact, if you use
> Photoshop or Paintshop Pro you can choose different dither
algorithms as you
> perform the bit reduction.
Exactly.
>Feel free to try them - if you succeed in finding
> one that matches the specific image you used you may find that the
edges
> between the bands become blurred. But ... if your monitor (or
printer) is
> good enough that picture is never going to look like the original
with just
> a little noise added.
The key phrase here is "a little noise". That's because most of the
color reductions we do are pretty gross. Therefore the consequence is
that we get pictures that look like the origin with LOTS of noise
added.
> Zoom in on the image and you will note that each pixel can only
take one of
> 256 colours. If you used dither (or the program added dither
automatically)
> you will see that adjacent pixels that take the same colour or were
maybe
> only one or two colour steps removed in the original now take some
random
> value - this is the noise that has been added.
Right.
> Zoom back out and compare several areas of the original image with
the same
> area in the bit-reduced copy and you will see a different colour.
> Occasionally the difference will be pronounced.
Right.
> I leave it to you to imagine what the process does to an audio
signal.
Again, its a matter of degree. If you reduce the resolution of a high
quality audio signal to 8 bits, there's no way that you won't hear
the difference. However, just about everybody prefers added noise to
added distortion. In either case you're going to hear it if you can
listen carefully. Pick your poison - the damage will be done! You
just get to pick which kind of damage.
> If there was anything in Arny's 'free lunch' theory we'd all be
listening to
> 4 bit encoded CDs.
My theory is not about "free lunch", it's about managing degradation
that has to happen. And it's not just my theory - it's a well-known
fact.
George has been told many times by many people that he's wrong. I've
cited 3 AES papers for him. At least two web sites besides my own.
But he keeps on like this. If John Atkinson didn't hate me so much, I
suspect that he'd also try to straighten George out.
In a photograph you can reduce the color or gray scale with dither
and get a grainy picture, or you can reduce it without dither and get
a banded picture. Most people seem to prefer the noise. In audio you
can reduce the resolution with dither and pick up some extra noise,
or you can reduce the resolution without dither and pick up some
distortion.
Chris Hornbeck
wrote:
>In a photograph you can reduce the color or gray scale with dither
>and get a grainy picture, or you can reduce it without dither and get
>a banded picture. Most people seem to prefer the noise. In audio you
>can reduce the resolution with dither and pick up some extra noise,
>or you can reduce the resolution without dither and pick up some
>distortion.
This reminds me that there is (are?) Photoshop plug-in(s) that apply
a fractal algorithum for scaling. I wonder if a similar technique
is ever used in audio mixing?
Ferinstance, someone with a bunch of 16/44.1 tracks and access to
a fancy 24/96 mixer. Might be some advantage to bring the originals
into the larger space with some "nonlinearities". Kinda like an
anti-dither; randomness/noise to round the edges. Just a thought.
BTW, solarization has a different meaning in photography.
Thanks for the insurrection,
Chris Hornbeck, Guyville
guyville{at}aristotle{dot}net
Arny Krueger
news:3c9dd089$0$8508$ed9e...@reading.news.pipex.net...
> Among a number of questionable statements ...
Ah here we go again.
> "Arny Krueger" wrote
> > If the recording is properly dithered, the reduction of bit depth
> > manifests itself only as decreased signal-to-noise ratio.
> Here's one we can all test for ourselves with ease.
> Take a 16 or 24 bit recording that has some fine detail in it.
> Now without changing the sampling rate convert it to 8 bits (4 bits
if you
> can - Arny sets no limits in his pronouncement) - don't forget the
dither
> now. I'll leave the choice of dither type to you - try all you have
> available if you want.
Been there, done that. You end up with one hell of a noisy recording.
There's no way you can reduce good clean audio to 4 or 8 bits and
hold the sample rate the same, and not degrade its sound.
> Listen to the original and listen to the reduced bit depth version.
If Arny
> is right all you will hear is a little more noise. I suggest you'll
hear
> something different. Try it.
It's a ludicrous experiment because 8 bits is very noisy. However, I
provide samples that degrade a 24 bit recording to 16, 15, 14, 13,
12... bits so that people can evaluate this issue for themselves. I
provide samples that use dither and don't use dither. I even provide
a sample that uses dither, but not enough. Here's the place to
listen: http://www.cdabx.com/technical/bits44/index.htm .
> You can also test Arny's statement this way. If you have a
reasonable
> graphics card in your computer (one that handles 16 bit or 24 bit
colour
> depth) load a 24 bit encoded ("milions of colours") photograph
which
> exhibits some finally graded colour variations - a picture
containing sky
> and clouds will often do. Now without changing the image size
(i.e., sample
> rate) convert the picture to 16 colour depth (4 bits).
Yet another ludicrous experiment.
Basically, George is arguing that you can surely taste the difference
caused by adding a tiny grain of salt to a barrel of water because
you can taste the difference caused by adding a pound of salt to a
teacup.
> Compare the original and 8 bit version of the pictures side by
side. See
> those bands in the graduated areas that weren't in the original?
They're not
> caused by noise but by lack of resolution.
That would be solarizing. There is such a thing as reducing color
depth or gray scale resolution with dither, and the result is that
you don't get the banding or solarization effect.
> This process is IDENTICAL to what happens when you reduce the bit
rate of an
> audio signal.
If you reduce the color or gray scale resolution in a similar
fashion, the outcome will be analogous to what happens with audio
signals.
Since George mentioned Paintshop Pro, let's talk about Paintshop Pro.
I have version 4.12 which I use for basic web art. I use Corel Draw
and Photopaint for more creative work. In Paintshop Pro you can
reduce color depth using the Colors, Decrease Color Depth Command bar
option. When you do this you get a dialog box that has an option
called "Error Diffusion". This is roughly like dithered
requantization, and eliminates the color banding that George claims
will always happen. Try converting a high resolution color
photograph both ways!
> Dither has nothing to do with it.
Dither has everything to do with it.
>In fact, if you use
> Photoshop or Paintshop Pro you can choose different dither
algorithms as you
> perform the bit reduction.
Exactly.
>Feel free to try them - if you succeed in finding
> one that matches the specific image you used you may find that the
edges
> between the bands become blurred. But ... if your monitor (or
printer) is
> good enough that picture is never going to look like the original
with just
> a little noise added.
The key phrase here is "a little noise". That's because most of the
color reductions we do are pretty gross. Therefore the consequence is
that we get pictures that look like the origin with LOTS of noise
added.
> Zoom in on the image and you will note that each pixel can only
take one of
> 256 colours. If you used dither (or the program added dither
automatically)
> you will see that adjacent pixels that take the same colour or were
maybe
> only one or two colour steps removed in the original now take some
random
> value - this is the noise that has been added.
Right.
> Zoom back out and compare several areas of the original image with
the same
> area in the bit-reduced copy and you will see a different colour.
> Occasionally the difference will be pronounced.
Right.
> I leave it to you to imagine what the process does to an audio
signal.
Again, its a matter of degree. If you reduce the resolution of a high
quality audio signal to 8 bits, there's no way that you won't hear
the difference. However, just about everybody prefers added noise to
added distortion. In either case you're going to hear it if you can
listen carefully. Pick your poison - the damage will be done! You
just get to pick which kind of damage.
> If there was anything in Arny's 'free lunch' theory we'd all be
listening to
> 4 bit encoded CDs.
My theory is not about "free lunch", it's about managing degradation
that has to happen. And it's not just my theory - it's a well-known
fact.
George has been told many times by many people that he's wrong. I've
cited 3 AES papers for him. At least two web sites besides my own.
But he keeps on like this. If John Atkinson didn't hate me so much, I
suspect that he'd also try to straighten George out.
In a photograph you can reduce the color or gray scale with dither
Arny Krueger
news:3c9dd089$0$8508$ed9e...@reading.news.pipex.net...
> Among a number of questionable statements ...
Ah here we go again.
> "Arny Krueger" wrote
> > If the recording is properly dithered, the reduction of bit depth
> > manifests itself only as decreased signal-to-noise ratio.
> Here's one we can all test for ourselves with ease.
> Take a 16 or 24 bit recording that has some fine detail in it.
