Louder IS Better (With Lossy)
Lord Hasenpfeffer
>>
>> I like watching a good train wreck.....
> YES! ROTFLOL! On second thought, my tolerance for gore may not be
> up to the spectacle. :-)
Laugh it up.
I spent time today browsing websites and reading FAQs, etc. to find out
more information about the Psycho-Acoustic Model employed by MP3. It
is indeed similar to MiniDisc's ATRAC compression scheme in that it
removes not only "masked frequencies" but also frequencies which
are determined to be "too quiet to be heard" by common human ears.
This proves to me that MP3s encoded from "quieter WAVs" which have
been ripped from CDs mastered with low, average amplitudes will suffer
more at the hands of a lossy audio data compression algorithm than
will MP3s encoded from WAVs which have first been "appropriately
normalized" (i.e. Linux: $normalize -ba -10dBFS ) prior to being encoded.
While my MFSL CD of Pink Floyd's, "Dark Side Of The Moon" may be
perfectly fine for playback directly from that CD it is inappropriate
for being encoded to MP3 because its amplitudes are too low to drive
many of its frequencies above their Absolute Thresholds of Hearing given
its existing, low-amp state.
"Batch normalizing" the WAVs from a "quiet CD", however, will raise
their _collective_ average amplitude to a significantly preferable level
capable of yielding better sounding MP3s than would otherwise be
achieved since fewer of their frequencies will be subject to being
discarded by the encoding process.
If applied sensibly and with appropriate grace for the sake of
preserving the original dynamics of the recording, such "batch
normalizing" will *not* bring clipping, limiting, compressing or adverse
dynamic-range-related effects to the WAVs in question.
The same holds true for MiniDisc recordings as well since the ATRAC
compression method too seeks to remove frequencies which are deemed "too
quiet to be heard".
Aggregate: Lower amplitudes cause higher frequency loss, therefore,...
...louder *is* better (with lossy).
Myke
--
-================================-
Windows...It's rebootylicious!!!
-================================-
John LeBlanc
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
news:3EFE8232...@spamsucks.ionet.net...
> ...louder *is* better (with lossy).
Seems to me that's like saying horse shit tastes better than dog shit.
John
Lord Hasenpfeffer
Hmmm... For some reason, I thought I'd posted to rec.audio.PRO.
I can get replies like that from the little kid next door.
Mark T. Wieczorek
> ...louder *is* better (with lossy).
Why not do a side-by-side comparison? All the theories in the world won't
change reality.
Regards,
Mark
--
http://www.marktaw.com/
http://www.prosoundreview.com/
User reviews of pro audio gear
Lord Hasenpfeffer
> Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote in
> news:3EFE8232...@spamsucks.ionet.net:
>
>>...louder *is* better (with lossy).
>
> Why not do a side-by-side comparison? All the theories in the world won't
> change reality.
Oh, I do all the time, believe me. My hypothesis as stated in my
original post is more than verified by my own personal experience,
however, I'm merely a "seasoned amateur" at boosting the loudnesses of
my WAVs before encoding them to MP3s and was curious to know if there
might be more I could learn about this from a few pros in the field.
I get the impression, however, that most "pros" dismiss MP3 as unworthy
of their time and attention, carte-blanche, therefore I'm uncertain as
to how much experience-based knowledge can be derived from such a forum.
Nevertheless, I was challenged to post my hypothesis here by a couple of
MP3-inexperienced buttheads(?) in another NG, so I did. I believe my
hypothesis is quite sound for the purpose of obtaining better results
with lossy MP3 encoding - although with _uncompressed_ CD/WAV audio,
merely increasing the amplitude does nothing to improve the quality of
the sound - even though Joe Sixpack might *think* it does when he hears
them - which is really beside the point as far as I'm concerned. :)
Does anybody in this NG have experience with attempting to create
higher-quality MP3s or do they all just rip-and-encode like everyone
else, taking whatever they can get from _unmodified_ WAVs extracted via DAE?
Artie Turner
> Does anybody in this NG have experience with attempting to create
> higher-quality MP3s or do they all just rip-and-encode like everyone
> else, taking whatever they can get from _unmodified_ WAVs extracted via
> DAE?
All MP3 encoders are not created equal. You might find that some
encoders sound better than others for a given .wav., but generally
speaking, louder is better regardless of what's being evaluated.
When I make MP3s of my material to send via email or website, I just use
the program material "as is." I don't recall any serious evaluations of
MP3 encoders or encoding levels here on RAP.
AT
Mark T. Wieczorek
Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote in
news:3EFF1A45...@spamsucks.ionet.net:
> Mark T. Wieczorek wrote:
>> Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote in
>> news:3EFE8232...@spamsucks.ionet.net:
>>
>>>...louder *is* better (with lossy).
>>
>> Why not do a side-by-side comparison? All the theories in the world
>> won't change reality.
>
> Oh, I do all the time, believe me. My hypothesis as stated in my
> original post is more than verified by my own personal experience,
> however, I'm merely a "seasoned amateur" at boosting the loudnesses of
> my WAVs before encoding them to MP3s and was curious to know if there
> might be more I could learn about this from a few pros in the field.
Unless you're increadibly knowledgable, I don't think the "pro's" encode
MP3's any different from the mere mortal... This isn't really a thing where
you need to know where to stick a mic or dial in a compressor. Nor is it a
thing where having more expensive equipment helps out out too much, though
I think the frounhoufer codec is something you have to pay for.
> I get the impression, however, that most "pros" dismiss MP3 as
> unworthy of their time and attention, carte-blanche, therefore I'm
> uncertain as to how much experience-based knowledge can be derived
> from such a forum.
Yes and no... If you're recording someone professionally, your job ends
when they have a CD in their hands and you go on to the next client. If
your client is a total luddite, you may encode the mp3's for them, or maybe
you do web hosting as part of your studio package (I guess partnering with
a design house) but I don't think many studios do that.
> Nevertheless, I was challenged to post my hypothesis here by a couple
> of MP3-inexperienced buttheads(?) in another NG, so I did. I believe
> my hypothesis is quite sound for the purpose of obtaining better
> results with lossy MP3 encoding - although with _uncompressed_ CD/WAV
> audio, merely increasing the amplitude does nothing to improve the
> quality of the sound - even though Joe Sixpack might *think* it does
> when he hears them - which is really beside the point as far as I'm
> concerned. :)
Ah... Dared to post to RAP eh? I bet the response here was little different
than the response in the other groups... just more authoritative sounding.
If you really want to test your hypothesis, you could do an extreme example
in a blind A/B comparison. take a .wav file and copy it to a 2nd file and
reduce the volume quit a bit (-30dB to -60dB maybe). Encode both to mp3,
then bring them back to .wav. Then bring the volume on the 2nd file back up
as much as you cut it.
You could post your results.
> Does anybody in this NG have experience with attempting to create
> higher-quality MP3s or do they all just rip-and-encode like everyone
> else, taking whatever they can get from _unmodified_ WAVs extracted
> via DAE?
I know someone who's really particular about his sound, really sort of anal
about it, does everything absolutely the best possible way he can afford,
does A/B tests all the time (and bugs the shit out of everyone he knows
with them too).
I don't know if he's looked at mp3 encoders yet - I think only briefly, but
when he does, you can be sure that he'll find the absolute best path to the
absolute best sounding mp3 he can make.
I'm the exact opposite. I have a "It's good enough for rock n' roll"
attitude. I use CDex because it's fast and easy, and sound pretty good,
even though I can always tell I'm listening to an mp3. Life's too short to
do too many A/B comparison tests.
Anyway, if you're interested in talking to this guy he hangs out at
prosoundreview.com, mostly in the computer tweaks section. I'm sure he'd be
very interested in talking to you.
> Myke
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3EFF0D3A...@spamsucks.ionet.net...
> John LeBlanc wrote:
> > "Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> > news:3EFE8232...@spamsucks.ionet.net...
> >
> >
> >>...louder *is* better (with lossy).
> >
> >
> > Seems to me that's like saying horse shit tastes better than dog shit.
>
> Hmmm... For some reason, I thought I'd posted to rec.audio.PRO.
>
> I can get replies like that from the little kid next door.
Smart kid.
Did he tell you that REC stands for "recreational" (like this post) and
that PRO stands for PROduction? There aren't too many people here
who's record production habits revolve around MP3... it's a byproduct.
--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com
flint
1. Choose a WAV.
2. Make a copy of the WAV.
3. Reduce the level of the 2nd WAV by 10dB.
4. Compress both WAVs with your MP3 codec.
5. Convert both MP3s back to WAV.
6. Increase the level of the adjusted WAV by 10dB
7. Compare the two with listening test.
8. Validate comparison with other listeners.
If you can hear a difference, overall level does afect quality.
-FLINT
Lord Hasenpfeffer
> I really liked the suggestion that you actually test this hypothesis.
Yes, me too. However, only since taking this to rec.audio.pro have I
encountered this form of "professional" advice. I certainly didn't get
any suggestions as good as this in the other newsgroups.
> 1. Choose a WAV.
> 2. Make a copy of the WAV.
> 3. Reduce the level of the 2nd WAV by 10dB.
> 4. Compress both WAVs with your MP3 codec.
> 5. Convert both MP3s back to WAV.
> 6. Increase the level of the adjusted WAV by 10dB
> 7. Compare the two with listening test.
> 8. Validate comparison with other listeners.
This is essentially the same test which David Morgan already suggested -
except his suggested, initial dB reduction values for the 2nd WAV were
much greater.
This is also the type of response and advice I've been seeking all
along. Thought I'd get something similar over in the other newsgroups,
but did not.
> If you can hear a difference, overall level does afect quality.
Agreed. And I definitely do expect to hear a difference - because the
effects of the psychoacoustic frequency/amplitude-based lossy
compression algorithms are real and not a figment of my imagination.
With that in mind, I suppose the real question is, how much of an
initial level reduction is needed before loss of fidelity is obvious.
-4dB? -10dB? -30dB? -60dB? I guess we'll see!
Suggestions such as these hardly equate with train wrecks and gore, so I
don't know what those boneheads(?) in the other newsgroup(s) were even
thinking when they said that. Based upon all but the first reply it
received, I'm *glad* to have posted my hypothesis in this newsgroup.
I hope things in here can remain as cordial as they have been so far.
Lord Hasenpfeffer
> Did he tell you that REC stands for "recreational" (like this post) and
> that PRO stands for PROduction? There aren't too many people here
> who's record production habits revolve around MP3... it's a byproduct.
Um, actually nobody told me anything about the meanings of the name...
Was I was stupid (again) to ASSuME that "rec.audio.pro" stands for
"recording.audio.professional" instead of
"recreational.audio.production"? :)
Myke
P.S. Thanks for your WAV test suggestions. That's exactly the kind of
advice I've been seeking as opposed to the stupid emotional outbursts
like "troll somewhere else you f**kwitted sockpuppet" blah blah blah.
Lord Hasenpfeffer
Lord Hasenpfeffer
> Lord Hasenpfeffer wrote:
>
>> Does anybody in this NG have experience with attempting to create
>> higher-quality MP3s or do they all just rip-and-encode like everyone
>> else, taking whatever they can get from _unmodified_ WAVs extracted
>> via DAE?
>
> All MP3 encoders are not created equal. You might find that some
> encoders sound better than others for a given .wav.,
I use "notlame" with Linux for all of my MP3 encoding. Last I knew,
several tests conducted by some so-and-so's-who're-in-the-know concluded
that "notlame" actually *beats* fraunhofer at its own game. From a
technical standpoint I'm in no real position to confirm or deny that
information. It's just some "crowing" that I read at the notlame
website about a year ago. Regardless of what tests may be being
conducted, however, I definitely get excellent results with notlame,
have never been disappointed by notlame, and have yet to find the
slightest reason to even consider working with any other encoder... but
others' mileages may vary.
> generally speaking, louder is better regardless of what's being evaluated.
Oh really? Well, I used to think so too (and kinda still do no matter
what the anal-techs would prefer that I believe) although such a belief
is obviously grounds for receiving a hefty spanking over in the other
newsgroups I've been contending with lately. I'm actually kinda
surprised to find anyone in here that' brave enough to say what you just
did! :)
> When I make MP3s of my material to send via email or website, I just use
> the program material "as is."
That makes sense, as I usually do the same. However, for the MP3s I'm
making today which I'd like to still like to be enjoying 5 years from
now, it's a slightly different story.
> I don't recall any serious evaluations of
> MP3 encoders or encoding levels here on RAP.
Would one be considered off-topic in this newsgroup?
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3EFE8232...@spamsucks.ionet.net...
> >>
> >> I like watching a good train wreck.....
>
> > YES! ROTFLOL! On second thought, my tolerance for gore may not be
> > up to the spectacle. :-)
>
>
> Laugh it up.
That's a RECreational way to start. :-)
> I spent time today browsing websites and reading FAQs, etc. to find out
> more information about the Psycho-Acoustic Model employed by MP3.
Well heck... so did I... just for something to do. (train wreck, step 1)
MP3 and MP3 codecs are not required for making standard audio recordings
or commercial CD’s.
> It is indeed similar to MiniDisc's ATRAC compression scheme in that it
> removes not only "masked frequencies" but also frequencies which
> are determined to be "too quiet to be heard" by common human ears.
And which coder / decoder did you study? There are many. Only the actual
developer's codec is considered toterable by most - it is not free and does
not come bundled with software packages. http://www.iis.fraunhofer.de/amm
> This proves to me that MP3s encoded from "quieter WAVs" which have
> been ripped from CDs mastered with low, average amplitudes will suffer
> more at the hands of a lossy audio data compression algorithm than
> will MP3s encoded from WAVs which have first been "appropriately
> normalized" (i.e. Linux: $normalize -ba -10dBFS ) prior to being encoded.
Depending on your initial source material, "normalization", as such, may
do absolutely *nothing* to your audio file.
Would you mind defining, "appropriately normalized." (?)
> While my MFSL CD of Pink Floyd's, "Dark Side Of The Moon" may be
> perfectly fine for playback directly from that CD it is inappropriate
> for being encoded to MP3 because its amplitudes are too low to drive
> many of its frequencies above their Absolute Thresholds of Hearing given
> its existing, low-amp state.