> Now without changing the sampling rate convert it to 8 bits (4 bits
if you
> can - Arny sets no limits in his pronouncement) - don't forget the
dither
> now. I'll leave the choice of dither type to you - try all you have
> available if you want.
Been there, done that. You end up with one hell of a noisy recording.
There's no way you can reduce good clean audio to 4 or 8 bits and
hold the sample rate the same, and not degrade its sound.
> Listen to the original and listen to the reduced bit depth version.
If Arny
> is right all you will hear is a little more noise. I suggest you'll
hear
> something different. Try it.
It's a ludicrous experiment because 8 bits is very noisy. However, I
provide samples that degrade a 24 bit recording to 16, 15, 14, 13,
12... bits so that people can evaluate this issue for themselves. I
provide samples that use dither and don't use dither. I even provide
a sample that uses dither, but not enough. Here's the place to
listen: http://www.cdabx.com/technical/bits44/index.htm .
> You can also test Arny's statement this way. If you have a
reasonable
> graphics card in your computer (one that handles 16 bit or 24 bit
colour
> depth) load a 24 bit encoded ("millions of colours") photograph
which
> exhibits some finally graded colour variations - a picture
containing sky
> and clouds will often do. Now without changing the image size
(i.e., sample
> rate) convert the picture to 16 colour depth (4 bits).
Yet another ludicrous experiment.
Basically, George is arguing that you can surely taste the difference
caused by adding a tiny grain of salt to a barrel of water because
you can taste the difference caused by adding a pound of salt to a
teacup.
> Compare the original and 8 bit version of the pictures side by
side. See
> those bands in the graduated areas that weren't in the original?
They're not
> caused by noise but by lack of resolution.
That would be solarizing. There is such a thing as reducing color
depth or gray scale resolution with dither, and the result is that
you don't get the banding or solarization effect.
> This process is IDENTICAL to what happens when you reduce the bit
rate of an
> audio signal.
If you reduce the color or gray scale resolution in a similar
fashion, the outcome will be analogous to what happens with audio
signals.
Since George mentioned Paintshop Pro, let's talk about Paintshop Pro.
I have version 4.12 which I use for basic web art. I use Corel Draw
and Photopaint for more creative work. In Paintshop Pro you can
reduce color depth using the Colors, Decrease Color Depth Command bar
option. When you do this you get a dialog box that has an option
called "Error Diffusion". This is roughly like dithered
requantization, and eliminates the color banding that George claims
will always happen. Try converting a high resolution color
photograph both ways!
> Dither has nothing to do with it.
Dither has everything to do with it.
>In fact, if you use
> Photoshop or Paintshop Pro you can choose different dither
algorithms as you
> perform the bit reduction.
Exactly.
>Feel free to try them - if you succeed in finding
> one that matches the specific image you used you may find that the
edges
> between the bands become blurred. But ... if your monitor (or
printer) is
> good enough that picture is never going to look like the original
with just
> a little noise added.
The key phrase here is "a little noise". That's because most of the
color reductions we do are pretty gross. Therefore the consequence is
that we get pictures that look like the origin with LOTS of noise
added.
> Zoom in on the image and you will note that each pixel can only
take one of
> 256 colours. If you used dither (or the program added dither
automatically)
> you will see that adjacent pixels that take the same colour or were
maybe
> only one or two colour steps removed in the original now take some
random
> value - this is the noise that has been added.
Right.
> Zoom back out and compare several areas of the original image with
the same
> area in the bit-reduced copy and you will see a different colour.
> Occasionally the difference will be pronounced.
Right.
> I leave it to you to imagine what the process does to an audio
signal.
Again, its a matter of degree. If you reduce the resolution of a high
quality audio signal to 8 bits, there's no way that you won't hear
the difference. However, just about everybody prefers added noise to
added distortion. In either case you're going to hear it if you can
listen carefully. Pick your poison - the damage will be done! You
just get to pick which kind of damage.
> If there was anything in Arny's 'free lunch' theory we'd all be
listening to
> 4 bit encoded CDs.
My theory is not about "free lunch", it's about managing degradation
Van
> Hi Van - welcome to am4-t!
Thank you, George!
>
> I've sent a short reply to your email to me - I'd rather answer some of your
> questions here in group.
Thx. Sadly, my news server is qwest which lags, but for such
discussions, suffices...
> It looks as if you are re-recording your older songs - strictly speaking
> 'mastering' is a process that comes after all the recording and mixing is
> finished and might apply here only if you were just trying to polish up the
> original recordings.
Newbie error; yes, I'm gutting all but the actual songs, so it's
reproduction. Much learned this week; larger project than I'd
anticipated....
> I listened to Wasted Tears and Violet Blues. They both sound good for songs
> that were recorded on a cassette multi-tracker - you've obviously got the
> basics well in hand.
>
> I like your songs and they deserve the better presentation you are trying to
> give them. Perhaps it would help if I set out how I would approach the task
> (with one eye firmly fixed on keeping the budget under control!)
>
> Let's first look at what I think the songs need:
>
> 1. Better instrumentation - the drum and keyboard sounds are very
> 'machine-like' - better quality sounds and a looser rhythm would both go a
> long way to letting these songs breathe.
If you have a spare band with a great drummer handy, I'm all ears.
Currently, I prioritize my practice schedule, which is showing some
returns during live recording. Point taken.
> 2. Better vocal quality. I have an idea of the vocal sound you are trying
> for. Nothing that a decent mic and a decent vocal channel (or pre-amp /
> compressor combination) wouldn't solve
If I'd done as much homework on recording gear I'd have known this
years ago. Mic's key. I actually thought the audio-technica was good
when I bought it.
> 3. Better effects - the reverb and other effects you are using betray their
> origins.
Attempting to use the newer gear in more effective ways. Just learned
aux loops this weekend...
>
> If I were you, I would beg, borrow or swap some of my existing equipment for
> a better computer as the starting point to constructing a better DAW. You
> don't need anything too fancy - you'd find a second-hand PII/300MHz enough
> for what I'm going to suggest - try for one with a decent motherboard (ASUS
> is my favourite) with at least UDMA 66 (pref. UDMA100) disk controllers.
I was slightly surprised by this suggestion. The machine I'm using is
an intel mobo with IBM 7200 rpm drives. PIII 733MHz and 256 MBytes
PC133 RAM.
<ANTI-M$ RANT BEGINS>
It is Windows on Intel hardware. It's Windows so there's a lot of
overhead. It's Windows, so it messes things up a lot. It's just
Windows, so it crashes during peak creative moments. That's just the
life you get as an artist when there's no competition in the consumer
DAW OS platform space. (READ MY SIG, BILL)
Someone want to port Cakewalk to Linux?
<ANTI-M$ RANT ENDS>
> Ask around (especially with your background) and you'll be sure to find
> someone throwing a machine with spec similar to this out or selling it off
> for $100 or less.
I saw a G4 800 for $1300 today. I'm watching...
>
> Your SoundBlaster is not ideal but it will do for now - and it has the
> advantage of its soundfont capability which you might find very useful.
Might be something worth exploring here...
>
> Computer in place, you now have the room to explore more of Cakewalk's
> functionality.
... of course ...
> BTW - don't over-compress during recording. It's an easy thing to do but
> compression (like reverb) can't be undone later. You want only enough
> compression to keep the levels in hand - you can add more compression later
> to get the sound you want.
Compression is amazing. I've used it in the past, but knowing what it
actually does has been beneficial. There's nothing better than a
clean signal on a multi-track recorder, tho'. That becomes obvious
quickly...
> b) multi-track and layer sounds without having to bounce as you do at
> present - you might, for example, find it useful to double-track your vocals
> on some songs.
There will always be bouncing/ping-ponging. I had blazes of fun using
MTC/MMC this weekend. Cool double-track suggestion...
> c) apply better effects - even Cakewalk's bundled plug-ins should sound
> better than those you use at present. As time and funds allow you can easily
> add better reverbs/delays/flangers. The quality of some of the free plug-ins
> is impressive in any case.
{At some point musicians will be replaced by plug-ins.} Good point,
though...
> If your version of Cakewalk is old, you might also consider n-Track software
If you only knew. (Ever heard of Dr. T's Sequencing Software for
Commodore 128?) Cakewalk 2.0 for Win 3.1 (16-bit). I also have
Cakewalk Express which came with the Platinum. Just the original
question relating to Intel vs. MAC platform.
> from FA Soft (www.fasoft.com) which is used with great affection by many
> here.
>
> I couldn't suggest you give even a passing glance at ProTools. It is a
> closed, proprietary system that requires very expensive hardware to run. It
> would not provide you with anything you can't do at least as well and far
> cheaper.
Highly encouraging. On paper, this is what I've thought all along.
More experiences will probably be useful/insightful...
> Hopefully this gives you at least some food for thought. I hope I have set
> out a road map you might be able to travel over time.
I dunno about a road map, but quite helpful, indeed!
> Again, welcome to am4t - let the adventure begin!
Or, continue, as it were. >:)
> --
> George
> Newcastle, England
> (please remove leading 'x' from email address to reply directly - thanks)
Thanks much for such an exhaustive and useful response, George.
You've given me a tremendous amount to go on.
Best Regards,
Van
=================================================================
http://www.dedserius.com/
=================================================================
George Perfect
news:UQtn8.71861$Gf.64...@bin2.nnrp.aus1.giganews.com.
>
> Been there, done that. You end up with one hell of a noisy recording.
> There's no way you can reduce good clean audio to 4 or 8 bits and
> hold the sample rate the same, and not degrade its sound.
>
Hhmmm - maybe we're getting somewhere. But don't be coy Arny - there's a
sight more than noise involved - as you admit further on.
>
> It's a ludicrous experiment because 8 bits is very noisy. However, I
> provide samples that degrade a 24 bit recording to 16, 15, 14, 13,
> 12... bits so that people can evaluate this issue for themselves. I
> provide samples that use dither and don't use dither. I even provide
> a sample that uses dither, but not enough. Here's the place to
> listen: http://www.cdabx.com/technical/bits44/index.htm .
>
No - yours was a ludicrous statement. You didn't specify any limits. You
stated
>>
>> If the recording is properly dithered, the reduction of bit depth
>> manifests itself only as decreased signal-to-noise ratio.
>>
I refer you to the word "only". This experiment shows that statement is not
true.
> > You can also test Arny's statement this way. If you have a
> reasonable
> > graphics card in your computer (one that handles 16 bit or 24 bit
> colour
>
> Yet another ludicrous experiment.
>
Not at all - the experiment shows what is happening when bit depths are
reduced
> Basically, George is arguing that you can surely taste the difference
> caused by adding a tiny grain of salt to a barrel of water because
> you can taste the difference caused by adding a pound of salt to a
> teacup.
No, I'm arguing that you are (once again) wrong.
>
> That would be solarizing. There is such a thing as reducing color
> depth or gray scale resolution with dither, and the result is that
> you don't get the banding or solarization effect.
>
Er ... "solarizing" is the name given to the effect I describe. The point
is, the effect exists and can be demonstrated - not that it has a name. I
recall describing the use of dither on images and explaining that it has no
effect on large, continuous coloured areas. What you will see are the
"steps" resulting from the quantisation process that I first described at
the very start of this argument.
>
> If you reduce the color or gray scale resolution in a similar
> fashion, the outcome will be analogous to what happens with audio
> signals.
>
Finally - you agree with me on something.
> Since George mentioned Paintshop Pro, let's talk about Paintshop Pro.
> I have version 4.12 which I use for basic web art. I use Corel Draw
> and Photopaint for more creative work. In Paintshop Pro you can
> reduce color depth using the Colors, Decrease Color Depth Command bar
> option. When you do this you get a dialog box that has an option
> called "Error Diffusion". This is roughly like dithered
> requantization, and eliminates the color banding that George claims
> will always happen. Try converting a high resolution color
> photograph both ways!
I suggested that people do this very thing. Colours across continuous areas
will no longer be the same. This is quantisation error remaining in the
analogue output. Exactly as I described it. And exactly as it would affect
an audio signal.
As I have said until I am blue in the face, the same effect applies when
reducing from (say) 24 bits to 16 bits. Dither will blur low level signals -
in the same way that it blurs the otherwise sharp steps between colour bands
in my imaging example. It will do precisely *nothing* for other signals and
the audible effects of the equivalent of the colour banding can be clearly
seen - and even at elevated bit and sample rates - will be heard by some on
some types of material.
The question 'at what point are we no longer able to hear the damage' is an
important one (and one that is far more relevant and interesting than the
matters we are discussing) but - and here's the important point - we only
hamper our ability to answer that question by pretending the damage does not
occur or by misrepresenting the mechanism that causes it.
A proper, controlled set of tests might establish the level at which people
can perceive audible differences of the kind caused by these mechanisms.
There remains the question of what to do with that knowledge. How do we
avoid audible distortions? What can be done to improve matters - maybe move
the hurdle that little bit higher? How can we use this knowledge to improve
the quality of our recordings?
Your approach does nothing to advance our knowledge or understanding. It is
aimed fairly and squarely at just one objective - the promotion of one Mr
Arny Krueger.
>
> > Dither has nothing to do with it.
>
> Dither has everything to do with it.
>
I do believe I have shown the limitations of the usefulness of dither. It is
far from being the 'magic cure-all' you claim it to be.
> >In fact, if you use
> > Photoshop or Paintshop Pro you can choose different dither
> algorithms as you
> > perform the bit reduction.
>
> Exactly.
>
Er ... didn't you just claim I hadn't said this?
> >Feel free to try them - if you succeed in finding
> > one that matches the specific image you used you may find that the
> edges
> > between the bands become blurred. But ... if your monitor (or
> printer) is
> > good enough that picture is never going to look like the original
> with just
> > a little noise added.
>
> The key phrase here is "a little noise". That's because most of the
> color reductions we do are pretty gross. Therefore the consequence is
> that we get pictures that look like the origin with LOTS of noise
> added.
No Arny. The point is that noise (in whatever quantity) is but one of the
effects of bit reduction. There are other, far more important and much more
visible/audible mechanisms at work.
>
> > Zoom in on the image and you will note that each pixel can only
> take one of
> > 256 colours. If you used dither (or the program added dither
> automatically)
> > you will see that adjacent pixels that take the same colour or were
> maybe
> > only one or two colour steps removed in the original now take some
> random
> > value - this is the noise that has been added.
>
> Right.
>
> > Zoom back out and compare several areas of the original image with
> the same
> > area in the bit-reduced copy and you will see a different colour.
> > Occasionally the difference will be pronounced.
>
> Right.
>
Er ... you agree that colours will have been changed? But you said
>>
>> If the recording is properly dithered, the reduction of bit depth
>> manifests itself only as decreased signal-to-noise ratio.
>>
As this is not "decreased signal to noise ratio" but a clearly visible
distortion in the signal I believe my point is proven and the lie of your
statement (and your claims on this topic right back to the start of this
thread) revealed.
> Again, its a matter of degree.
No Arny, it isn't. Your statement was absolute. You claimed that the *only*
effect of bit reduction is a reduction in signal to noise ratio. In fact the
original signal is also corrupted / distorted. The extent to which this
occurs will vary with the amount of bit reduction applied but it will still
happen. It is more significant than any noise introduced. You denied its
existence.
> If you reduce the resolution of a high
> quality audio signal to 8 bits, there's no way that you won't hear
> the difference. However, just about everybody prefers added noise to
> added distortion. In either case you're going to hear it if you can
> listen carefully. Pick your poison - the damage will be done! You
> just get to pick which kind of damage.
No you don't. You get to pick a masking effect that might cover a small part
of the problem. That's all. And that is all dither does.
> George has been told many times by many people that he's wrong. I've
> cited 3 AES papers for him. At least two web sites besides my own.
> But he keeps on like this. If John Atkinson didn't hate me so much, I
> suspect that he'd also try to straighten George out.
>
Name the "many people".
The AES papers you cited do not support - let alone "prove" - your idiotic
views.
The web sites you quoted illustrate and describe the effect I am
describing - when I brought this to your attention you claimed it was the
"artwork" that was at fault.
John Atkinson has written to me privately, not to "straighten me out" but to
support the statements I have made. He seems well balanced to me. Like me,
he doesn't "hate" you - just the misinformation you spread and the damage
you do to people on Usenet who dare to show how wrong you are.