(train wreck, step 2)
> "Batch normalizing" the WAVs from a "quiet CD", however, will raise
> their _collective_ average amplitude <split>
(train wreck, step 3)
Whoops. "Normalization" alone is based on peak amplitude - hence my
earlier statement. You are now referring to "Average amplitude", a much
different beast entirely. If you are "normalizing" to an "average" RMS level,
you are merely wreaking havok on the peak portions of your program that
exceed your specified RMS level as compared to 0dBFS. (I think someone
already mentioned the dogshit results... your neighbor's kid, was it?)
> <end split> to a significantly preferable level
> capable of yielding better sounding MP3s than would otherwise be
> achieved since fewer of their frequencies will be subject to being
> discarded by the encoding process.
I would venture to say that by using the method you describe (that is,
RMS Normalization), you have destroyed significantly important audible
portions of extremely usable audio, in favor of saving what may be near
inaudible (and by most codecs = 'diposable') horseshit. Congratulations.
I believe the phrase was something like, you believe that your "Horseshit"
sounds better than plain old MP3 encoded "Dogshit".
http://www.mp3-converter.com/mp3codec/mp3_anatomy.htm might
explain why the MP3 suffers, although it is an acceptable method of
transferring data within a confined bandwidth.
Discussing the 'pros' and 'cons' of your suggested approach would be
better suited to this forum.
> If applied sensibly and with appropriate grace for the sake of
> preserving the original dynamics of the recording, such "batch
> normalizing" will *not* bring clipping, limiting, compressing or adverse
> dynamic-range-related effects to the WAVs in question.
(train wreck, step 4)
Totally Impossible.... at least in the manner which you describe.
> The same holds true for MiniDisc recordings as well since the ATRAC
> compression method too seeks to remove frequencies which are deemed
> "too quiet to be heard".
I don't think so. A variety of tools might need to be, or could be, applied
to the soundfile before "normalization" ever enters the picture - - but all
of these tools *DO* bring clipping, limiting, and compression into play
and most definitely create "adverse dynamic-range-related effects".
> Aggregate: Lower amplitudes cause higher frequency loss, therefore,...
>
> ...louder *is* better (with lossy).
(train wreck, step 5)
In probably 95% of all cases, "louder" will be almost *always* deemed as
"better" by the average listener as a perceived increased in 'clarity' occurs.
This is a natural, psychoacoustical phenomenon which occurs at different
levels with different listeners. Hopefully somewhere below the threshold of
pain.
Geoff Wood
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> Nevertheless, I was challenged to post my hypothesis here by a couple of
> MP3-inexperienced buttheads(?) in another NG, so I did. I believe my
> hypothesis is quite sound for the purpose of obtaining better results
Thanks for the compliment.
For those who don't know, My Myke here insists on normalising every piece of
his 2100-strong CD collection to -10dB average RMS, and considers any tracks
that don't meet this criteria to be flawed, and incompetently produced.
Despite serious attempts to clue him up he clings to total misconceptions
regarding levels, amplification, attenuation, normalisation, the mastering
process, etc.
He dismisses MFSLs Dark Side Of The Moon as being a peice of excrement
because the highest peak is -4dB or so, and that buyers have been ripped
off. (they didn't get all the bits they paid for ?).
geoff (an mp3-inexperienced butthead, evidently)
Artie Turner
> Artie Turner wrote:
and have yet to find the
> slightest reason to even consider working with any other encoder... but
> others' mileages may vary.
If you're really that interested in MP3s, it seems to me you'd try some
other encoders just out of curiosity.
>
>> generally speaking, louder is better regardless of what's being
>> evaluated.
>
>
> Oh really? Well, I used to think so too (and kinda still do no matter
> what the anal-techs would prefer that I believe) although such a belief
> is obviously grounds for receiving a hefty spanking over in the other
> newsgroups I've been contending with lately. I'm actually kinda
> surprised to find anyone in here that' brave enough to say what you just
> did! :)
Sounds like you stumbled into rec.audio.opinion. Don't drop your soap.
I'm guessing that everyone in here would probably agree in general with
the "louder is better" truism , but most people here don't give much
thought to MP3s unless it's with regard to the P2P/Napster/IP quagmire.
>
>> When I make MP3s of my material to send via email or website, I just
>> use the program material "as is."
>
>
> That makes sense, as I usually do the same. However, for the MP3s I'm
> making today which I'd like to still like to be enjoying 5 years from
> now, it's a slightly different story.
Well, I doubt your hearing acuity will improve, so unless you're sick of
the tune by that time, I see no reason you'd stop enjoying it if you
like it now.
>
>> I don't recall any serious evaluations of MP3 encoders or encoding
>> levels here on RAP.
>
>
> Would one be considered off-topic in this newsgroup?
Not when you consider some of the other threads we've beaten into dog
food here, but again, I doubt if many people here care to quibble over
the fidelity of this MP3 vs that one, *unless* you pirated it from their
CD, in which case that MP3 will be indistinguishable from the source
regardless of how it was encoded.
AT
You know, after further consideration, this sounds like a job for Arny
Kruger and Double Blind Boys. Take it away, Arny...
>
> Myke
>
Mike Rivers
> I get the impression, however, that most "pros" dismiss MP3 as unworthy
> of their time and attention, carte-blanche, therefore I'm uncertain as
> to how much experience-based knowledge can be derived from such a forum.
Oh, there are plenty of "pro" applications for MP3 files, it's just
that evaluating the technical quality of a microphone or A/D converter
isn't one of them. And sometimes we get preoccupied with degradations
of sound quality and forget that not everyone who hears our recordings
will have as good a playback and listening environment as we do (and
we try not to say anything about those who will have a BETTER one).
Professional musicians working on songs and arrangements often
exchange MP3 files (often of MIDI synths). Studios often offer MP3s of
mixes or edits for client approval in the same spirit that we used to
give them a cassette. And I'll bet a few of us even have MP3 jukeboxes
on our computers for home listening and party music.
> Does anybody in this NG have experience with attempting to create
> higher-quality MP3s or do they all just rip-and-encode like everyone
> else, taking whatever they can get from _unmodified_ WAVs extracted via DAE?
Not me. I just let 'er rip. When loading up my Jukebox 3 for travel, I
even commit the unpardonable sin of using its own internal encoder
(which I've been told isn't so hot) rather than rip and encode a CD on
my computer and then transfer it to the Jukebox. I use it because it's
simple and I choose not to complicate it for the sake of a slight
improvement in fidelity that will be lost on the plane or in the car
anyway.
--
I'm really Mike Rivers - (mri...@d-and-d.com)
Lord Hasenpfeffer
> MP3 and MP3 codecs are not required for making standard audio recordings
> or commercial CD’s.
Yes, and therein is where I believe I made my first mistake; attempting
to discuss lossy-related-matters with a lossless-only crowd! :) The
effect of increasing amplitude in an attempt to obtain "better sound"
with uncompressed audio data is negligible. However, the effect of
increasing the average level of a WAV prior to "munging" it with an
amp/freq-based lossy compression algorithm such as ATRAC or MP3 seems at
a glance to me to be rather substantial... and until I know better, I am
taking a "true until proven false" approach with all my MP3 encoding
practices.
> And which coder / decoder did you study? There are many. Only the actual
> developer's codec is considered toterable by most - it is not free and does
> not come bundled with software packages. http://www.iis.fraunhofer.de/amm
As I have stated in another post within this thread which you may or may
not have read already, (I'll repost it here for your convenience):
I use "notlame" with Linux for all of my MP3 encoding. Last I knew,
several tests conducted by some so-and-so's-who're-in-the-know concluded
that "notlame" actually *beats* fraunhofer at its own game. From a
technical standpoint I'm in no real position to confirm or deny that
information. It's just some "crowing" that I read at the notlame
website about a year ago. Regardless of what tests may be being
conducted, however, I definitely get excellent results with notlame,
have never been disappointed by notlame, and have yet to find the
slightest reason to even consider working with any other encoder... but
others' mileages may vary.
> Depending on your initial source material, "normalization", as such, may
> do absolutely *nothing* to your audio file.
To the quality of the source audio WAV, yes, it may indeed do nothing,
however, if "normalization", as such, helps to boost more of the
original WAV's frequencies into a more audible range, surely that would
cause those frequencies to be preserved rather than discarded by a
psycho-acoustic lossy compression scheme, no?
> Would you mind defining, "appropriately normalized." (?)
Yes, and thank you for asking before making invalid assumptions that I'm
a neanderthal about such matters like those in the other newsgroup were
prone to doing.
I've posted some screenshots of various WAVs to my website which, I
believe, can help to explain this as well, but first a little
textplanation is in order.
I am a Linux user. For Linux there is a command-line application called
"normalize" which is currently installed on my machine. Despite its
name, this application can do more than simply normalize a WAV according
to the strict, textbook definition of the term. So, right there, a
stumbling block exists when it comes to discussing the use of
"normalize" with people who are not familiar with it.
As I understand it, the technical definition of normalization indicates
that it is a process of simply boosting the amplitude of a digital audio
recording by the amount necessary to cause its loudest peak to reach
"Full Scale". Nothing more, nothing less.
At this point it is important for you to realize as a participant in
this discussion that while the Linux application called "normalize" can
indeed readily perform the task of normalization according to its strict
textbook definition, it can and will do more (e.g. employ limiting to
avoid clipping) if given user-supplied instructions which would require
it to do so.
> Whoops. "Normalization" alone is based on peak amplitude - hence my
> earlier statement. You are now referring to "Average amplitude", a much
> different beast entirely.
Correct. I still refer to it as "normalization" though (note the
quotes), because (1) that is the name of the application and (2) I don't
know what else to call it. I am not a highly trained professional in
this field. I am merely a seasoned amateur who sometimes uses terms
which I *think* I understand when I actually do not. This can and has
led to some unfortunate errors of miscommunication in recent days
between me and some who are more familiar with the terms and processes
with which I am fiddling around. The typical response this garners is
usually more along the lines of "go read a book you stupid jackass"
rather than something that will actually assist me sooner to understand
the correct definitions of the terms for the sake of advancing of the
discussion that is already underway. Anyway...
> If you are "normalizing" to an "average" RMS level, you are merely
> wreaking havok on the peak portions of your program that exceed your
> specified RMS level as compared to 0dBFS.
In theory, this would seem to be the case, yes, however, in practice, I
cannot visually discern in my WAV editor where the sins I'm accused of
committing are occurring as a result of having used "normalize" to boost
the amplitude of a WAV. In addition, I have posted to my website
several screenshots which visually demonstrate the effects of using this
application. Some who see these images say, "Oh, OK, that's different.
You aren't doing what I thought you were doing and no harm has been
done." while others who claim to have seen the images say, "You bastard!
You've mangled the original WAV beyond all recognition!". Naturally,
such polar opposite reactions to my images leave me more confused than
anything else.
> (I think someone already mentioned the dogshit results...
> your neighbor's kid, was it?)
No. I told that guy in effect that I'd come here seeking a more
professional analysis of my hypothesis, but instead, the first reply I
received from this newsgroup turned out to be some stupid emotional
"analogy" which I could have received just as well from the little kid
next door.
> I would venture to say that by using the method you describe (that is,
> RMS Normalization), you have destroyed significantly important audible
> portions of extremely usable audio, in favor of saving what may be near
> inaudible (and by most codecs = 'diposable') horseshit. Congratulations.
OK, I think I see the misunderstanding I've caused (again) by describing
it the way I did in my original post - and yes, if I had been talking
about pumping the average RMS level up to Full Scale - gee whiz! - that
*would* be a disaster! :-D Fortunately, that's not at all what is
happening.
To better illustrate, let's say we have a WAV with a maximum peak
amplitude of -6dBFS. Using "normalize" to normalize this WAV, a simple
boost of +6dB applied to its average amplitude (RMS?) would cause the
whole recording to become 6dB louder and send that maximum peak right up
to Full Scale. Correct me if I'm wrong about that.
"Batch normalize" is the term used by the author of "normalize" to
describe the exact same process being applied to an entire set of WAVs
rather than to just one WAV. If a set of WAVs is simply "normalized"
they are all made to sound "equally loud". However, if a set of WAVs
are "batch normalized", the loudnesses of all the files in question are
boosted by a single common value, in effect treating the whole set as if
it were a single WAV. "Batch normalizing" is essential for preserving
the original _relative loudnesses_ of the full set of WAVs as they
originally were in relation to each other.
In short, when creating a "mix" of songs from various sources, simple
"normalization" of all the files would help to bring them all to a
fairly uniform loudness which would save the listener from constantly
having to manually adjust his volume control to compensate for the
varying loudnesses carried forth from the various original sources.
Meanwhile, "batch normalizing" is something that would be useful for
processing a set of WAVs that are related to each other from having been
derived from a single, common original source.
> Discussing the 'pros' and 'cons' of your suggested approach would be
> better suited to this forum.
So by posting in rec.audio.pro I've actually come to the right place?
> I don't think so. A variety of tools might need to be, or could be, applied
> to the soundfile before "normalization" ever enters the picture - - but all
> of these tools *DO* bring clipping, limiting, and compression into play
> and most definitely create "adverse dynamic-range-related effects".
I think I'll hold off on this point for now until I know that we're on
the same page with all that has gone before so far.
> In probably 95% of all cases, "louder" will be almost *always* deemed as
> "better" by the average listener as a perceived increased in 'clarity' occurs.
> This is a natural, psychoacoustical phenomenon which occurs at different
> levels with different listeners.
Precisely. I believe you 100%. But for some reason, the tech-heads
over where I was before sure don't seem to know about it. While they
readily admit to the existence of psycho-acoustic philosophies, to me
they seem completely oblivious to any knowledge of really what it's all
about or how it works. (Probably because in their non-MP3, CD-only,
perfect world, they never have to deal with it.)
> Hopefully somewhere below the threshold of pain.
Indeed. :-)
Lord Hasenpfeffer
train-wreck here because of this it will be you not me. You are mud in
these waters and have proven yourself to be so on more occasions than I
prefer to count.
Lord Hasenpfeffer
>> Nevertheless, I was challenged to post my hypothesis here by a couple of
>> MP3-inexperienced buttheads(?) in another NG, so I did. I believe my
>> hypothesis is quite sound for the purpose of obtaining better results
> Thanks for the compliment.
At least I was kind enough to stick a question mark in there afterwards
when I knew you weren't yet looking. Now, however, as far as you are
concerned, I am removing that question mark.
> For those who don't know, My Myke here insists on normalising every piece of
> his 2100-strong CD collection to -10dB average RMS, and considers any tracks
> that don't meet this criteria to be flawed, and incompetently produced.
> Despite serious attempts to clue him up he clings to total misconceptions
> regarding levels, amplification, attenuation, normalisation, the mastering
> process, etc.
And you dare to call me a troll! You are grossly misrepresenting both
me and my position because you do not understand either me or my
position. Go back to the goo from whence you came you friggin' TROLL!
> geoff (an mp3-inexperienced butthead, evidently)
Yes, that's what you are.
John LeBlanc
> John LeBlanc wrote:
> > "Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> > news:3EFE8232...@spamsucks.ionet.net...
> >
> >
> >>...louder *is* better (with lossy).
> >
> >
> > Seems to me that's like saying horse shit tastes better than dog shit.
>
> Hmmm... For some reason, I thought I'd posted to rec.audio.PRO.
It is rec.audio.pro
> I can get replies like that from the little kid next door.
Oh, he's interested in professional audio, too? Why come here, then, when
you have such a resource nearby? <g>
Since my response seemed be a bit enigmatic to you, I'll elucidate.
You already know that MP3 encoding -- regardless the encoder used -- is a
lossy format. So discussing the merits of one encoding technique in
comparison to another is simply discussing which encoding scheme screws up
the sound less than another. Much like comparing the relative tastes of one
form of shit to another; it may be a better tasting shit, but it's still
shit.
Lossy compression schemes certainly have their usefullness, but MP3 encoding
it isn't a "quality audio format" in my book.
Clearly, YMMV.
John
John LeBlanc
"Mike Rivers" <mri...@d-and-d.com> wrote in message
news:znr1056913819k@trad...
> Not me. I just let 'er rip. When loading up my Jukebox 3 for travel, I
> even commit the unpardonable sin of using its own internal encoder
> (which I've been told isn't so hot) rather than rip and encode a CD on
> my computer and then transfer it to the Jukebox. I use it because it's
> simple and I choose not to complicate it for the sake of a slight
> improvement in fidelity that will be lost on the plane or in the car
> anyway.
I use my Jukebox in the same way, and for pretty much the same reason. It's
adequate for a rip-and-run, which will just be deleted next round anyway.
John
Bob Olhsson
<my...@spamsucks.ionet.net> wrote:
>
>I spent time today browsing websites and reading FAQs, etc. to find out
>more information about the Psycho-Acoustic Model employed by MP3. It
>is indeed similar to MiniDisc's ATRAC compression scheme in that it
>removes not only "masked frequencies" but also frequencies which
>are determined to be "too quiet to be heard" by common human ears.
You are making a very fundamental mistake in your conclusion. But don't
take my word for it, just try different levels and listen.
--
Bob Olhsson Audio Mastery Recording Project Design and Consulting
Box 90412, Nashville TN 37209 Tracking, Mixing, Mastering, Audio for Picture
615.385.8051 FAX: 615.385.8196 Mix Evaluation and Quality Control
40 years of making people sound better than they ever imagined!
Lord Hasenpfeffer
> Lossy compression schemes certainly have their usefullness, but MP3 encoding
> it isn't a "quality audio format" in my book.
We are not discussing the merits of MP3 as a "quality audio format" in
this thread, therefore, your input to that end is, completely irrelevant
to this discussion. All attempts on your behalf to convert it into
something that it is not are completely misguided and entirely unwanted.
If you would prefer to discuss that topic in this newsgroup, start a
new thread of your own about and get it out of mine.
> discussing the merits of one encoding technique in comparison to
> another is simply discussing which encoding scheme screws up the
> sound less than another.
We are not discussing whether the merits of different encoding
techniques in this thread, therefore, your input to that end is,
completely irrelevant to this discussion. All attempts on your behalf
to convert it into something that it is not are completely misguided and
entirely unwanted. If you would prefer to discuss that topic in this
newsgroup, start a new thread of your own about and get it out of mine.
Lord Hasenpfeffer
> I use my Jukebox in the same way, and for pretty much the same reason. It's
> adequate for a rip-and-run, which will just be deleted next round anyway.
Therefore, you obviously have nothing constructive to contribute to this
thread. Please take your pollution elsewhere and stop diluting this thread.
Lord Hasenpfeffer
>> I spent time today browsing websites and reading FAQs, etc. to find out
>> more information about the Psycho-Acoustic Model employed by MP3. It
>> is indeed similar to MiniDisc's ATRAC compression scheme in that it
>> removes not only "masked frequencies" but also frequencies which
>> are determined to be "too quiet to be heard" by common human ears.
> You are making a very fundamental mistake in your conclusion. But don't
> take my word for it, just try different levels and listen.
The sky is a certain colour. But don't take my word for it, just go
outside and look up.
In other words, why bother telling me that I am making a "very
fundamental mistake" if you aren't also willing to cite what it is?
Lord Hasenpfeffer
John LeBlanc wrote:
> Lossy compression schemes certainly have their usefullness, but MP3
> encoding it isn't a "quality audio format" in my book.
We are not discussing the merits of MP3 as a "quality audio format" in
this thread, therefore, your input to that end is completely irrelevant
to this discussion. All attempts on your behalf to convert it into
something that it isn't are completely misguided and entirely unwanted.
If you would prefer to discuss that topic in this newsgroup, start a
new thread of your own about it and get out of mine.
> discussing the merits of one encoding technique in comparison to
> another is simply discussing which encoding scheme screws up the
> sound less than another.
We are not discussing the merits of different encoding techniques in
relation to each other in this thread, therefore, your input to that end
is completely irrelevant to this discussion. All attempts on your
behalf to convert it into something that it isn't are completely
misguided and entirely unwanted. If you would prefer to discuss that
topic in this newsgroup, start a new thread of your own about it and get
out of mine.
dmi...@spamblock.demon.co.uk
> >> why not add rec.audio.pro to the list and see what they think?
> >>
> >> I like watching a good train wreck.....
> > YES! ROTFLOL! On second thought, my tolerance for gore may not be
> > up to the spectacle. :-)
> Laugh it up.
> I spent time today browsing websites and reading FAQs, etc. to find out
> more information about the Psycho-Acoustic Model employed by MP3. It
> is indeed similar to MiniDisc's ATRAC compression scheme in that it
> removes not only "masked frequencies" but also frequencies which
> are determined to be "too quiet to be heard" by common human ears.
> This proves to me that MP3s encoded from "quieter WAVs" which have
> been ripped from CDs mastered with low, average amplitudes will suffer
> more at the hands of a lossy audio data compression algorithm than
> will MP3s encoded from WAVs which have first been "appropriately
> normalized" (i.e. Linux: $normalize -ba -10dBFS ) prior to being encoded.
> While my MFSL CD of Pink Floyd's, "Dark Side Of The Moon" may be
> perfectly fine for playback directly from that CD it is inappropriate
> for being encoded to MP3 because its amplitudes are too low to drive
> many of its frequencies above their Absolute Thresholds of Hearing given
> its existing, low-amp state.
I would be somewhat surprised that any current CD was not NORMALISED to
somewhere around (just below) 0dbfs. Note that this says nothing about
average level, and where something intended to be listened to as a whole
is split into tracks may say nothing about individual track levels.
All the Normalising does is to ensure that the sample with the highest
magnetude over the range being normalised (track of whole disk) is at the
relevant level.
This is (almost) totally a non contentious thing to do when mastering a disk,
the place where judgement comes into it is when limiting and compression are
applied to raise the RMS level closer to the normalised level.
Further the phycoacoustic model IIRC derives 'too quiet' threshold from
the RMS level of the signal, not an absolute threshold (IE at some
frequency it may be 30db below curent RMS level, NOT 40db below 0dbFS).
It is worth noting that whatever you do to the input data, it is only
possible to fit a fixed amount of information into the output stream,
if you force the encoder to include more frequency data (higher resolution
or more bands active, then time or ampletude resolution MUST necassarily
suffer).
Regards, Dan.
--
** The email address *IS* valid, do NOT remove the spamblock
And on the evening of the first day the lord said...........
.... LX 1, GO!; and there was light.
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message ...
> So by posting in rec.audio.pro I've actually come to the right place?
That depends on whether or not you are here to state that you are
without flaw in any of your reasoning and will never change your
mind or approach in any way, regardless...... or, if you are in search
of some facts regarding the methodology you say you employ.
It would appear thus far, you are merely espousing to an opinion
which may, however unfortunate, be scientifically unjustifiable
outside of the psychoacoustical realm. No slam intended, but
if you can not hear what this process may be doing to the quality
of the recorded audio (most of which is probably already butchered
to some degree if released within the past 8 years), you may be up
against a tough set of facts to attempt to rationalize... then attempt
to train your ears to perceive. Especially, when reducing the tested
data to MP3 quality.
> In addition, I have posted to my website
> several screenshots which visually demonstrate the effects of using this
> application.
I have searched this thread for a link.... I don't see one. Maybe my server
dropped something. I don't (won't) do any newsgroup but this one.
Do you also have this audio (source for screen shots) posted for streaming
or download in a high bit rate file ?
______________________________________________________
Taken from the primary source for the free Linux Not_Lame codec
site : http://www.idiap.ch/~sanders/not_lame/
"This encoder produces MP3 files of much higher quality (less
distortion) than every other *free* encoder".
______________________________________________________
We may only be able to influence your methodology by training your ears,
but I don't think you came here for that.
If I don't see that link by morning, I'll follow up with a decent reply to
your response anyway.
Mark T. Wieczorek
> With that in mind, I suppose the real question is, how much of an
> initial level reduction is needed before loss of fidelity is obvious.
> -4dB? -10dB? -30dB? -60dB? I guess we'll see!
What's the dynamic range of a 16 bit file? You don't want to go that far. I
think -30dB should be sufficient, but what do I know? It's probably not
difficult to do a test with each of the above.
> Whoops! It was Mark who suggested this test. Not David! Sorry! :)
=)
Don't worry about the flak here, no matter what your station in life -
"professional" "amateur" "demigod" "urchin" you'll find religious wars
wherever people gather.
You may be interested in the sleightly unrelated "Great Dither Shootout"
http://www.24-96.net/dither/
Lord Hasenpfeffer
>> So by posting in rec.audio.pro I've actually come to the right
>> place?
> That depends on whether or not you are here to state that you are
> without flaw in any of your reasoning and will never change your mind
> or approach in any way, regardless...... or, if you are in search of
> some facts regarding the methodology you say you employ.
Hehe!! Good answer. It shows that you've encountered both on more than
one occasion! :)
I'm definitely here for the latter, though, sometimes my often too-bold
personality causes others to incorrectly assume the former (until they
see proof otherwise).
> It would appear thus far, you are merely espousing to an opinion
> which may, however unfortunate, be scientifically unjustifiable
> outside of the psychoacoustical realm.
Espousing to an hypothesis? And yes, it's almost guaranteed that
outside of the psychoacoustical realm the basic principles of my
hypothesis are all but unworthy of any practical contemplation at all.
> No slam intended, if you can not hear what this process may be doing
> to the quality of the recorded audio (most of which is probably
> already butchered to some degree if released within the past 8
> years), <split, to borrow a technique>
Yes, I am aware of this "past 8 years butchering" thing to which you
refer. I have a personal library of over 2,100 compact discs mastered
obviously throughout all years from the beginning of CD to the present
day. I have been ripping and encoding my library for over 2 years now.
It didn't take long for me to notice curious, nearly-always consistent
"differences in the sounds" of my older CDs from the 80s vs. those from
the early 90s vs. those from about 1994 on to the present day. I now
understand that CD mastering techniques in the 80s were based on setting
peak levels upwards in relation to 0dB whereas in the 90s the focus was
reversed to being measured downwards in relation to 0dBFS.
One of several tasks I'm currently undertaking involves ripping and
encoding MP3s of every CD I own. My discovery of the application
"normalize", I believe, is helping me to better equalize these
differences in loudnesses by boosting the RMS levels of my older CDs to
where they're more closely resemble those of my 24-bit digitally
remastered KISS and Billy Joel CDs. I have definitely encountered
"grunge/rock" CDs from the 90s which are pumped *way* beyond anything I
would ever do or even suggest doing on my own with the RMS level of
*any* CD regardless of its age. In other words, I do my best to focus
on "slightly higher than middle ground" whenever I make my own
adjustments prior to encoding my final files.
> you may be up against a tough set of facts to attempt to
> rationalize... then attempt to train your ears to perceive.
> Especially, when reducing the tested data to MP3 quality.
I may be. Nevertheless, as a fairly well-seasoned, both amateur and
professional computer programmer throughout the past 20 years, I'm very
well accustomed to accepting and subsequently acting upon knowledge
derived from facts and/or conditions which can only be proven and/or
demonstrated at abstract levels.
>> In addition, I have posted to my website several screenshots which
>> visually demonstrate the effects of using this application.
>
> I have searched this thread for a link.... I don't see one. Maybe
> my server dropped something. I don't (won't) do any newsgroup but
> this one.
I have not yet provided links to my screenshots yet because, based upon
my experience with the other newsgroups during the past week, they have
a tendency to cause flamers who do not understand my hypothesis to crawl
out of the woodwork and hurl vicious insults at me in their ignorance.
I want to make certain that my stated purpose/hypothesis, the function
of the "normalize" application and the jargon necessary to effectively
communicate its results are all understood by those who are a part of
the discussion before I post them again.
> Do you also have this audio (source for screen shots) posted for
> streaming or download in a high bit rate file?
I am working here with only a 56Kbps modem so "high bit rate" files are
difficult at best for me to use. But I do the best I can with what I
have, yes.
> We may only be able to influence your methodology by training your
> ears, but I don't think you came here for that.
If you have good advice for how I might better train my ears, I'm
willing to take it.
> If I don't see that link by morning, I'll follow up with a decent
> reply to your response anyway.
Thanks.
Tomorrow I would like to "begin" by demonstrating as best I can both
visually and aurally the *default behaviour* of the "normalize"
application. The way things have gone so far in the other newsgroups, a
lack of understanding and/or knowledge of what this application does can
do a lot to throw a thread off-topic. I will need to create a couple of
new sample MP3s and screenshots for this, however, I'm too tired to
begin doing that at this particular time. :)
Chris Johnson
> perfectly fine for playback directly from that CD it is inappropriate
> for being encoded to MP3 because its amplitudes are too low to drive
> many of its frequencies above their Absolute Thresholds of Hearing given
> its existing, low-amp state.