I'll let John respond to this latest baseless allegation - that he would in
some way pervert what he knows to be true (in the very area on which he
writes and makes his living, no less) just to score a cheap point against
you.
I have in the past called you paranoid. I suggest this is further evidence
to support that opinion.
Anyone can set up a web site and publish whatever they want. Anyone can spam
every newsgroup on the planet with self-promoting puff. Anyone can set
themselves out as an expert. But none of this actually makes you an expert.
All the bluster, name-calling and self-proclamation in the world will not
make you what you so desperately want to be - an authority on digital audio.
Time and again you have been told that the only way to establish your
credentials is to publish and subject yourself to peer review. This is the
only way that science is conducted. This isn't a challenge (in the
aggressive sense that you interpret almost everything said to you) but
something that you must do if you are to be acknowledged as someone with
expertise.
I suspect you do not make the attempt because you know fine well what will
happen. Your "research" will be shown for what it is (repetition and
tabulation of simple tests already conducted every day by others with far
less fuss) and your conclusions daft and idiotic. If you wish to prove me
wrong in these opinions, please feel free to submit your papers to the JAES.
As matters stand, your views published on your web site are seen as
ridiculous by your peers so I suggest you alter them to something more
credible before formal submission.
Your views on digital audio can most kindly be described as simplistic. It
is easy to see how you reach your conclusions by conveniently ignoring the
detail surrounding the issues you focus on. The mistakes you make are those
of a first year college undergraduate. Whether you are quoting AES papers,
Nyquist, Shannon or the Maharaji you have a great skill for taking one
small, barely understood part of what they have to say and leaping to some
unique, personal revelation. That no-one else appears to share in this
revelation does not mean we are all out to get you. It means that you are -
in almost certain probability - wrong.
Of course, in your self-appointed role of ultimate authority, you can never
be wrong. It is the tactics you employ to insult and denigrate those who
correct your errors that make you such an objectionable character. You may
have discovered that seeking to damage the reputations of respected figures
wins you some friends among those who know no better but other than in your
delusional world such actions prove nothing.
And, let's not forget that your authority does not stop at digital audio. It
extends to recording techniques, musicology and computing among other
diverse fields. In each of these fields, your pronouncements (always right,
huh?) litter Usenet.
Whenever you are cornered with some inarguable proof of your
wrong-headedness you seek to divert and confuse or (as we have recently seen
here) set ridiculous conditions to avoid having to answer your critics.
Anything, in fact, to avoid having to accept that you are wrong.
I see that three of the people you attacked on rec.audio.opinion finally
resorted to threats of legal action against you - one a respected doctor who
faced the loss of a patient who was alarmed by the unfounded, nasty and
spiteful accusations you threw at him. Faced with this final arbiter of
proof, did you rise to the challenge? Have you had your day in court and
"proven" the scientific basis of your self-pronounced wisdom?
No. You high-tailed it out of the discussion pronto.
You have called me some high old names here too Arny and attempted to damage
my reputation too with your mud-slinging. Well, my reputation is important
to me and I do not take kindly to having the stains you have sought to
spread on my character and my own humble expertise left in the public record
unchallenged.
Perhaps you should be hearing from my lawyers? How about it? Would you like
an opportunity to establish once and for all the veracity of your
pronouncements and your unchallengeable role as ultimate authority in the
field of digital audio? I'll happily provide the platform.
All you have to do is continue slandering me and I'll issue the
instructions.
Arny Krueger
> "Arny Krueger" <ar...@comcast.net> wrote in message
> news:UQtn8.71861$Gf.64...@bin2.nnrp.aus1.giganews.com.
> >
> > Been there, done that. You end up with one hell of a noisy
recording.
> > There's no way you can reduce good clean audio to 4 or 8 bits and
> > hold the sample rate the same, and not degrade its sound.
> >
>
> Hhmmm - maybe we're getting somewhere.
If progress is being made, it's in the area of understanding of what
I've been saying all along,
>But don't be coy Arny - there's a sight more than noise involved -
as you admit further on.
No such thing!
> > It's a ludicrous experiment because 8 bits is very noisy.
However, I
> > provide samples that degrade a 24 bit recording to 16, 15, 14,
13,
> > 12... bits so that people can evaluate this issue for themselves.
I
> > provide samples that use dither and don't use dither. I even
provide
> > a sample that uses dither, but not enough. Here's the place to
> > listen: http://www.cdabx.com/technical/bits44/index.htm .
> No - yours was a ludicrous statement. You didn't specify any
limits. You
> stated
> >> If the recording is properly dithered, the reduction of bit
depth
> >> manifests itself only as decreased signal-to-noise ratio.
That's right, and it happens without limits. Even in your ludicrous
examples, George, if the recording is properly dithered the reduction
of bit depth manifests itself only as decreased signal-to-noise
ratio.
> I refer you to the word "only". This experiment shows that
statement is not true.
The experiment shows that even in ludicrous cases, if the recording
is properly dithered the reduction of bit depth manifests itself
only as decreased signal-to-noise ratio. This takes place even if
the recording is made with one bit. In fact, oversampled 1 bit
through 8 bit converters are widely used these days to provide high
quality results. The noise issue is dealty with by means of
oversampling followed by decimation.
> > > You can also test Arny's statement this way. If you have a
> > reasonable
> > > graphics card in your computer (one that handles 16 bit or 24
bit
> > colour
>
> > Yet another ludicrous experiment.
> Not at all - the experiment shows what is happening when bit depths
are
> reduced.
The experiment shows that even in ludicrous cases, if the recording
is properly dithered the reduction of bit depth manifests itself
only as decreased signal-to-noise ratio.
> > Basically, George is arguing that you can surely taste the
difference
> > caused by adding a tiny grain of salt to a barrel of water
because
> > you can taste the difference caused by adding a pound of salt to
a
> > teacup.
> No, I'm arguing that you are (once again) wrong.
The experiment shows that even in ludicrous cases, if the recording
is properly dithered the reduction of bit depth manifests itself
only as decreased signal-to-noise ratio.
> > That would be solarizing. There is such a thing as reducing
color
> > depth or gray scale resolution with dither, and the result is
that
> > you don't get the banding or solarization effect.
> Er ... "solarizing" is the name given to the effect I describe. The
point
> is, the effect exists and can be demonstrated - not that it has a
name. I
> recall describing the use of dither on images and explaining that
it has no
> effect on large, continuous coloured areas. What you will see are
the
> "steps" resulting from the quantisation process that I first
described at
> the very start of this argument.
It's typical of what happens when you don't dither while
requantizing.
> > If you reduce the color or gray scale resolution in a similar
> > fashion, the outcome will be analogous to what happens with audio
> > signals.
> Finally - you agree with me on something.
If you want me to agree with you George, say something that is
correct!
> > Since George mentioned Paintshop Pro, let's talk about Paintshop
Pro.
> > I have version 4.12 which I use for basic web art. I use Corel
Draw
> > and Photopaint for more creative work. In Paintshop Pro you can
> > reduce color depth using the Colors, Decrease Color Depth Command
bar
> > option. When you do this you get a dialog box that has an option
> > called "Error Diffusion". This is roughly like dithered
> > requantization, and eliminates the color banding that George
claims
> > will always happen. Try converting a high resolution color
> > photograph both ways!
> I suggested that people do this very thing. Colours across
continuous areas
> will no longer be the same.
That's right, they are dithered.
>This is quantisation error remaining in the
> analogue output. Exactly as I described it. And exactly as it would
affect
> an audio signal.
Right, but the banding you predicted George, doesn't happen when the
"Error Diffusion" option (which is something like dithering) is
selected.
> As I have said until I am blue in the face, the same effect applies
when
> reducing from (say) 24 bits to 16 bits.
That's right, if the recording is properly dithered the reduction of
bit depth manifests itself only as decreased signal-to-noise ratio.
> Dither will blur low level signals -
> in the same way that it blurs the otherwise sharp steps between
colour bands
> in my imaging example. It will do precisely *nothing* for other
signals and
> the audible effects of the equivalent of the colour banding can be
clearly
> seen - and even at elevated bit and sample rates - will be heard by
some on
> some types of material.
That's right, if the recording is properly dithered the reduction of
bit depth manifests itself only as decreased signal-to-noise ratio.
> The question 'at what point are we no longer able to hear the
damage' is an
> important one (and one that is far more relevant and interesting
than the
> matters we are discussing) but - and here's the important point -
we only
> hamper our ability to answer that question by pretending the damage
does not
> occur or by misrepresenting the mechanism that causes it.
That's right,if the recording is properly dithered the reduction of
bit depth manifests itself only as decreased signal-to-noise ratio.
> A proper, controlled set of tests might establish the level at
which people
> can perceive audible differences of the kind caused by these
mechanisms.
> There remains the question of what to do with that knowledge. How
do we
> avoid audible distortions? What can be done to improve matters -
maybe move
> the hurdle that little bit higher? How can we use this knowledge to
improve
> the quality of our recordings?
That's right, if the recording is properly dithered the reduction of
bit depth manifests itself only as decreased signal-to-noise ratio.
But it does manifest itself as a reduction in signal-to-noise ratio
which can be audible in extreme cases. I never said otherwise.
<gratuitous personal attacks snipped>
Arny Krueger
"Chris Hornbeck" <guyv...@removethisaristotle.net> wrote in message
news:25ss9u0m60h8coi5m...@4ax.com...
> On Mon, 25 Mar 2002 00:05:09 GMT, "Arny Krueger"
<ar...@comcast.net>
> wrote:
> >In a photograph you can reduce the color or gray scale with dither
> >and get a grainy picture, or you can reduce it without dither and
get
> >a banded picture. Most people seem to prefer the noise. In audio
you
> >can reduce the resolution with dither and pick up some extra
noise,
> >or you can reduce the resolution without dither and pick up some
> >distortion.
> This reminds me that there is (are?) Photoshop plug-in(s) that
apply
> a fractal algorithm for scaling. I wonder if a similar technique
> is ever used in audio mixing?
It is not used that I know of. Not that it hasn't been tried.
> Ferinstance, someone with a bunch of 16/44.1 tracks and access to
> a fancy 24/96 mixer. Might be some advantage to bring the originals
> into the larger space with some "nonlinearities".
One serious problem here is the adequacy of 16/44 audio. In audio it
isn't that tough to digitize signals with overkill levels of bit
depth and sample rate. It is a lot harder to digitize photographs
with overkill levels of bit depth and sample rate.
>Kinda like an
> anti-dither; randomness/noise to round the edges. Just a thought.
Its a thought that has been looked at before. If it was only that
easy...
> BTW, solarization has a different meaning in photography.
> Thanks for the insurrection,
;-)
Ethan Winer
> I agree that the way you did the test makes sense, but the 13
bit test is ultimately measuring a 3 bit change, not an 11 bit
change. <
It wasn't the amount of change I was after, but rather the
minimum number of bits needed to reproduce a musical instrument
faithfully.
> But now I'm curious if you took the 24 bit file and
did -/+60db (or whatever amount was appropriate) to get it down
to 9 bits, then dithered it from 24 (with padded zeros) down to
16 bits, how would that sound? <
You can start with 24 bits or 124 bits. If you reduce a
full-scale recording to 9 bits, then that establishes the
quality.
> If [dithering] does affect more than 1 bit, would dithering
from 16 bit to 16 bit be a valid option? <
Somehow I don't think so, but Arny is in a better position to
answer that than I am.
> Mostly, I learned I really need a quieter computer. <
Yeah, this seems to be a common problem. My last computer had a
PC Power & Cooling brand power supply which helped a lot. My
current computer is a Dell and its (stock) power supply is even
quieter. I have it under a corner of my desk, and the top,
corner, and back "walls" under the desk are lined with one-inch
thick 703 fiberglass board. While it is not totally quiet, the
computer noise is low enough that I can sit with my back to the
computer and play my cello, and record that with mikes pointing
right at me and the computer.
--Ethan
Ethan Winer
Arny Krueger
> > I agree that the way you did the test makes sense, but the 13
> bit test is ultimately measuring a 3 bit change, not an 11 bit
> change. <
> It wasn't the amount of change I was after, but rather the
> minimum number of bits needed to reproduce a musical instrument
> faithfully.
> > But now I'm curious if you took the 24 bit file and
> did -/+60db (or whatever amount was appropriate) to get it down
> to 9 bits, then dithered it from 24 (with padded zeros) down to
> 16 bits, how would that sound? <
> You can start with 24 bits or 124 bits. If you reduce a
> full-scale recording to 9 bits, then that establishes the
> quality.
Right.
> > If [dithering] does affect more than 1 bit, would dithering
> from 16 bit to 16 bit be a valid option? <
> Somehow I don't think so, but Arny is in a better position to
> answer that than I am.
Once something has been truncated, after-the-fact dithering is pretty
futile.
> > Mostly, I learned I really need a quieter computer.
How about that!
;-)
I put mine in another room, and run extensions for the keyboard,
mouse, and audio cables.
> Yeah, this seems to be a common problem. My last computer had a
> PC Power & Cooling brand power supply which helped a lot. My
> current computer is a Dell and its (stock) power supply is even
> quieter.
Dell and Gateway sometimes use el-humongous CPU heat sinks that I
wish I could buy to put in the PCs I build. This can get you down to
just a power supply fan, which obviously reduces noise. However,
there is still the possibility of noise from the drives.
>I have it under a corner of my desk, and the top,
> corner, and back "walls" under the desk are lined with one-inch
> thick 703 fiberglass board. While it is not totally quiet, the
> computer noise is low enough that I can sit with my back to the
> computer and play my cello, and record that with mikes pointing
> right at me and the computer.
IME, there is nothing like having the computer out of the room. It
makes working with CDs a bit awkward, though. I often do that with a
second, networked computer that is also someplace else.
Ethan Winer
> Dell and Gateway ... This can get you down to just a power
supply fan ... there is still the possibility of noise from the
drives. <
My drives are both dead silent. Only the fan is an issue, and
as I said it's quiet enough to be workable even with a cello.
And when I really care about the noise, I sit in the middle of
the room and point the mike away from the computer instead of
toward it. Of course, when I mike my guitar amp the computer
fan noise doesn't matter. :->)
> IME, there is nothing like having the computer out of the
room. It makes working with CDs a bit awkward, though. <
Yeah, inaccessibility is the show-stopper for me.
--Ethan
Glenn Booth
"George Perfect" <xgeo...@byline.co.uk> wrote in message
news:3c9dd089$0$8508$ed9e...@reading.news.pipex.net...
> You can also test Arny's statement this way. If you have a reasonable
> graphics card in your computer (one that handles 16 bit or 24 bit colour
> depth) load a 24 bit encoded ("milions of colours") photograph which
> exhibits some finally graded colour variations - a picture containing sky
> and clouds will often do. Now without changing the image size (ie, sample
> rate) convert the picture to 16 colour depth (4 bits).
Sorry, no cigar. You simply cannot test Arny's statement this way,
for the simple reason that there are no monitors that can display
those 24 bits worth of colours at once. Note that I say "at once";
even a huge resolution like 2048*1536 (about the highest I
regularly encounter) is only around 3.14 million pixels. If I can find
an image that will drive this display so that every pixel receives
a unique RGB triple ('colour'), I can only represent about one
in every five colour values that are in the file. The 'true colour'
file isn't true colour on the output device. The are bunch of
other problems, like 8 bit colour modes being read from a look
up table and 24 bit modes not, but I won't get into that. It's
too much like work.
The test isn't valid, as the output device places an upper limit
on the perceived number of colours in the 'real world'. i.e.
it's throwing data away in order to let us see the image. As
luck would have it, this usually involves dithering.
Add to this the fact that the human visual system is non-linear (in
different ways to the auditory system) and the analogy just
doesn't hold. We can differentiate more shades of green, for
example, than we can of blue, so the results will vary with the
type of image. Here, there is a valid analogy, as we have different
perceived sensitivities to audio of different frequencies.
> This process is IDENTICAL to what happens when you reduce the bit rate of
an
> audio signal. Dither has nothing to do with it.
First, it's not IDENTICAL at all, due to the nature of the
human visual system and the way we have to view the
image. Second, most graphics cards these days have dither
built in to the hardware, and for the most part you can't disable it,
so again the display system screws the experiment up for you.