>
> achieved since fewer of their frequencies will be subject to being
> discarded by the encoding process.
No, wrong. Sorry man :)
128K (or whatever it is) means that it's throwing away a very large
amount of low-level information anyhow.
All you're doing, _all_ you're doing, is taking that point where
information's getting thrown away, and making it louder and easier to
hear where it's happening.
I take it you are not a coder? Because if you were, you'd be able to
experiment with this sort of thing yourself. I have. The aspect you are
looking to adjust is 'ATH', short for Audible Threshold of Hearing, and
it sets the weighting of that threshold.
When you lower the ATH, here is what happens: the same amount of
information is distributed differently. You get a greatly enhanced
ability to represent reverb tails, soundstage depth. You lose punch,
clarity and distinctness, and it all turns to mud. It ain't really worth
it, take it from someone who has tried this.
Nice thought, though. If there was no limit to the data encoding, you
might even have a point there. But lossy coding is designed to run out
of bits at a certain point and stop. It doesn't really CARE whether the
loudest component is -3 db or -30. It'll just keep adding information
until it reaches 128K or whatever you're using, regardless.
If you feel that strongly about this, use an encoder like LAME in
which you can retain the psychoacoustic model but disable ATH. That's
what you're really trying to do.
Chris Johnson
Chris Johnson
"Geoff Wood" <ge...@paf.co.nz-nospam> wrote:
> For those who don't know, My Myke here insists on normalising every piece of
> his 2100-strong CD collection to -10dB average RMS, and considers any tracks
> that don't meet this criteria to be flawed, and incompetently produced.
> Despite serious attempts to clue him up he clings to total misconceptions
> regarding levels, amplification, attenuation, normalisation, the mastering
> process, etc.
I prefer -15db RMS myself, -10 is getting to be kinda squashed and
undynamic. But then, I master stuff, so it makes sense that I would have
a preference. Funny, a lot of massive hit records have had way, way more
headroom than that, even (especially?) in the vinyl days...
> He dismisses MFSLs Dark Side Of The Moon as being a peice of excrement
> because the highest peak is -4dB or so, and that buyers have been ripped
> off. (they didn't get all the bits they paid for ?).
Dark Side is a little weird that way. It's got extraordinarily hot
peaks, is unusually undistorted. I can see why MFSL would want to
represent that accurately.
To me, it's not so much that Dark Side is wrong, just that it might
sound awfully quiet if Myke normalizes everything else to -10. Could be
worse, though- 'The Wall' and Fleetwood Mac's 'Rumours' are even more
peak-happy. That's the sound- and it takes incredible playing to make it
happen, too. It's not happening by mistake or laziness.
Chris Johnson
Chris Johnson
Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote:
> Bob Olhsson wrote:
> > You are making a very fundamental mistake in your conclusion. But don't
> > take my word for it, just try different levels and listen.
>
> The sky is a certain colour. But don't take my word for it, just go
> outside and look up.
>
> In other words, why bother telling me that I am making a "very
> fundamental mistake" if you aren't also willing to cite what it is?
>
> Myke
Be respectful to Bob, he helped make some of the music in your
personal record collection if you have any taste ;)
Seriously- I'd bet what Bob means is this: 'no, it doesn't actually
sound better, just more squashed and with the peaks missing'.
I agree there's a fundamental mistake but I'm seeing a more
technical-to-mp3-encoding one, and I'd be interested to know if Bob
instead meant this: mp3 is already throwing out stuff above the ATH just
because it's only got so much information to work with. If you cause it
to weight such faint material equally with loud material (which you can
do by disabling ATH in some encoders), you'll get a much more
reverberant mp3, but it'll be real vanilla and bland, because it's still
throwing out most of the information. You're just telling it to keep
faint information, so it's throwing out louder information- or,
technically, it's using up data-rate on faint stuff and running out
before it's really dealt with the loud stuff.
Same result: 'no, it doesn't actually sound better'. You just like
stuff to be more consistently loud. That's fine. Doesn't mean it's
better by any stretch of the imagination, just different.
If you filter out everything above 12K, the mp3 encoder rejoices
because it has lots more data to expend on the crucial midrange, and
doesn't have to burn data trying impossibly to encode very high
frequencies anymore. Does that make it 'better'? It's certainly going to
be richer, 'warmer', etc that way. But again it's a judgement call, this
time for people who don't care about treble.
Chris Johnson
John LeBlanc
"Chris Johnson" <jinx...@sover.net> wrote in message
news:jinx6568-2471A8...@fe01.atl2.webusenet.com...
> To me, it's not so much that Dark Side is wrong, just that it might
> sound awfully quiet if Myke normalizes everything else to -10. Could be
> worse, though- 'The Wall' and Fleetwood Mac's 'Rumours' are even more
> peak-happy. That's the sound- and it takes incredible playing to make it
> happen, too. It's not happening by mistake or laziness.
Hello, Chris.
Same goes for the voice instrument. I've talked with many engineers who
actually remember when a vocalist knew how to "work" a mic well.
This ties in nicely with one of the longer threads from a few months ago
discussing the lack of dynamic program matter in today's pop releases.
John
Lord Hasenpfeffer
> I would be somewhat surprised that any current CD was not NORMALISED
> to somewhere around (just below) 0dbfs.
Yes, as that has apparently been "all the rage" for several years
running now. Actually, *abuse* of the 0dBFS threshold is really "all
the rage" these days. I personally am only interested in "normalizing"
older, quieter CDs mastered during the stone and middle ages before
encoding them as MP3s so that my final files sound somewhat "modern" to
my ears and not all tinny and weak when juxtaposed with MP3s encoded
from more recently mastered CDs.
In no way am I interested in brutally forcing any RMS levels to go
"through the roof". Anywhere from strictly normalized to "slightly hot"
levels (depending on genre) are all that I seek. Many times I've
encountered WAVs from CDs that are, according to my standards, mastered
too loudly. When this is the case, I _do_ simply "rip-and-encode" the
unmodified, original WAVs (contrary to what dipshit said of me earlier
when he said I seek to indiscriminately normalize every CD I own). If
that were the case, I'd end up making half of what's in my library
*quieter* than it is on the original CDs, not louder! :)
> Note that this says nothing about average level,
Correct.
> and where something intended to be listened to as a whole is split
> into tracks may say nothing about individual track levels.
Correct.
If normalization is "needed" on an individual file with which I'm
working out of context, I will "normalize" that stand-alone file.
If normalization is "needed" on an entire album, I will "batch
normalize" the files by a common value to preserve their original,
relative loudnesses.
If normalization is _needed_ for a "mix-CD" sporting tracks by various
artists from various sources, I will "normalize" all the tracks
individually to bring them to a common loudness, across-the-board.
> All the Normalising does is to ensure that the sample with the
> highest magnetude over the range being normalised (track or whole
> disk) is at the relevant level.
Correct.
However, depsite its name, the "normalize" application I use is capable
of doing more than just "textbook normalization" if I tell it to do so.
This has been a major stumbling block for me when I've previously
attempted to discuss its behaviour amongst others who've not used nor
even heard of it before. "Normalize" can be made to limit the peaks
when instructed by the user to boost an RMS level beyond the textbook
normalization level. One of the guys over in the other newsgroup
suggested that we call this "limitizing" because there is no other,
readily available, predefined textbook term to describe it.
> This is (almost) totally a non contentious thing to do when mastering
> a disk, the place where judgement comes into it is when limiting and
> compression are applied to raise the RMS level closer to the
> normalised level.
Bingo!! I couldn't have said that any better than you have. Naturally
you can see why when someone mistakenly believes my stated goal is to
push the average RMS level of "Dark Side Of The Moon" all the way up to
0dBFS, they freak out in horror. :) I assure you, such is *not* my
intention.
> Further the phycoacoustic model IIRC derives 'too quiet' threshold
> from the RMS level of the signal, not an absolute threshold (IE at
> some frequency it may be 30db below curent RMS level, NOT 40db below
> 0dbFS).
Mmm-HMMM.... Now *that's* something of which I was not previously aware.
If that's true then my hypothesis for boosting amps to save freqs may
indeed be fatally flawed.
Are you sure you're not thinking of frequency masking when you say this?
> It is worth noting that whatever you do to the input data, it is only
> possible to fit a fixed amount of information into the output
> stream,
Now, I do understand that, however, I haven't contemplated it very much.
> if you force the encoder to include more frequency data (higher
> resolution or more bands active, then time or ampletude resolution
> MUST necassarily suffer).
Excellent information. By time resolution you mean in that the file
would have to be made to play slower in order to accommodate the
increased amount of data?
Also, would not such effects be greater at lower bitrates than at higher
ones?
And if so, would this not mean that my hypothesis would actually become
more appropriately applied towards higher bitrate MP3s than at lower ones?
Because I'm guessing here, by reversing your logic, that a large enough
bitrate could eventually be employed which would cause the encoder to
either "pad the file with zeroes" or store the additional data depending
on the normalized status of the WAV being encoded.
And if that's true, what bitrate may I be talking about?
Stephen Raiman
made it a point to do some study on the subject.
The compression algorithm still needs to take a given amount of data
out of the original wav file. Don't forget, using CBR (constant bit
rate) encoding, the file must still fit into the selected MP3 bit
rate. Increasing or decreasing the overall level of the wav file would
increase or decrease the threshhold at which the algorithm removes
data in order to maintain a fixed file size. In other words, lowering
the overall level would also lower the level below which the algorithm
deems data "masked".
This brings us to VBR (variable bit rate) MP3s. VBR is a newer
technology which aims to give more bits to complex passages, and fewer
bits to simple passages.
How does this affect your hypothesis? Damned if I know. Were you were
basing it on CBR or VBR?
Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote in message news:<3EFE8232...@spamsucks.ionet.net>...
> >> why not add rec.audio.pro to the list and see what they think?
> >>
> >> I like watching a good train wreck.....
>
> > YES! ROTFLOL! On second thought, my tolerance for gore may not be
> > up to the spectacle. :-)
>
>
> Laugh it up.
>
> I spent time today browsing websites and reading FAQs, etc. to find out
> more information about the Psycho-Acoustic Model employed by MP3. It
> is indeed similar to MiniDisc's ATRAC compression scheme in that it
> removes not only "masked frequencies" but also frequencies which
> are determined to be "too quiet to be heard" by common human ears.
>
> This proves to me that MP3s encoded from "quieter WAVs" which have
> been ripped from CDs mastered with low, average amplitudes will suffer
> more at the hands of a lossy audio data compression algorithm than
> will MP3s encoded from WAVs which have first been "appropriately
> normalized" (i.e. Linux: $normalize -ba -10dBFS ) prior to being encoded.
>
> While my MFSL CD of Pink Floyd's, "Dark Side Of The Moon" may be
> perfectly fine for playback directly from that CD it is inappropriate
> for being encoded to MP3 because its amplitudes are too low to drive
> many of its frequencies above their Absolute Thresholds of Hearing given
> its existing, low-amp state.
>
> "Batch normalizing" the WAVs from a "quiet CD", however, will raise
> their _collective_ average amplitude to a significantly preferable level
> capable of yielding better sounding MP3s than would otherwise be
> achieved since fewer of their frequencies will be subject to being
> discarded by the encoding process.
>
> If applied sensibly and with appropriate grace for the sake of
> preserving the original dynamics of the recording, such "batch
> normalizing" will *not* bring clipping, limiting, compressing or adverse
> dynamic-range-related effects to the WAVs in question.
>
> The same holds true for MiniDisc recordings as well since the ATRAC
> compression method too seeks to remove frequencies which are deemed "too
> quiet to be heard".
>
> Aggregate: Lower amplitudes cause higher frequency loss, therefore,...
>
> ...louder *is* better (with lossy).
>
> Myke
Lord Hasenpfeffer
> Be respectful to Bob,
I was not being disrespectful to Bob. I have no reason to be.
He simply took the time to tell me I was making a fundamental flaw but
didn't cite the fundamental flaw. I don't understand why anyone would
bother posting such a message without further explanation.
Another analogy would be:
Patient: "Doctor, is there something wrong with me?"
Doctor: "Yes."
To this there is no meaningful information in this; no closure; no
resolution.
> he helped make some of the music in your
> personal record collection if you have any taste ;)
Hmmm... Namely? (Just out of curiosity.)
And yes, I do have taste - although John LeBlanc seems convinced that
it's a tossup between horseshit and dogshit, about which I think it's
"very deep" of him to say so.
> Seriously- I'd bet what Bob means is this: 'no, it doesn't actually
> sound better, just more squashed and with the peaks missing'.
Well, while that *could* be an accurate assessment of the situation, I
assure you, that it is not.
> I agree there's a fundamental mistake but I'm seeing a more
> technical-to-mp3-encoding one, and I'd be interested to know if Bob
> instead meant this: mp3 is already throwing out stuff above the ATH just
> because it's only got so much information to work with.
Bob didn't really say what he meant, so unless he returns to do so, we
cannot be certain of anything in that regard; one way or another.
A technical-to-mp3-encoding response is why I'm here, so if you've got
one, by all means, don't keep it to yourself.
> If you cause it to weight such faint material equally with loud
> material (which you can do by disabling ATH in some encoders),
> you'll get a much more reverberant mp3, but it'll be real vanilla
> and bland, because it's still throwing out most of the information.
I'm not seeking to cause an encoder to weight faint material equally.
I personally have no problem with the standard way in which ATH-based
lossy compression algorithms are made to do their dirty work. I think
it's a very kewl idea and I've been recording to great effect with
MiniDiscs for 6 years.
My main concern goes a little sump'm like this...
Like ta hear it? Hear it go...