Were it otherwise, the algorithms used in audio processing would
be completely re-useable for video and graphics. They aren't,
despite similarities. In photoshop, whatever Adobe claim,
what you see is NOT what you have on the disk.
I understand the point of the argument, however, and while it's
intuitive, it's fundamentally flawed.
> I leave it to you to imagine what the process does to an audio signal.
It's impossible to extrapolate your argument from the visual
world to the auditory world.
> If there was anything in Arny's 'free lunch' theory we'd all be listening
to
> 4 bit encoded CDs.
...and viewing monitors that could simultaneously display over
16 million colours? I'm with Arny on this one.
Regards,
Glenn.
Glenn Booth
"Chris G." <chri...@txdirect.net> wrote in message
news:u9rq22k...@corp.supernews.com...
> herbal remedies for various ailments. I swear by my Echinacea for
example.
> Science be damned it shortens or sometimes halts colds and flu in its
tracks
> when I take it.
If it works for you, then do it. No argument from me. However, it's
fair to point out that Echinacea is far from a homeopathic treatment.
It has active ingredients that can be identified and studied.
I have to ask how you know that the echinacea shortened your
cold, though. How long would it have lasted if you had not
taken it?
A *real* homeopathic treatment has *no* measurable active
ingredients whatever - the whole point is to dilute, dilute, dilute
and then dilute some more, until you end up taking, erm, water
that had an ancestor that once contained some active ingredient.
My wife (a doctor) has been looking hard for real evidence of
a provably effective homeopathic treatment for years, and has yet
to find any. That is not to say that homeopathy does not work. Placebos
have been proven to work many times, especially in (gulp) double
blind trials, but you can't measure and study the reasons why.
(dons asbestos underwear).
Regards,
Glenn.
George Perfect
> Sorry, no cigar. You simply cannot test Arny's statement this way,
> for the simple reason that there are no monitors that can display
> those 24 bits worth of colours at once. Note that I say "at once";
> even a huge resolution like 2048*1536 (about the highest I
> regularly encounter) is only around 3.14 million pixels. ...
Please read up on the difference between image resolution (pixels) and
colour depth.
Then try the test
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Glenn Booth" <do...@removeqtlg.demon.co.uk> wrote in message
news:1017431502.24401....@news.demon.co.uk...
Arny Krueger
> You lose this one in the first paragraph:
>
> > Sorry, no cigar. You simply cannot test Arny's statement this
way,
> > for the simple reason that there are no monitors that can display
> > those 24 bits worth of colours at once. Note that I say "at
once";
> > even a huge resolution like 2048*1536 (about the highest I
> > regularly encounter) is only around 3.14 million pixels. ...
>
> Please read up on the difference between image resolution (pixels)
and
> colour depth.
>
No George, you just lost it. You completely missed his point! I'd
explain it to you, but that would be redundant since you claim to
know it all!
;-)
Denny Fohringer
> A week ago I invited folks here to compare a series of audio
> files, to see how audible are the differences between dithering
> from 24 bits to 16 versus truncating...The short answer is that
>most folks were not able to distinguish
>any of the files.
<snip>
I'm not in the least surprised. With regard to this comment: "So it seems
that grit - whether
caused by analog tape or digital artifacts - can make a
recording sound "better" ; I am reminded of how the early Aphex "Aural
Exciter" worked (and how they may still work, for all I know). Basically
they just split the signal with a hi-pass filter, distorted the hell out of
the high band part to "add harmonics" then mixed a little of this clipped
portion back in the full-bandwidth signal. This is how they added "air" and
"brilliance". Lots of hi frequency fuzz. I guess my point is a lot of
perceived "fidelity" is a psycho-acoustic thang.
> Failings of this test: Several people pointed out that the real
> advantage of 24-bit recording is when an entire project is
> recorded and mixed at 24 bits, because the increased resolution
> helps minimize errors that accumulate from gain changes and
> plug-in processing.
I can see how greater bit depth would help reduce "rounding errors" which
would inevitably be introget duced by processing algorithms. But I think you
would pretty much the same results by dithering "up" a 16-bit file before
you do any processing (like Cool Edit allows you to do), then dither back
down when you're done.
I'm pretty sure the 96Khz, 24-bit hardware issue is one of marketing new
hardware to a saturated market rather than real-world audio improvement.
But then I'm a cynical bastard . . .
Still recording at 44.1Khz, 16-bit . . .
--------------------------------------------------
Denny Fohringer
Itinerant guitarist
--------------------------------------------------
Lessons and music:
http://surf.to/dennyf http://mp3.com/dennyf
Bands:
http://listen.to/bluepearl http://listen.to/doubletake
--------------------------------------------------
Ethan Winer
> I am reminded of ... the early Aphex "Aural Exciter"
Yes, and when that was first offered in the 1970s, it too was
accompanied by outrageous BS technical explanations. I remember
a full-page graph in R-e/p magazine showing how the Aphex
manipulates the phase to achieve its stunning clarity, or
whatever nonsense term they used. Of course, the Aphex Aural
Exciter does no such thing. Now BBE makes the exact same BS
claims for their Sonic Maximizer, stating that it re-aligns the
phase to compensate for loudspeaker inaccuracies or some such
nonsense.
> I can see how greater bit depth would help reduce "rounding
errors" <
Yeah, except those errors amount to like 0.00001 % added
distortion, so who cares.
> But then I'm a cynical bastard . . .
There needs to be more of us. :->)
--Ethan
Suso Ramallo
If you know any headphone amplifier with 96 dB of dinamic range or better
(be carefull with your ears), maybe
Regards
"Chris G." <chri...@txdirect.net> escribió en el mensaje
news:u9rp4n2...@corp.supernews.com...
> Headphones anyone? (uh oh..I better duck).
>
> Chris G.
>
> --
> "Getting eaten by a crocodile is JUST like falling asleep... in a
blender!"
> - Homer Simpson
Glenn Booth
"George Perfect" <xgeo...@byline.co.uk> wrote in message
news:3ca5fe55$0$8507$ed9e...@reading.news.pipex.net...
> You lose this one in the first paragraph:
>
> > Sorry, no cigar. You simply cannot test Arny's statement this way,
> > for the simple reason that there are no monitors that can display
> > those 24 bits worth of colours at once. Note that I say "at once";
> > even a huge resolution like 2048*1536 (about the highest I
> > regularly encounter) is only around 3.14 million pixels. ...
>
> Please read up on the difference between image resolution (pixels) and
> colour depth.
I have done. Foley and Van Dam's "Computer Graphics" is my usual
reference when I'm training people about this stuff.
I'll try again.
Image resolution is just the number of pixels, for the sake of this
argument. (Photographers would disagree, but that doesn't matter
here, as we're not concerned with the distance between 'dots',
merely the number of them).
Colour depth is the number of bits used to store the colour value
of each pixel.
Back to my point. Take an image that contains 24 bits per pixel of
colour information. Since each pixel has one and only one colour
value, an image that contains EVERY possible colour value must
be at least 2^24 pixels at the output. That means around 16.8 million
pixels are required to represent one pixel at each possible colour.
Now try to find a monitor that can display this. Can't? No, you can't,
because they are extremely hard to find . I've worked for graphics
card manufacturers as an engineer for more than 10 years, and
I've never seen one. 24 bit colour depths are only meaningful
as a measure of the available choices of colour that can be sent
to the display, not the number of colours on screen at a certain
time.
The 'control' for your experiment is no control at all, since it is
compromised at the display device. The display is not showing
all the colours in the file; it can't.
> Then try the test
I don't need to. I know what will happen, as I've done it hundreds,
or thousands, of times over the last ten years. The maximum number
of colours available on a monitor displaying an image is
equal to the number of pixels the monitor is drawing, since each pixel
must be one colour. If any colour is reproduced more than once, then
the maximum number goes down correspondingly.
We could add that the gamut of the display may not match the gamut
of the image, adding further problems. We could also talk about
sub-pixel rendering, hardware dithering, gamma, the white point of the
monitor
and a bunch of other limitations on the experiment, but since the starting
point of the argument is flawed (A complete 24 bit image can't be viewed
in 24 bit colour with available display technology), these other
restrictions
don't matter until we can get an output device that can draw 5000*3500
pixels or thereabouts.
Bringing printers into the argument changes things somewhat, but then
you have to convert between two, or more than two, colour spaces,
so there is still a problem.
Regards,
Glenn
George Perfect
> Hi,
>
>
> Colour depth is the number of bits used to store the colour value
> of each pixel.
>
Correct
> Back to my point. Take an image that contains 24 bits per pixel of
> colour information. Since each pixel has one and only one colour
> value, an image that contains EVERY possible colour value must
> be at least 2^24 pixels at the output. That means around 16.8 million
> pixels are required to represent one pixel at each possible colour.
>
Que? I'd agree that each pixel would have to be capable of displaying any of
2^24 possible colour values but why on earth do you need 2^24 pixels at each
resolution point to do this?
This is simply not the way that monitors work. If this were the case then
even at 8 bit colour depth you would need 256 (what shall we call them?)
colour dots(??) at each pixel to display the image.
> Now try to find a monitor that can display this. Can't? No, you can't,
> because they are extremely hard to find . I've worked for graphics
> card manufacturers as an engineer for more than 10 years, and
> I've never seen one. 24 bit colour depths are only meaningful
> as a measure of the available choices of colour that can be sent
> to the display, not the number of colours on screen at a certain
> time.
You are missing the point of my test. The point is that when reduced to 4
bit (or even 8 bit) colour depth an image containing graduated colour will
display as bands of that colour. I could as easily have chosen 48 bits, 27
bits or any much higher value for the 'control'.
If (as you argue) the graphics card applies signal shaping automaticaly then
this is akin to what I said already goes on in some audio DACs in order to
improve their measured performance at the expense of their broad-range audio
performance.
>
> The 'control' for your experiment is no control at all, since it is
> compromised at the display device. The display is not showing
> all the colours in the file; it can't.
>
Any more than you might argue that a loudspeaker can't reproduce 65,536
amplitude values from a 16 bit encoded audio signal? Or - if it can, perhaps
it can reproduce 65,537 or 128,000 or 256,000 or ... where should we draw
the line? Or is that question too close to the point of all this?
> > Then try the test
>
> I don't need to. I know what will happen, as I've done it hundreds,
> or thousands, of times over the last ten years. The maximum number
> of colours available on a monitor displaying an image is
> equal to the number of pixels the monitor is drawing, since each pixel
> must be one colour. If any colour is reproduced more than once, then
> the maximum number goes down correspondingly.
But my experiment only requires a much reduced number of possible colours to
be displayed.
My point is that when this number is greatly reduced the display *WILL*
start to show colour bands where none existed in the original image.
If this is too much to grasp then do this instead.
Load your favourite graphics program, open a blank image and flood fill it
with a colour chosen from a 24 bit palette. Save the image as a 24 bit file.
Now set your monitor to "16 colours" and display the image again.
Its displayed colour will have changed.
What yo have just observed is the quantisation process at work.
>
> We could add that the gamut of the display may not match the gamut
> of the image, adding further problems. We could also talk about
> sub-pixel rendering, hardware dithering, gamma, the white point of the
> monitor
> and a bunch of other limitations on the experiment, but since the starting
> point of the argument is flawed (A complete 24 bit image can't be viewed
> in 24 bit colour with available display technology), these other
> restrictions
> don't matter until we can get an output device that can draw 5000*3500
> pixels or thereabouts.
>
> Bringing printers into the argument changes things somewhat, but then
> you have to convert between two, or more than two, colour spaces,
> so there is still a problem.
If it makes you any happier, do the experiments in grey-scale.
My point still applies. And the example I gave will display the bands I
described.
Glenn Booth
"George Perfect" <xgeo...@byline.co.uk> wrote in message
news:3ca8a56a$0$8514$ed9e...@reading.news.pipex.net...
> "Glenn Booth" <do...@removeqtlg.demon.co.uk> wrote in message
> news:1017682924.27121....@news.demon.co.uk...
> > Hi,
> > Back to my point. Take an image that contains 24 bits per pixel of
> > colour information. Since each pixel has one and only one colour
> > value, an image that contains EVERY possible colour value must
> > be at least 2^24 pixels at the output. That means around 16.8 million
> > pixels are required to represent one pixel at each possible colour.
> >
> Que? I'd agree that each pixel would have to be capable of displaying any
of
> 2^24 possible colour values but why on earth do you need 2^24 pixels at
each
> resolution point to do this?
You don't. However, that wasn't my point (which has now become clouded!).
You need one piece of phosphor for each of red, green, and blue (The
RGB primaries). However, my point is that a 24 bit file (our reference) is
not 24 bit when it gets to the monitor; any monitor. What you see when
you display a 24 bit file on a high-res modern monitor is nearer 20 or
21 bits of colour, assuming every pixel is a unique colour (i.e.that colour
appears in only one pixel of the image). You can't show more colours
than there are pixels on the screen.
> This is simply not the way that monitors work. If this were the case then
> even at 8 bit colour depth you would need 256 (what shall we call them?)
> colour dots(??) at each pixel to display the image.
No. You wouldn't. You would need one triad of red, green and blue
phosphors to make a single 'pixel'. That's how traditional CRT monitors
work. There are three electron guns, corresponding to each flavour of
screen element. According to the number of electrons hitting each of R,
G and B phosphors you can build a number of colours equal to R * G * B
values. For 24 bit, R, G and B each have 8 bits, or 256 levels. That's
around 16.8 million combinations...2^8 * 2^8 * 2^8 = 2^24.
> You are missing the point of my test. The point is that when reduced to 4
> bit (or even 8 bit) colour depth an image containing graduated colour will
> display as bands of that colour. I could as easily have chosen 48 bits, 27
> bits or any much higher value for the 'control'.
I understood the test, and it shows the point, but it's bad science, as the
'control' image is not being displayed as it is being stored, just as the
output
system (everything between the storage device and our ears)
in audio, shapes what comes off the CD/vinyl/whatever. There is an error
introduced in your analogy. In this case, the error actually lends itself
to making the point more obvious, not less, but it's error nonetheless.
For a typical monitor, you can show probably 0.8 million pixels, and
therefore 0.8 million colours, but the image contains over 20 times this
number. Reducing the 'bandwidth' (don't flame me for that) of the
colour spectrum to 256 colours from 0.8 million is a lot less lossy than
doing so from 16.8 million, and the affect is still very apparent - the
banding is visible. But in changing your display from 24 bits to
4 bits, you reconfigured the hardware, and changed the DAC setup,
among other things - you can't say 'I changed the bit depth,
but all else is equal'.
It's not possible to attribute this banding to simply the
reduction of bits-per-sample. There is too much other stuff going on
in the graphics chip for us to be certain that this is what's going on.
> If (as you argue) the graphics card applies signal shaping automaticaly
then
> this is akin to what I said already goes on in some audio DACs in order to
> improve their measured performance at the expense of their broad-range
audio
> performance.
Yes. But in this case, there is shaping *before* the DAC (actually, DACs for
graphics, as there need to be three, and they may not be identical).
For example, you mentioned 16 bit colour. Here, there are 5 bits for
Red, 5 for blue and 6 for green. To get to the point where we can feed
this type of signal to the DACs, we may dither and play with the
colour space, thereby changing the stored image before displaying it.
This is transparent, and independent of the software, but it affects
the output. Also, it isn't the same for all colour depths.
> > The 'control' for your experiment is no control at all, since it is
> > compromised at the display device. The display is not showing
> > all the colours in the file; it can't.
> Any more than you might argue that a loudspeaker can't reproduce 65,536
> amplitude values from a 16 bit encoded audio signal? Or - if it can,
perhaps
> it can reproduce 65,537 or 128,000 or 256,000 or ... where should we draw
> the line? Or is that question too close to the point of all this?
I think that last sentence hits the nail on the head. As a 'graphical
representation'
your analogy works - it shows the effect, in broad terms. However, it's not
really the truth. Quantitatively, the banding introduced doesn't really
represent the change in the stored image, it just tries to approximate it,
using
dither and various other techniques that can't be measured outside the
graphics
chip.
> But my experiment only requires a much reduced number of possible colours
to
> be displayed.
Granted. Unfortunately, with a graphics chip containing over 30 million
gates, there is no easy way to tell which of the many processes is really
causing the banding. Hence my assertion that the experiment is intuitive,
but not good science. By comparison, the critical path through an
audio system is relatively simple (though still horribly complex at the
lowest level, at least to me).