I own 2 different CD versions of Pink Floyd, "Dark Side Of The Moon":
Pink Floyd, "Dark Side Of The Moon" (1973)
Mobile Fidelity Sound Lab / Ultradisc II
==========================================
level peak
-22.5712dBFS -10.6335dBFS track01.cdda.wav
-20.5430dBFS -9.3433dBFS track02.cdda.wav
-17.6751dBFS -8.1199dBFS track03.cdda.wav
-18.9479dBFS -7.8896dBFS track04.cdda.wav
-18.5242dBFS -7.6373dBFS track05.cdda.wav
-15.7739dBFS -4.9832dBFS track06.cdda.wav
-17.1860dBFS -6.1509dBFS track07.cdda.wav
-16.3045dBFS -4.7547dBFS track08.cdda.wav
-16.4466dBFS -5.5207dBFS track09.cdda.wav
-14.5302dBFS -4.4229dBFS track10.cdda.wav <<<<< highest peak
-17.1589dBFS average level
Pink Floyd, "Dark Side Of The Moon" (1973)
Capitol, Digital Remaster; (P)1992,(C)1994
==========================================
level peak
-15.5109dBFS -3.3526dBFS track01.cdda.wav < MFSL's tracks 1&2 combined
-12.5016dBFS -3.2019dBFS track02.cdda.wav
-13.9399dBFS -2.2710dBFS track03.cdda.wav
-13.4255dBFS -2.4928dBFS track04.cdda.wav
-11.6618dBFS -0.2804dBFS track05.cdda.wav <<<<< highest peak
-13.7878dBFS -3.1469dBFS track06.cdda.wav
-12.8113dBFS -0.9773dBFS track07.cdda.wav
-13.8615dBFS -2.4098dBFS track08.cdda.wav
-11.7779dBFS -1.6673dBFS track09.cdda.wav
-12.9274dBFS average level
Obviously, there are some significant differences between these two
versions of this album.
If I were to rip 2 huge WAV files, each containing the full contents
from each CD, and then encode them, as is, directly to MP3s, would not
more of the frequencies in the Ultradisc II version be in greater danger
of being discarded by MP3's lossy psychoacoustical filtering process
than would be the case with Capitol's version simply because Capitol's
version is mastered, on average, approximately -4.25dB louder than
MFSL's version?
If no, why not?
Myke
P.S. Also as a side note, it is quite obvious from just looking at the
level and peak readings of all of the tracks from both of these CDs that
Capitol's version is significantly sonically different from MFSL's. In
so doing, and in your estimation, has Capitol completely trashed this
classic or has Capitol done Pink Floyd and Alan Parsons a favour? In a
side-by-side comparison test with his amp's volume set at the same
level, Joe Sixpack would surely accept Capitol's remaster and probably
think something is seriously wrong with the older, quieter one from MFSL.
Lord Hasenpfeffer
> This is how I understand MP3 compression. I'm not an expert, I've only
> made it a point to do some study on the subject.
Welcome aboard!
I too am not an expert and have only made it a point to do some study on
the subject. And, in fact, my asking about this in this forum is just
another addition to all of that.
> The compression algorithm still needs to take a given amount of data
> out of the original wav file. Don't forget, using CBR (constant bit
> rate) encoding, the file must still fit into the selected MP3 bit
> rate. Increasing or decreasing the overall level of the wav file would
> increase or decrease the threshhold at which the algorithm removes
> data in order to maintain a fixed file size. In other words, lowering
> the overall level would also lower the level below which the algorithm
> deems data "masked".
OK. Unless there's something about "masking" that I don't yet understand
which links it to RMS levels in a *static* sense (as opposed to in a
dynamic sense), then I don't see how this applies to my original
hypothesis which is actually focused on the static relationship(s), if
any, between RMS levels and the "Absolute Threshold of Hearing" aspects
of psychoacoustical lossy audio data compression techniques.
> This brings us to VBR (variable bit rate) MP3s. VBR is a newer
> technology which aims to give more bits to complex passages, and fewer
> bits to simple passages.
Yeah, I'm hearing you now... And indeed you are the first to mention
this in any of the discussions I've shared with others on this topic so
far. Good call!
At least to me on the surface, VBR does appear to be a viable candidate
option for addressing the limitations imposed by CBR encoding in cases
where heightened RMS levels might cause a need for more data to be
retained than could otherwise be. That is, if heightened RMS levels do
indeed cause more frequencies in the WAV being encoded to survive the
ATH chopping block... Interesting! Thanks for that, definitely.
> How does this affect your hypothesis? Damned if I know. Were you were
> basing it on CBR or VBR?
Actually I was thinking only in terms of CBR with both 128kb/s and
192kb/s bitrates in mind. With this you have certainly added a
colourful new dimension to this discussion as far as I'm concerned. And
that's the kind of relevant informational exchange I'm definitely hoping
to create by way of this thread. I think a lot of people could really
learn a lot from this discussion if the original topic can be
maintained. I eagerly await the day that we all can bring it to a solid
conclusion.
Myke
Lord Hasenpfeffer
> 128K (or whatever it is) means that it's throwing away a very large
> amount of low-level information anyhow.
Correct. This discussion makes no real, practical sense unless lossy
compression is assumed ( hence the modifier, "(With Lossy)", in the
Subject: line! *lol* ). After that, the bitrate and even the format in
question (e.g. OggVorbis, MP3 or even Sony's ATRAC method for MiniDiscs
) is really not relevant, AFAICS. I think everyone can agree that lower
bitrates equal lesser quality; the differences between mono and stereo
files notwithstanding.
> All you're doing, _all_ you're doing, is taking that point where
> information's getting thrown away, and making it louder and easier to
> hear where it's happening.
OK, this point as you've described it is not entirely clear to me.
> I take it you are not a coder?
I am a 20-year veteran coder whose first program was written in BASIC on
a Timex/Sinclair 1000 in April 1983, however I am not a coder of audio
processing software (yet?). :)
> Because if you were, you'd be able to experiment with this sort of
> thing yourself. I have.
I've been offered a couple of good suggestions for tests which I can
perform using particular applications which are currently installed on
my Linux box. I just haven't found time to conduct them yet.
> The aspect you are looking to adjust is 'ATH',
> short for Audible Threshold of Hearing, and
> it sets the weighting of that threshold.
Actually, I'm more interested in leaving the ATH engaged and alone right
where it has been set by the developers of my encoder. Not only that,
I'm not *able* to switch out the ATH on any of my MiniDisc portables and
home decks.
> Nice thought, though. If there was no limit to the data encoding, you
> might even have a point there. But lossy coding is designed to run out
> of bits at a certain point and stop. It doesn't really CARE whether the
> loudest component is -3 db or -30. It'll just keep adding information
> until it reaches 128K or whatever you're using, regardless.
Does not VBR seek to enable MP3's ability to use more bits as needed
would otherwise be the case with CBR?
> If you feel that strongly about this, use an encoder like LAME
I use "notlame".
> in which you can retain the psychoacoustic model but disable ATH.
> That's what you're really trying to do.
With all due respect, I really don't believe that it is. And I think
that if we were to head off in that direction, the discussion would only
serve to throw everyone here off-track with regard to the original
question which assumes a fixed presence of ATH.
The object here is not to defeat ATH because I believe ATH is a wholly
worthwhile and entirely sensible element of the psychoacoustical lossy
compression concept. The primary concern of this discussion is to
determine whether or not it is sensible or senseless too accept without
question the RMS levels of all commercial CDs - both old and new - as
adequate for being encoded or transferred in an environment where
ATH-based lossy compression schemes are present to affect the result?
Lord Hasenpfeffer
>> With that in mind, I suppose the real question is, how much of an
>> initial level reduction is needed before loss of fidelity is obvious.
>> -4dB? -10dB? -30dB? -60dB? I guess we'll see!
> What's the dynamic range of a 16 bit file? You don't want to go that far. I
> think -30dB should be sufficient, but what do I know? It's probably not
> difficult to do a test with each of the above.
Actually, I was asking that in a rhetorical sense.
(i.e. Which one? We'll see!)
> You may be interested in the sleightly unrelated "Great Dither Shootout"
> http://www.24-96.net/dither/
Looks interesting! :) Gimme some time to actually do it.
Lord Hasenpfeffer
> If you're really that interested in MP3s,
I am.
> it seems to me you'd try some other encoders just out of curiosity.
After 2+ years of working with it, notlame's quality leaves me with
quiet *incurious* about other encoders. That's how well I like it! :)
...with the exception, of course, of the actual fraunhofer encoder which
is not available to me for free as is notlame.
Last I read anyway, some tests were conducted which demonstrated notlame
to actually surpass the fraunhofer encoder at its own game.
>>> generally speaking, louder is better regardless of what's being
>>> evaluated.
>> Oh really? <snip!> I'm actually kinda surprised to find anyone in
>> here that's brave enough to say what you just did! :)
> Sounds like you stumbled into rec.audio.opinion. Don't drop your soap.
Ha! That's "no SHIT"! :)
> I'm guessing that everyone in here would probably agree in general with
> the "louder is better" truism,
Well, my ears tell me that it is, but my intellect can actually see
things from the other side of the fence as well. Still, my natural
tendency is to go with my ears since music is so much an emotional
rather than intellectual 'thang'. :)
> but most people here don't give much thought to MP3s unless
> it's with regard to the P2P/Napster/IP quagmire.
Hmmm... That's their loss then, I suppose. The RIAA's got nothing to
complain about where I'm concerned. I make nearly 100% of my own MP3s
from CDs I and my wife have personally purchased since we first began
buying CDs as teenagers in the mid-late 80s.
Regarding the MP3s I'm making today which I'd like to still be enjoying
in 2008:
> Well, I doubt your hearing acuity will improve, so unless you're sick of
> the tune by that time, I see no reason you'd stop enjoying it if you
> like it now.
You have a truly uncommon common sense approach to this! I like it! :)
> I doubt if many people here care to quibble over the fidelity of this
> MP3 vs that one, *unless* you pirated it from their CD, in which case
> that MP3 will be indistinguishable from the source regardless of how
> it was encoded.
But of course! You've just described me "to a T"! :-D
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message ...
> The effect of increasing amplitude in an attempt to obtain "better sound"
> with uncompressed audio data is negligible.
Although "better sound" is a completely subjective term, I dissagree.
This is done daily and you have acknowledged this as fact in your
subsequent reply to me. It can be anything but "negligible"... it can be
blatantly obvious, reveal severe change, be quite impressive or even
highly offensive - depending on both the listener and the skills of the
person performing the manipulation of the audio.
> However, the effect of
> increasing the average level of a WAV prior to "munging" it with an
> amp/freq-based lossy compression algorithm such as ATRAC or MP3
> seems at a glance to me to be rather substantial...
"At a glance", the Mona Lisa seems like a rather trivial piece of art. Either
we listen 'at a glance' (common use of MP3 codec), or we listen with a
keen sense of critique. To do the latter, requires a certain degree of
knowledge and experience, that of which you admit to having little.
In fairness, I would at least agree that it would be best to alter, or otherwise
'process', only the most data rich audio files as opposed to the ATRAC or
MP3 skeletons thereof.
However, I would l suggest getting to know a great deal more about the tools
of the trade in audio before blindly using such functions as RMS normalization.
> > Depending on your initial source material, "normalization", as such, may
> > do absolutely *nothing* to your audio file.
>
> To the quality of the source audio WAV, yes, it may indeed do nothing,
> however, if "normalization", as such, helps to boost more of the
> original WAV's frequencies into a more audible range, surely that would
> cause those frequencies to be preserved rather than discarded by a
> psycho-acoustic lossy compression scheme, no?
You still seem to be saying that you are willing to inflict damage caused
by RMS normalization in order to have seperate files that play back in a
fashion that prevents you from having to change the volume during the
playback - while at the same time implying that there must be some
underlying set of frequencies, that if boosted far enough, will somehow
suddenly reveal themselves or enhance the remaining audio so that
they can be preserved on a lossy format.
See not only my previous analogy :
"by using the method you describe (that is,
RMS Normalization), you have destroyed significantly important audible
portions of extremely usable audio, in favor of saving what may be near
inaudible (and by most codecs = 'diposable')"
...but also that of Chris Johnson :
"Chris Johnson" wrote in message news:jinx6568-B6AEDC...@fe01.atl2.webusenet.com...
...and that of Dan Mills :
"Dan Mills" wrote news:4qbndb...@spamblock.demon.co.uk...
The common answer would seem to approach, "no".
Both of these data compression schemes are actually quite impressive
in the manner by which they have the ability to be discerning enough
to discard portions of the audio that are essentially (given it's popular
focus of usage; casual and lo-fi) non-discernable and theoretically
inaudible, thus disposable. Rather ingenious actually.
What is it that you believe to be hidden within the original audio that
cannot be found by turning up the volume knob?
> > Would you mind defining, "appropriately normalized." (?)
> As I understand it, the technical definition of normalization indicates
> that it is a process of simply boosting the amplitude of a digital audio
> recording by the amount necessary to cause its loudest peak to reach
> "Full Scale". Nothing more, nothing less.
Yes.
> At this point it is important for you to realize as a participant in
> this discussion that while the Linux application called "normalize" can
> indeed readily perform the task of normalization according to its strict
> textbook definition, it can and will do more (e.g. employ limiting to
> avoid clipping) if given user-supplied instructions which would require
> it to do so.
Essentially, limiting *is* clipping.
> > If you are "normalizing" to an "average" RMS level, you are merely
> > wreaking havok on the peak portions of your program that exceed your
> > specified RMS level as compared to 0dBFS.
>
> In theory, this would seem to be the case, yes, however, in practice, I
> cannot visually discern in my WAV editor where the sins I'm accused of
> committing are occurring as a result of having used "normalize" to boost
> the amplitude of a WAV.
Just to interject... one generally *listens* to audio rather than looking at it.
In this forum, listening is everything. If you get lost in waveforms rather
than soundwaves, it is easy to become confused.
> In addition, I have posted to my website
> several screenshots which visually demonstrate the effects of using this
> application.
I found them. They _look_ rather harmless... much more like a standard
normalization based on peaks only, not RMS.
> To better illustrate, let's say we have a WAV with a maximum peak
> amplitude of -6dBFS. Using "normalize" to normalize this WAV, a simple
> boost of +6dB applied to its average amplitude (RMS?) would cause the
> whole recording to become 6dB louder and send that maximum peak right up
> to Full Scale. Correct me if I'm wrong about that.
Average RMS amplitude is determined by the entire file content *selected*.
Some sections of the song may be higher, some parts lower... it is not a
constant, therefore the term "average".
You would use "Peak" normalization to accomplish the goal you describe
above. Why? There could be specific segments of the program that
are actually much higher in RMS than the "average" would indicate. It
is these sections that would suffer the most traumatic damage.
> "Batch normalize" is the term used by the author of "normalize" to
> describe the exact same process being applied to an entire set of WAVs
> rather than to just one WAV. If a set of WAVs is simply "normalized"
> they are all made to sound "equally loud". However, if a set of WAVs
> are "batch normalized", the loudnesses of all the files in question are
> boosted by a single common value, in effect treating the whole set as if
> it were a single WAV. "Batch normalizing" is essential for preserving
> the original _relative loudnesses_ of the full set of WAVs as they
> originally were in relation to each other.
Apply that mathematical 'assumption' (for example) to the samples you
gave elsewhere of the two Pink Floyd CDs. Does it appear they simply
performed a "Batch Normalize" routine on the entire earlier version?
NO. Each file had to be treated seperately because each file had both
a different peak and a different average level. Those points aside, they
all had to mesh together to *appear* to be relative in volume, and that
is a function of the human ear, not a batch script.
> In short, when creating a "mix" of songs from various sources, simple
> "normalization" of all the files would help to bring them all to a
> fairly uniform loudness which would save the listener from constantly
> having to manually adjust his volume control to compensate for the
> varying loudnesses carried forth from the various original sources.
Oh please. <g> If this is why you are putting forth this rather forceful
"Hypothesis" of louder is better, you have just lost a great deal of credibility.
Try a new header: "Convenience and self-gratification is better".
Better than *what* ?
> > A variety of tools might need to be, or could be, applied
> > to the soundfile before "normalization" ever enters the picture - - but all
> > of these tools *DO* bring clipping, limiting, and compression into play
> > and most definitely create "adverse dynamic-range-related effects".
>
> I think I'll hold off on this point for now until I know that we're on
> the same page with all that has gone before so far.
If you're asking will I, or anyone here, turn to *your* page...... well..... (?)
David Morgan (MAMS)
Again, short & sweet... some snippage for brevity....
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3EFFA8B5...@spamsucks.ionet.net...
> David Morgan (MAMS) wrote:
>
> >> So by posting in rec.audio.pro I've actually come to the right
> >> place?
>
> > That depends on whether or not you are here to state that you are
> > without flaw in any of your reasoning and will never change your mind
> > or approach in any way, regardless...... or, if you are in search of
> > some facts regarding the methodology you say you employ.
>
> I'm definitely here for the latter, though, sometimes my often too-bold
> personality causes others to incorrectly assume the former (until they
> see proof otherwise).
What if some of these people have been studying said "proof" since it's
inception and fail to agree with you? Your personality may have nothing
to do with it unless you fail to be willing to hear *their* proof as well.
Because they may have "proof" of their own, does that make them
"too bold" as well? (no answer needed) Let's just get along here.
> And yes, it's almost guaranteed that
> outside of the psychoacoustical realm the basic principles of my
> hypothesis are all but unworthy of any practical contemplation at all.
I'm glad you at least consider that this is a possibility. ;-)
> Yes, I am aware of this "past 8 years butchering" thing to which you
> I have a personal library of over 2,100 compact discs mastered
> obviously throughout all years from the beginning of CD to the present
> day. I have been ripping and encoding my library for over 2 years now.
Good heavens !!! What on earth for ?? You are going through all of this
simply to place the audio onto a temporary and an essentially 'disposable'
medium !!!
> It didn't take long for me to notice curious, nearly-always consistent
> "differences in the sounds" of my older CDs from the 80s vs. those from
> the early 90s vs. those from about 1994 on to the present day. I now
> understand that CD mastering techniques in the 80s were based on setting
> peak levels upwards in relation to 0dB whereas in the 90s the focus was
> reversed to being measured downwards in relation to 0dBFS.
Those are simply two different 'scales'. Volume is still a product created
by the end user when referencing only these scales.
> One of several tasks I'm currently undertaking involves ripping and
> encoding MP3s of every CD I own.
2100, eh? Do you work ?? <g> I really hope your heart is in very good
condition when MP3 suddenly becomes a thing of the past, like Laser Disc,
like 8-tracks, like Casettes and LPs are rapidly becoming, like 8-bit sampling,
4-bit video games and ISA slot cards. I give it until 2008.
> >> In addition, I have posted to my website several screenshots which
> >> visually demonstrate the effects of using this application.
> >
> > I have searched this thread for a link.... I don't see one. Maybe
> > my server dropped something. I don't (won't) do any newsgroup but
> > this one.
>
> I have not yet provided links to my screenshots yet because, based upon
> my experience with the other newsgroups during the past week, they have
> a tendency to cause flamers who do not understand my hypothesis to crawl
> out of the woodwork and hurl vicious insults at me in their ignorance.
Well, I tend to research things a bit before inserting my foot, so I found your
admittedly 'old' screenshots and an audio sample.
<from alt.audio.minidisc - 06/29/03>
> http://www.mykec.com/mykec/images/All_Right_Before.png
>
> http://www.mykec.com/mykec/images/All_Right_After.png
This screenshot looks relatively harmless, as if only peak normalization
were applied, but it does little to define your assumption that a "Batch"
processing of RMS normalized files will guarantee you a decent 'relative'
loudness between different sources. It does nothing to assure that you
maintain the *integrity* (or what's left of it on newer material) of your
original files.
> I want to make certain that my stated purpose/hypothesis, the function
> of the "normalize" application and the jargon necessary to effectively
> communicate its results are all understood by those who are a part of
> the discussion before I post them again.
I trust that you will be making some new ones. If you feel you are
presenting some 'new' revelation to this group regarding what happens
when one disregards certain file attributes in favor of RMS normalization
solely for the sake of not having to reach for the volume control, what the
other groups countered you with will pale in comparison. At least one of
those who offerred you the best logic is a regular contributor here and
is considered by many as quite valuable. I notice you have already told
him to "bug out"... but this is, after all, usenet, and escape is futile... you
will be assimilated. <g>
The problem with this discussion thus far, in each place I reviewed any
of your posts, is that you are trying to assimilate others. Even with my
tiny amount of knowledge I should ask that you give up now... before
the rest of the 'collective' awakens and chimes in. It's Monday morning
and it would be nice to see that you exist at least until Wednesday. ;-)
I respect your apparently infinite curiosity of your teenage years.
alt.culture.us.1980s
alt.audio.minidisc
alt.music.duran-duran
rec.audio.tech
ok.tulsa.general
rec.music.beatles.moderated
rec.music.artists.paul-mccartney
alt.fan.kroq
You appear to be quite intelligent, so you should be open to learning
enough about audio to rationalize that what you are doing with RMS
normalization to speed up your 'archiving to MP3' process is likely
damaging the source material and doing little to improve the sound
of the encoded files other than to make them loud.
Dozens of people have told you so, but you continue to search for
some one or some place that will reassure you or be influenced by
you for some reason. I don't quite follow that part, although I do see
you opening up in recent posts that relate to encoding, but remain in
denial to those that pertain to standard, uncompressed audio - and
that is where your errors may rest.
> > Do you also have this audio (source for screen shots) posted for
> > streaming or download in a high bit rate file?
<from alt.audio.minidisc - 06/29/03>
> http://www.mykec.com/mykec/audio/All_Right.zip
I'm afraid that I didn't download the zipped file.
> Tomorrow I would like to "begin" ...
Well... you 'began' yesterday with a blanket statement and the reaction
was somewhat less than hospitable, overall. I wish you better luck.
Check the archives.... this group has covered the devastating effects
of destroying dynamic range in favor of volume for many years.
> ... by demonstrating as best I can both
> visually and aurally the *default behaviour* of the "normalize"
> application.
I cannot speak for others,most of which are far more advanced than I,
but I have been using the "normalize" function and various alterations
thereof in a large variety of audio software for at least 15 years.
I'm afraid that I won't be attending the class.
> The way things have gone so far in the other newsgroups, a
> lack of understanding and/or knowledge of what this application does can
> do a lot to throw a thread off-topic.
You will find NO shortage of people here that *fully* comprehend the
normalization process... My advice would be to ask questions, not to
walk into the Dean's office and tell him that you are about to begin
teaching his classes.
David Morgan (MAMS)
a flame-fest, so I'm skipping today's class....
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3EFFCEE6...@spamsucks.ionet.net...
> Chris Johnson wrote:
> > Be respectful to Bob,
>
> I was not being disrespectful to Bob. I have no reason to be.
>
> He simply took the time to tell me I was making a fundamental flaw but
> didn't cite the fundamental flaw. I don't understand why anyone would
> bother posting such a message without further explanation.
I think it's the most kind manner feasable to inform someone that
they were either slightly hard of hearing or not very 'discerning'. <g>
> Another analogy would be:
>
> Patient: "Doctor, is there something wrong with me?"
> Doctor: "Yes."
In a 'round about way, people here will say :
Patient : "Doc, It hurts when I do this...."
Doctor : "Well then, don't do that."
The problem is, you aren't discerning the pain inflicted on the source
material _if_ you indescriminately apply RMS normalization without
serious analysis of the waveform to determine it's peak to average
ratio and how that is distributed throughout the material - say nothing
of lumping a bunch of various source files into a single "batch" script.
> > Seriously- I'd bet what Bob means is this: 'no, it doesn't actually
> > sound better, just more squashed and with the peaks missing'.
>
> Well, while that *could* be an accurate assessment of the situation, I
> assure you, that it is not.
I believe that it *is* an accurate assessment. Pretty bold of you to
ASSuME - to the point of assuring us all - that it is not.
> Bob didn't really say what he meant, so unless he returns to do so, we
> cannot be certain of anything in that regard; one way or another.
If you knew Bob (I don't, I've only read about him, listened to hours
upon hours of recordings he has been affiliated with over the years,
and paid great heed to his posts here) you might be a little more
open to his style of discussion. I'm very certain that he is quite
coherent enough to say what he means.
> A technical-to-mp3-encoding response is why I'm here, so if you've got
> one, by all means, don't keep it to yourself.
Use the best codecs at a bit rate you can be satisfied with on disc
consumption. Above all, and if it matters (and it would appear thus
far *not* to matter) apply a descriminating ear to some part of music
perception other than just VOLUME.
> I'm not seeking to cause an encoder to weight faint material equally.
Thus far, this is the only substantially stated (perhaps I've misinterpreted)
point you have made... repeatedly stating that by bringing up the average
RMS level you will have more data to work with before encoding. Other
than this, volume matching has been your only other reasoning.
The quibble from this forum will be that you are potentially damaging the
audio before converting - and doing so for the purpose of 'standardizing'
playback levels. The other quibble would be that due to the mediums
which you have been exposed to (that lowered bar, if you will) you may
not even be able to hear what you have done.
Yup... the industry started learning how to "crank it up" and began their
adventure into lowering the bar for standards in audio distributed to the
public. <g>
Now look at us... we're so lazy we can't even rationalize the thought of
reaching for the remote, let alone the volume control itself.
--
David Morgan (MAMS)
http://www.m-a-m-s.com
Morgan Audio Media Service
Dallas, Texas (214) 662-9901
_______________________________________
http://www.artisan-recordingstudio.com
Erwin Timmerman
> John LeBlanc wrote:
>
> > I use my Jukebox in the same way, and for pretty much the same reason. It's
> > adequate for a rip-and-run, which will just be deleted next round anyway.
>
> Therefore, you obviously have nothing constructive to contribute to this
> thread. Please take your pollution elsewhere and stop diluting this thread.
>
I don't think you're aware that this is a public forum, where anyone can respond
in any way. The fact that you started a thread doesn't mean you own it, or that
people cannot take it into other directions. Despite your self-inflicted Lord
status, you are not the one who demands what, and in what way, should be answered
to your question. If someone wants to discuss a completely different topic in a
thread he is free to do so, if the discussion was started in that particular
thread. Not changing the subject may not be wise, because the people who are
actually interested in the new topic may skip the messages because they are not
interested in the original subject. But it isn't absolutely required to do so.
Let alone to be demanded by the thread starter. Who are you to demand this?
Pretty arrogant I'd say.
Calling people from other groups (or people in this one who don't respond to your
liking, some of who have produced some pretty impressive albums) names, or
arrogantly stating that an answer wasn't within your narrowly defined scope of an
answer, doens't gain you any respect. If you need to talk down to other people
just to make yourself look better, it doesn't make the others look bad, it makes
YOU look bad.
This message of course, has not yet helped you any further with your mp3
hypothesis and thus, too, doesn't comply with your demands of how my response
should have been. Bad, bad me.
Well fuck you Myke.
It is I and I alone who chooses how to respond (as long as I don't break any
laws). By sitting behind your computer screen all the time you might not have
become aware that this world is inhabited by very different human beings, who
have totally different ways of communicating. You are not the one to define the
right way. You may have your preferences, and are freely to express them, but
talking down to people who don't comply is just downright arrogant, and will get
you killfiled soon by people who might have something very interesting to say to
your questions. If someone acts different from you, it is not stupid, it is just
different.
I DID have something to say about an extra interesting step in the wav-comparison
test, and about listening levels, about which "not hearable" sounds might be
dumped and which might not and why, about average levels and normalizing, but
with your attitude I just don't feel like answering those questions any more.
Since I'm not paid by you, I'm not obliged to either.
I'm sorry for you that this message didn't help you any further with your
original question, and that it polluted your precious thread. Well, that's life.
Get over it.
I won't respond to your reply if it contains any flame bait. Not because you have
"won", but because I don't feel the need to be in a pissing contest with you.
You're not worth the effort. No one is, actually. I don't care. Maybe when you
take some time to look at that other screen, the one with a pretty remarkable
resolution and contrast, AKA "the outside window", you might get some perspective
on life. You're still young. ENJOY life. Respect yourself, AND respect others.
That's the key.
Good luck (and I mean that),
Erwin Timmerman
David Morgan (MAMS)
you *really* feel about this ? I mean, don't hold any punches....
;-)
DM
"Erwin Timmerman" <erw...@stack.no> wrote in message news:3F0003D9...@stack.no...
Geoff Wood
No. There is an extremely minor difference between these two versions (in
the statistics sense that you quantify).
As before in r.a.t, a difference in around 3dB in tthe maximum peak which
will make diddly-squat difference in each resultant MP3. No frequencies are
going to appear or dissappear with the two versions. If there is any
difference, it would be hugely swamped by the damage done primarily in the
MP3 coding in the first place.