> My point is that when this number is greatly reduced the display *WILL*
> start to show colour bands where none existed in the original image.
But you couldn't really see the original image...not all of it, anyhow :-)
This is pedantry, and I didn't want to inflame the situation, honestly.
The bands certainly will exist, but as a graphics engineer I have to
say that I could not attribute that banding solely to reduction in bit
depth. There's more going on than that.
> If this is too much to grasp then do this instead.
>
> Load your favourite graphics program, open a blank image and flood fill it
> with a colour chosen from a 24 bit palette. Save the image as a 24 bit
file.
> Now set your monitor to "16 colours" and display the image again.
> Its displayed colour will have changed.
Not if I deliberately pick one of the 16 available colours for my
flood fill. It will remain identical. Okay, so it's a one in a million
choice, but
it's still there, and this very technique is used by design in graphics
applications
all the time. However, since most new graphics cards
use a look up table for all modes less than 16 bpp, this is irrelevant -
the card won't use the DAC in the same way when you do this in 16
or 256 colours. There is no 'palette' for 24 bit modes, just a combination
of R, G, and B voltage levels, each having up to 256 possible states. Each
combination gives a different colour. The monitor then shapes that colour
according to its state of tune, white point etc. In 256 colour mode the
card simply grabs a value from a table stored in hardware (the 'closest
match') and offers it to the monitor. Different technique, different result.
> What yo have just observed is the quantisation process at work.
Among other things. Not JUST the quantisation process, and that
is the crux.
> My point still applies. And the example I gave will display the bands I
> described.
Yes, it will display the bands, but not *solely* due to reduction in
bit depth. Your example, while offering a reasonable way to
envisage the process, neglects a great deal of other influencing
factors. I realise that your point is just to demonstrate quantisation
effects at work in an easy to understand way (commendable,
incidentally), but there are lots of other factors here. Had you
not used the word "exactly" in your original post, I wouldn't
have replied at all. I've used the example myself, but with
caveats, and never using the word 'exactly' to describe the
analogy - there are just too many unmeasurables for that.
Maybe I just need to lighten up. Usenet breeds pedantry.
Regards,
Glenn.
George Perfect
account.
I confess I was indeed going for a simple illustration. If my use of the
word "exactly" in some context has upset you then let me withdraw it as I'm
aiming for a simple life here.
While I am not dismissing all the factors you describe, I'm content that the
exercise demonstrates - in its banding - quantisation error in what I hope
is an easier to understand manner.
Your description of the sort of processing that goes on in and around
graphics DACs is useful in itself. One of the points I made (sooo long ago,
now) concerns the processing that goes on in and around audio DACs. In
effect, what we are both describing is the difference between theory and
practise. Theory suggests that (left to itself) a high bit depth image
displayed via a lower bit depth mode will band or colour shift. Ditto, an
audio waveform (if theory holds true) once sampled should be reproduced
consistently. That neither statement holds true in practise illustrates the
sometimes marked difference between theory and practise.
--
George
Newcastle, England
(please remove leading 'x' from email address to reply directly - thanks)
"Glenn Booth" <do...@removeqtlg.demon.co.uk> wrote in message
news:1017692163.3047.0...@news.demon.co.uk...
Glenn Booth
"George Perfect" <xgeo...@byline.co.uk> wrote in message
news:3ca9728e$0$8508$ed9e...@reading.news.pipex.net...
> Glenn - thanks for the dissertation. I hope you didn't write it on my
> account.
Well...actually, it was the result of a bad attitude arising from a
shitty Easter holiday. I needed to vent :-)
> I confess I was indeed going for a simple illustration. If my use of the
> word "exactly" in some context has upset you then let me withdraw it as
I'm
> aiming for a simple life here.
For that, it serves the purpose, and it does 'draw the picture' nicely. As I
said,
I've used similar analogies (with some modification) myself.
> While I am not dismissing all the factors you describe, I'm content that
the
> exercise demonstrates - in its banding - quantisation error in what I hope
> is an easier to understand manner.
Point taken. As I said, it was a commendable goal. I just had my
engineer's hat on for a while there. Getting paid to be pendatic is bad for
the social life. It also makes watching TV hell in these days of bad
MPEG, but that's another story.
> Your description of the sort of processing that goes on in and around
> graphics DACs is useful in itself. One of the points I made (sooo long
ago,
> now) concerns the processing that goes on in and around audio DACs. In
> effect, what we are both describing is the difference between theory and
> practise. Theory suggests that (left to itself) a high bit depth image
> displayed via a lower bit depth mode will band or colour shift. Ditto, an
> audio waveform (if theory holds true) once sampled should be reproduced
> consistently. That neither statement holds true in practise illustrates
the
> sometimes marked difference between theory and practise.
That works for me. In the end, both digital audio and digital video systems
are just our best efforts at getting back out whatever it was we tried to
put in.
Regards,
Glenn.
Arny Krueger
> "George Perfect" <xgeo...@byline.co.uk> wrote in message
> news:3ca9728e$0$8508$ed9e...@reading.news.pipex.net...
> > Your description of the sort of processing that goes on in and
around
> > graphics DACs is useful in itself. One of the points I made (sooo
long ago,
> > now) concerns the processing that goes on in and around audio
DACs. In
> > effect, what we are both describing is the difference between
theory and
> > practise.
Theory and practice are different, and I've got a web site that
documents that pretty thoroughly - www.pcavtech.com
> >Theory suggests that (left to itself) a high bit depth image
> > displayed via a lower bit depth mode will band or colour shift.
Not necessarily. Theory also says that there are ways to manage
resolution reduction in such a way that banding doesn't happen. One
methodology is called dithering.
Dithering is an inherent part of digital audio. For example the AES
standard for testing digital audio equipment says that with a few
exceptions, digitally-generated waves for testing digital audio gear
should always be generated with dither.
> >Ditto, an
> > audio waveform (if theory holds true) once sampled should be
reproduced
> > consistently.
Theory allows for the presence of all kinds of processing and
degradation. Only a highly incomplete and overly-simplified verison
of digital theory says that an audio waveform once sampled should be
reproduced consistently.
Since digital theory says that dither is an inherent part of digital
processing, and dither is basically an inconsistent signal whose
purpose is to create an inconsistent signal, digital theory
essentially states that once sampled, audio signals should be
reproduced inconsistently, not consistently as Mr. Perfect
erroneously suggests.
> >That neither statement holds true in practise illustrates the
> > sometimes marked difference between theory and practise.
No, both statements are true if you understand enough details about
the theory.
Chris G.
throughout my whole life generally things like the flue would last about 4-5
days. When I take echinaciea flues are reduced to 1-2 days or often if I
take them when symptoms first start appearing (like a sore throat or
achiness) it goes away before I even start getting a temperature. It's good
stuff as long as it's not abused. I imagine that it's probably not a good
idea to take it every day either.
What's really funny though is that scientific tests have shown prayer to
make an actual measurable difference on people who are ill. I don't know
the details of the study but I remember a while back that it was widely
talked about in the media.
In Chinese herbal medicines I think that there are mountains of stuff that
can be learned. The same for many indigenous forms of medicine practiced by
curanderos all over Latin America.... a reason why a lot of ethnomedical
research is being done in Latin America... some of it sponsored by large
pharmaceutical corporations who want to get their hands on that stuff.
There was also some studies of Raki practiced by registered nurses, but I
can't remember whether it was conclusive or not.
Anyways interesting stuff. I basically don't think that science holds all
their answers...still plenty of stuff science can't explain. No reason to
try alternative medical treatments out as long as they are relatively safe.
Chris G.
--
"Getting eaten by a crocodile is JUST like falling asleep... in a blender!"
- Homer Simpson
"Glenn Booth" <do...@removeqtlg.demon.co.uk> wrote in message
news:1017433530.25675....@news.demon.co.uk...
Jeez
<chri...@txdirect.net> wrote:
> What's really funny though is that scientific tests have shown prayer to
> make an actual measurable difference on people who are ill.
"Here, take this placebo - you'll feel better right away..."
Forever,
Kim.
--
My only love sprung from my only hate!
Too early seen unknown, and known too late!
-- William Shakespeare, "Romeo and Juliet"