But what would I know ...
geoff
Geoff Wood
> didn't cite the fundamental flaw. I don't understand why anyone would
> bother posting such a message without further explanation.
>
> Another analogy would be:
>
> Patient: "Doctor, is there something wrong with me?"
> Doctor: "Yes."
Maybe he has seen the threads where you have been offered concise and
accurate help, but have just gone on steamrolling ahead.
geoff
Roger W. Norman
--
Roger W. Norman
SirMusic Studio
Ro...@SirMusicStudio.com
301-585-4681
"John LeBlanc" <john__...@hotmail.com> wrote in message
news:vdqcnX69BKT...@giganews.com...
>
> "Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> news:3EFE8232...@spamsucks.ionet.net...
>
> > ...louder *is* better (with lossy).
>
> Seems to me that's like saying horse shit tastes better than dog shit.
>
> John
>
>
Roger W. Norman
leaves you with a turd. And about 100 other neat little sayings.
You ARE posting to rec.audio.PRO and the reason you get this back is simply
because it is PRO. We don't do mp3 for a living. We do real audio with the
client's acceptance of the product as job one. What you're talking about
doing is about what some kid would do. So sorry if you don't like John's
post, but I thought it was absolutely correct.
Besides, if you want to know how to go about making music work with mp3
well, then dig up Stephen Paul. He's got engineering credits out the butt,
and he knows exactly how to make a mix sound good enough to work well as an
mp3. But he doesn't take an existing mix and expect it to sound like
anything decent as an mp3. It has to be MIXED to be an mp3.
So louder ISN'T better.
--
Roger W. Norman
SirMusic Studio
Ro...@SirMusicStudio.com
301-585-4681
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
news:3EFF0D3A...@spamsucks.ionet.net...
> John LeBlanc wrote:
> > "Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> > news:3EFE8232...@spamsucks.ionet.net...
> >
> >
> >>...louder *is* better (with lossy).
> >
> >
> > Seems to me that's like saying horse shit tastes better than dog shit.
>
> Hmmm... For some reason, I thought I'd posted to rec.audio.PRO.
>
> I can get replies like that from the little kid next door.
Roger W. Norman
group is for professionals who happen to be in audio production. No big
deal. But it does put a damper on any real conversation about such things
as lossy compression when we're spending 10s and 100s of thousands of
dollars to get the most realistic and accurate recordings we can. At that
point, who wants to then degrade the audio by putting it through a lossy
process?
Don't get me wrong. As a tool, mp3 works fine. I've sent mixes out as mp3s
all the time to get client's opinions or to keep them abreast on how a mix
may be coming, but it's just that - a tool. And not a very good one,
either, except in it's usefullness in quick checks. There are far better
compression methods coming down the pike, including lossless. As the
internet infrastructure improves, better methods of making smaller file
sizes for audio will become readily available and cheap. But none of it is
our goal in recording. Perhaps on the distribution side, but that's outside
the studio.
--
Roger W. Norman
SirMusic Studio
Ro...@SirMusicStudio.com
301-585-4681
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
news:3EFF3CF6...@spamsucks.ionet.net...
> David Morgan (MAMS) wrote:
>
> > Did he tell you that REC stands for "recreational" (like this post) and
> > that PRO stands for PROduction? There aren't too many people here
> > who's record production habits revolve around MP3... it's a byproduct.
>
> Um, actually nobody told me anything about the meanings of the name...
>
> Was I was stupid (again) to ASSuME that "rec.audio.pro" stands for
> "recording.audio.professional" instead of
> "recreational.audio.production"? :)
>
> Myke
>
> P.S. Thanks for your WAV test suggestions. That's exactly the kind of
> advice I've been seeking as opposed to the stupid emotional outbursts
> like "troll somewhere else you f**kwitted sockpuppet" blah blah blah.
Roger W. Norman
thinking. No one here said anything of the sort. If your skin is that thin
you'll be in for some serious problems around here. If you really want to
discuss mp3s, then go to the Motion Picture Expert Group (MPEG - get it?)
and discuss it with the guys who developed the math.
Now before you complain that I'm trying to kick you off of this group, I
didn't say that. I said this conversation would be better to have with the
guys that developed the technology. If I wanted to talk about a Marshall
amp's sound, then I'd look to places like this, but if I wanted to talk
about the design I'd talk to Jim Marshall.
Lord Hasenpfeffer
>> I'm definitely here for the latter, though, sometimes my often too-bold
>> personality causes others to incorrectly assume the former (until they
>> see proof otherwise).
> What if some of these people have been studying said "proof" since it's
> inception and fail to agree with you?
Allow me to restate my position, Mr. Morgan, as my previous word choice
was flawed.
I'm definitely here in search of some facts and not to state that I am
without flaw. Sometimes my often too-bold personality causes others to
incorrectly assume that I am here to state that I am without flaw (at
least until they see proof that I am actually here in search of facts).
> And yes, it's almost guaranteed that outside of the psychoacoustical
> realm the basic principles of my hypothesis are all but unworthy
> of any practical contemplation at all.
>
> I'm glad you at least consider that this is a possibility. ;-)
I am here in search of facts not to state that I am without flaw.
> I have been ripping and encoding my library for over 2 years now.
> Good heavens !!! What on earth for ??
Professional reasons.
> You are going through all of this simply to place the audio
> onto a temporary and an essentially 'disposable' medium !!!
All Things Must Pass.
> Those are simply two different 'scales'. Volume is still a product created
> by the end user when referencing only these scales.
Very well.
>> One of several tasks I'm currently undertaking involves ripping and
>> encoding MP3s of every CD I own.
> 2100, eh?
Yes.
> Do you work ?? <g>
I enjoy what I do, so, you tell me! <g>
> I really hope your heart is in very good condition when MP3
> suddenly becomes a thing of the past, like Laser Disc, like 8-tracks,
> like Casettes and LPs are rapidly becoming, like 8-bit sampling,
> 4-bit video games and ISA slot cards. I give it until 2008.
All Things Muss Pass II.
> This screenshot looks relatively harmless, as if only peak normalization
> were applied,
To achieve the effect that's visible in that screenshot I elevated the
"level" of that file to, I believe, -10dBFS.
> but it does little to define your assumption that a "Batch"
> processing of RMS normalized files will guarantee you a decent 'relative'
> loudness between different sources. It does nothing to assure that you
> maintain the *integrity* (or what's left of it on newer material) of your
> original files.
You may be right. I may be crazy. But what I need to know is whether
or not my decision to amplify the level of that file saved any of the
frequencies in that recording from being discarded by the ATH-based
psychoacoustic filtering mechanisms of the MP3 lossy compression algorithm.
> I trust that you will be making some new ones.
If it serves the purpose for which this thread was created, yes.
> If you feel you are presenting some 'new' revelation to this group
> regarding what happens when one disregards certain file attributes
> in favor of RMS normalization solely for the sake of not having to
> reach for the volume control, what the other groups countered you
> with will pale in comparison.
I am here in search of facts related to the stated mission of this
thread as outlined in my initial post to this newsgroup. The
unfortunate events which transpired in the other newsgroups are the
result of my having prepared no mission statement prior to my
engagements therein. Provided that the original purpose of this thread
is maintained, I see no reason why the same unfortunate circumstances
which occurred there should reoccur here.
> At least one of those who offerred you the best logic is a regular
> contributor here and is considered by many as quite valuable. I notice
> you have already told him to "bug out"...
I don't recall. Nevertheless, while my online personality may often be
bold and free-spirited, I am rarely intentionally draw first blood. My
mission was initially undeclared in those newsgroups and you see how
things can go. My mission in this newsgroup is well-defined partly so
as to avoid repeating my past mistake(s).
> but this is, after all,
> usenet, and escape is futile... you will be assimilated. <g>
<g>
> The problem with this discussion thus far, in each place I reviewed any
> of your posts, is that you are trying to assimilate others.
My posts in the other newsgroups are snapshots taken in the heat of
battle following what I perceived to be a surprise attack. The first
victim in every war is the truth.
> I respect your apparently infinite curiosity of your teenage years.
By and large, 1964-1969 were phenomenal years.
> You appear to be quite intelligent, so you should be open to learning
> enough about audio to rationalize that what you are doing with RMS
> normalization to speed up your 'archiving to MP3' process is likely
> damaging the source material and doing little to improve the sound
> of the encoded files other than to make them loud.
My purpose with this thread is to discuss and to learn the truth
regarding whether or not adjusting the RMS level of a WAV file increases
the likelihood that more of its frequencies will survive the ATH-based
filtering processes employed by lossy compression schemes such as ATRAC
and MP3.
> Dozens of people have told you so, but you continue to search for
> some one or some place that will reassure you or be influenced by
> you for some reason.
When I can positively influence the lives of others I am pleased to do
so. Flame wars and unmoderated, undefined discussions on Usenet do not
positively influence the lives of others and can cause me to say and do
strange things online which some people, sometimes even I, do fully
comprehend.
> I don't quite follow that part, although I do see
> you opening up in recent posts that relate to encoding,
Those posts are inline with achieving the goal which defines my purpose
in being here in the first place.
> but remain in denial to those that pertain to standard, uncompressed
> audio - and that is where your errors may rest.
My goal lies not in producing the best uncompressed audio.
> I'm afraid that I didn't download the zipped file.
That's OK. They're older than these discussions anyway. There are no
records of the exact peak and level readings which existed at that time.
Recreations *may* be necessary.
> Well... you 'began' yesterday with a blanket statement and the reaction
> was somewhat less than hospitable, overall. I wish you better luck.
> Check the archives.... this group has covered the devastating effects
> of destroying dynamic range in favor of volume for many years.
I have never denied that my method of "RMS normalization" *could*
adversely affect the dynamic range of a recording. However, I have
certainly been accused of having done so in no uncertain terms. I have
in response to these claims posted screenshots in which I can see no
adverse effects to that end. Some agreed with me. Some vehemently
disagreed with me. Much of the disagreement was caused by false
assumptions that I had been shoving the RMS level up to Full Scale which
is ludicrous. None of all nonsensical bickering helped to clarify
anything related to my actual goal - which is largely my fault because
in the beginning I really was not familiar enough with the jargon
necessary to accurately describe my goal. If you return to the first
post in this thread, you will find a rather high degree of careful
wording which is the full measure of my desire to avoid my past mistakes
elsewhere while participating in *this* newsgroup.
I'm depending a lot on the "PRO" suffix in the name of this newsgroup to
assist me in achieving my goal for being here in the first place. I
reasoned that the *professional* thing to do would be to initially
explain my purpose for arriving. No flames. No false allegations
against my character. No off-topic posts. Hopefully a valid
conclusion. I am not a troll. If after an appropriate amount of time I
see that a conclusion either proving or disproving my hypothesis cannot
be reached, I will quietly vacate.
> You will find NO shortage of people here that *fully* comprehend the
> normalization process...
What I found before was that people who *fully* comprehend the
normalization process also failed to, at least initially, comprehend
that "normalize" is probably a gross misnomer for that particular
application. The lack of definition of terms early on caused a *lot* of
confusion, hence my desire here to first make certain that everyone
(including I) was "straight on the terms". That was the only purpose
for the "lecture" on demonstrating what the application I use can and
will do with a WAV. The purpose of said demonstration was *not* to
elighten others with regard to the concept of normalization itself.
> My advice would be to ask questions, not to walk into the Dean's
> office and tell him that you are about to begin teaching his classes.
What are the questions I should ask?
Lord Hasenpfeffer
> Ummm... Say there Mr. Timmerman, why don't you tell us how
> you *really* feel about this ? I mean, don't hold any punches....
>
> ;-)
His response is justified.
Peter Kaersaa
my...@spamsucks.ionet.net says...
> Actually I was thinking only in terms of CBR with both 128kb/s and
> 192kb/s bitrates in mind. With this you have certainly added a
> colourful new dimension to this discussion as far as I'm concerned. And
> that's the kind of relevant informational exchange I'm definitely hoping
> to create by way of this thread. I think a lot of people could really
> learn a lot from this discussion if the original topic can be
> maintained. I eagerly await the day that we all can bring it to a solid
> conclusion.
Take a CDDA rip a song encode it in VBR format.
Take the same song normalize it save it in the same VBR format.
Compare file size, the normalized file is bigger containing more
information.
Now does the Bigger VBR version of the song sound better ?
Different Yes, but better ?
You decide.
BTW the RAP 5 collection is a good source for testing various
compressing formats and bit rates.
Some of the more dynamic songs really sounds like shit, when
mp3 compressed at any bitrate.
--
/ Peter Kaersaa
Lord Hasenpfeffer
> will make diddly-squat difference in each resultant MP3. No frequencies are
> going to appear or dissappear with the two versions. If there is any
> difference, it would be hugely swamped by the damage done primarily in the
> MP3 coding in the first place.
>
> But what would I know ...
Frequencies from the WAV will not be included in the resultant MP3 if
their amplitudes are deemed to be "too quiet to be heard".
Is this flawed reasoning?
Myke
Lord Hasenpfeffer
Lord Hasenpfeffer
> You ARE posting to rec.audio.PRO and the reason you get this back is simply
> because it is PRO. We don't do mp3 for a living.
Well as a webmaster, I do!
Lord Hasenpfeffer
> Actually, were one to read the faqs, the mission statement is that this
> group is for professionals who happen to be in audio production.
My knowledge regarding MP3 production is directly related to a business
project on which I am working. Just because you don't do this too
doesn't mean that nobody else does.
*Pretty arrogant I'd say!*
Myke
Lord Hasenpfeffer
> Now before you complain that I'm trying to kick you off of this group, I
> didn't say that.
I have already stated quite cordially that if what happens in this
thread turns out to be a stupid repeat of the occurrences in other
similar past threads in other newsgroups, then I will vacate.
I never expect to find polite people on Usenet just like I never expect
to find people who are not on drugs. If a polite person appears I am
impressed with him because he is unusual in that respect. If a
drug-free person appears, I am equally impressed with him for the same
reason.
The *very first response* I received from this newsgroup included
unwarranted references to "horseshit" and "dogshit". And you guys think
that *I* can cause a "train wreck" by making my presence known in this
newsgroup.
With responses like most of the ones I've received so far, there is *no*
professionalism on display here at all
You misinterpret my experience-based effort to avoid conflicts by
perverting it into some misguided accusation of self-righteous pomposity
and "Lordship" and in the process turn this thread into a snowball
from hell.
Nearly every post here so far has been OFF-TOPIC. Yes I understand your
freedoms. Yes I understand my position here in relation to others.
Nevertheless, without adequate moderation communication cannot progress
- and so far I'm the only one who's shown interest in handling that
responsibilty.
Can we continue this discussion on-topic or not? If so, let's. If not,
goodbye. It's as simple as that.
I do not have time to dick around with off-topic bullshit and have done
so enough already anyway.
Myke
Vladan
<my...@spamsucks.ionet.net> wrote:
>Please stay on topic.
Never did it with Lossy, but louder must be better, or they wouldn't
make porn that way.
Vladan
www.geocities.com/vla_dan_l
www.mp3.com/lesly , www.mp3.com/shook , www.mp3.com/lesly2
www.kunsttick.com/artists/vuskovic/indexdat.htm
Lord Hasenpfeffer
> Take a CDDA rip a song encode it in VBR format.
> Take the same song normalize it save it in the same VBR format.
> Compare file size, the normalized file is bigger containing more
> information.
> Now does the Bigger VBR version of the song sound better ?
> Different Yes, but better ?
> You decide.
Thank you for keeping this discussion moving in a forward direction,
Peter. You are uncommon in this regard and for that you have my deepest
appreciation and utmost respect.
I have never used VBR because of the various things I have read about it
in the past which essentially turned me off to it (mostly
incompatibility issues and such, IIRC). However, if my hypothesis is
true, then that will be, for me anyway, a sound reason (no pun intended)
to consider changing my past encoding habits for the better.
Your test seems to indicate that you've already tried this. When you
say "Different Yes, but better ?" what exactly are you implying? It's
given that only individual opinions can define what is "better" but in
what way were they "Different" when you compared them in your own test(s)?
What possible alternative explanations do you suspect might also be able
to account for the larger filesizes? You really have me going at this
point but I'm afraid of making too hasty assumptions prior to conducting
my own tests as well.
> BTW the RAP 5 collection is a good source for testing various
> compressing formats and bit rates.
> Some of the more dynamic songs really sounds like shit, when
> mp3 compressed at any bitrate.
First of all, what *is* the RAP 5 collection?
Lord Hasenpfeffer
> On Mon, 30 Jun 2003 09:19:14 -0500, Lord Hasenpfeffer
> <my...@spamsucks.ionet.net> wrote:
>
>> Please stay on topic.
>
> Never did it with Lossy, but louder must be better, or they wouldn't
> make porn that way.
>
> Vladan
Ba-da-boom! Ba-da-bing! ;-D
Myke
Artie Turner
> With responses like most of the ones I've received so far, there is *no*
> professionalism on display here at all
I think I mentioned that most of the "pros" reading this NG don't care
about MP3s regardless of how they're encoded, and therefore they don't
reply at all. Some of them might rip MP3s for parties or casual,
jukebox-style listening, but I'd guess that speed and convenience would
be their primary objective, and adding the extra step of normalizing
would not be worth the extra time and effort.
Most of the ones who *are* replying view your question as comparing one
lossy recording to another louder, lossy recording.
Attempting to form some concensus here that your normalized/louder MP3s
are objectively "better" than another MP3 is futile, IMO. You'd get more
traction of you were comparing one hi-res format or technique to another.
But, if for your purposes as a programmer/web master/site designer you
find that your normalized MP3s are more effective, then, by all means,
go for it. I'm sure if you conducted your own listening tests and posted
the results there would be some interest.
AT
Lord Hasenpfeffer
> BTW the RAP 5 collection is a good source for testing various
> compressing formats and bit rates.
> Some of the more dynamic songs really sounds like shit, when
> mp3 compressed at any bitrate.
Something to remember, as much as you may read that I am in the business
of shoving RMS levels to Full Scale for the sheer sake of maintaining a
"louder is better" position, rest assured that I do not. :)
To my (admittedly limited) knowledge, I am unaware of any instance when
I have reduced the dynamic range of a recording by bumping up its RMS
level a few notches. I only tamper with RMS levels whenever they fail
to "meet spec" according my current purpose or taste. Many of the WAVs
I've ripped are already louder than I would ever consider making them be
if left to my own devices.
I would suspect that the "more dynamic songs" to which you have
referred are more along the lines of these songs which I personally
never adjust since "the music industry" has already done so for me (and
then some).
GIGO,
Erwin Timmerman
> Peter Kaersaa wrote:
>
> > BTW the RAP 5 collection is a good source for testing various
> > compressing formats and bit rates.
> > Some of the more dynamic songs really sounds like shit, when
> > mp3 compressed at any bitrate.
>
> I would suspect that the "more dynamic songs" to which you have
> referred are more along the lines of these songs which I personally
> never adjust since "the music industry" has already done so for me (and
> then some).
No. With "more dynamic" he means "more dynamic". Which means, very soft
parts as well as very loud parts. The RAP5 collection is a collection of
songs recorded by recording engineers on this newsgroup. Most of them
despise the loud louder loudest war that is going on in commercial releases,
and won't contribute to it. The way their recordings sound, shows it. Really
nice to listen to.
Read more about the collection at http://www.recaudiopro.net/
Erwin Timmerman
Erwin Timmerman
> BTW the RAP 5 collection is a good source for testing various
> compressing formats and bit rates.
> Some of the more dynamic songs really sounds like shit, when
> mp3 compressed at any bitrate.
Can you name a few tracks? Or maybe one at least? I'd like to try this myself.
Erwin Timmerman
Lord Hasenpfeffer
> Lord Hasenpfeffer wrote:
>
>> With responses like most of the ones I've received so far, there is
>> *no* professionalism on display here at all
>
> I think I mentioned that most of the "pros" reading this NG don't care
> about MP3s regardless of how they're encoded, and therefore they don't
> reply at all.
Yes you did mention that early on. However, no warning was provided
that most "pros" reading this NG see no viable professional use and/or
commercial potential for the format either! :) Then again, by no means
do I hold you accountable for any preponderances for presuppositions on
their part regarding the format; my purposes for using MP3s are quite
beyond the realm of the ordinary indeed.
> Most of the ones who *are* replying view your question as comparing one
> lossy recording to another louder, lossy recording.
And if that is the case, I do not understand why. I have done my best
to (1) define my purpose for being here quite clearly and (2) *try* and
keep this thread on-topic as much as possible so that such
misunderstandings of its reason for being could be avoided - as was the
case with disastrous effects in my previous, open-ended and generally
less- or undefined attempts in other newsgroups.
> Attempting to form some concensus here that your normalized/louder MP3s
> are objectively "better" than another MP3 is futile, IMO. You'd get more
> traction of you were comparing one hi-res format or technique to another.
I have attempted also to make it quite clear that this discussion isn't
strictly limited to MP3 encoding techniques but also to *all* ATH-based
lossy compression schemes - of which MP3 is only one. Nevertheless, I
see your point and feel compelled to conclude that you are correct in
your assessment.
> But, if for your purposes as a programmer/web master/site designer you
> find that your normalized MP3s are more effective, then, by all means,
> go for it.
Trust me, I'm *trying* to. :-)
> I'm sure if you conducted your own listening tests and posted
> the results there would be some interest.
Well, now that I have not just one but *two* excellent test suggestions
to work with, I think I'll first get some more sleep and then get back
to it and see what happens! :)
Thanks,
Myke
P.S. And Windows really *is* rebootylicious, btw!
Peter Kaersaa
my...@spamsucks.ionet.net says...
> I have never used VBR because of the various things I have read about it
> in the past which essentially turned me off to it (mostly
VBR has come a long way the last year or so, compared to the old
buggy Xing encoder years back it is a waste improvement.
> Your test seems to indicate that you've already tried this. When you
On a peak normalized song yes, I don't like to RMS normalize.
> say "Different Yes, but better ?" what exactly are you implying? It's
That you should be the judge, if it sounds better to You, then
do it. Just do It :)
> given that only individual opinions can define what is "better" but in
> what way were they "Different" when you compared them in your own test(s)?
Well I could have converted them back to Wav and peak normalized
the encoding of the original track to match the levels, and do
an A - B test, but i didn't.
> What possible alternative explanations do you suspect might also be able
> to account for the larger filesizes? You really have me going at this
The file size was 9.8mb (max peak 0dB) for the normalized and
9.4mb (max peak -5.5dB) for the original. I doubt I would be
able to hear any difference, it's a lossy format after all.
> point but I'm afraid of making too hasty assumptions prior to conducting
> my own tests as well.
My point is use your own ears.
Or You could try to contact the author(s) of your encoding
engine and ask what the encoder does at different levels.
> First of all, what *is* the RAP 5 collection?
--
/ Peter Kaersaa
Erwin Timmerman
> Roger W. Norman wrote:
>
> > Now before you complain that I'm trying to kick you off of this group, I
> > didn't say that.
> You misinterpret my experience-based effort to avoid conflicts by
> perverting it into some misguided accusation of self-righteous pomposity
> and "Lordship" and in the process turn this thread into a snowball
> from hell.
I think you confuse Roger with me, he never talked about your "Lordship". Even
more because most of the things you say in this message are a response to what
I wrote. So I'll reply as well.
> I have already stated quite cordially that if what happens in this
> thread turns out to be a stupid repeat of the occurrences in other
> similar past threads in other newsgroups, then I will vacate.
You also stated that you came in here to seek answers to your questions, and
after having gotten them you will vacate as well. So the end result for the
group (you vacating) will be identical either way.
> I never expect to find polite people on Usenet
Oh, there are plenty. Usually I myself am quite polite as well. Just not when
some people whom I highly respect and have been on this group for many many
years are ordered around by someone who comes barging in (and intends to leave
later on anyway no less). The fact that people type *anything* in response to
a question of you is worthy of praise, because that means they took you
seriously enough to take the effort to write something. It might not be to
your liking, it might not be your style, but they did it anyway. They have put
effort into you. Respect them for that.
> If a polite person appears I am
> impressed with him because he is unusual in that respect.
Many people are polite if you manage to press the right buttons. If you find
politeness to be unusual, maybe you haven't learned quite well how to push
those buttons correctly.
> The *very first response* I received from this newsgroup included
> unwarranted references to "horseshit" and "dogshit". And you guys think
> that *I* can cause a "train wreck" by making my presence known in this
> newsgroup.
Well, it was just a direct way to say he didn't like MP3. In other words: "why
are you wasting your time with this? It doesn't matter because it will sound
like shit anyway". You might not agree with him, but that doesn't make the
response any less on topic.
Furthermore, one response doesn't make the group, as you have already seen.
You've gotten plenty of other useful responses since the first one. Patience
can be a good thing.
> Nearly every post here so far has been OFF-TOPIC. Yes I understand your
> freedoms. Yes I understand my position here in relation to others.
> Nevertheless, without adequate moderation communication cannot progress
> - and so far I'm the only one who's shown interest in handling that
> responsibilty.
If you want moderation, go to a moderated newsgroup. This one is unmoderated.
The people that reside here seem to like that, otherwise they'd be on a
moderated group (quite a few of them are on both btw). People who are here
aren't here to serve you, they are just here. They will say something when
they feel like it, and say nothing otherwise. Trying to actively moderate a
thread yourself is futile, looks childish and arrogant, and only backfires, as
you have seen. Had you not responded so pedantic, this thread would have been
spared of many off-topic posts, among which at least mine. Even more so, the
excessive number of your own moderating posts is polluting the thread heavily
as well.
> Can we continue this discussion on-topic or not? If so, let's. If not,
> goodbye. It's as simple as that.
You say it as if it would be a loss to the group if you went. You said before
that your leaving is imminent anyway. A few days more or less will be hardly
noticed.
However, I find the topic itself interesting enough to contribute some
on-topic stuff as well.
First of all, everything is dependent on the psychoacoustic model applied. The
LAME encoder comes with source code. Since the notlame encoder works with
Linux, chances are pretty big that the source code of that is available as
well. As you're a coder yourself, it shouldn't be too hard to derive the
psychoacoustic model applied in this particular program. And then you know the
truth.
I still have to wonder about the "threshold of human hearing" they speak
about, because with playback systems that have a volume knob, it just doesn't
apply. If the CD is so damn soft that you can't hear a thing, turn up the
volume. Given that fact, what should I, if I were defining the model, pick as
the absolute threshold level?? It would just be a wild guess. Do they pick the
threshold level relative to the loudest peak? If that were the case,
normalizing a wave wouldn't do anything for getting better encoding. The
threshold level would be increased just as much as the quiet waveforms get
louder. Do they pick it relative to the RMS level of a song? If so,
"normalizing" it your way would actually only INCREASE the level under which
data was dropped, so it would actually hurt the encoding result. Do they pick
it relative to the loudness of the moment? That way the result wouldn't change
because with every low volume part the threshold would decrease as well. All
the more reason just to duck into the source code and find out for real. Or
maybe a message on
http://board.mp3-tech.org/w-agora.php3?site=agora&bn=agora_mp3techorg would
help.
Anyway, there are several ways to test the behaviour of an encoder. The test
suggested by Mark (if I remember correctly) is quite good. In addition to his
suggestions I would also keep the original "low volume" wav, and re-amplify
with the same amount you will re-amplify the encoded-decoded wav with, and
compare those as well for differences. That will eliminate any effect that has
been introduced by the lowering and highering in volume itself.
The applied threshold of hearing thing might be tested by generating a very
low volume sinus in a wave editor. Encode it, decode it. If it disappears then
it has been killed by the encoding. Maybe you have to run a frequency analyzer
instead of looking at the wave image, as a very low volume sinus won't look
pretty all by itself anyway. This way you can also test the relative masking
by generating a loud and a very soft sinus together, seeing what is left of
them after endecoding etc etc. That doesn't give you any answer as to how it
sounds, but it might give you an idea of what is actually happening.
BTW, I don't need praise for saying something on topic any more than I need to
be spanked for saying something off topic. Keep your reactions to just staying
on topic yourself (and not trying to make others do that) and the flames will
die by themselves.
Good luck with your quest.
Greetings,
Erwin Timmerman
Kurt Albershardt
>
> "Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message ...
>
>> To the quality of the source audio WAV, yes, it may indeed do nothing,
>> however, if "normalization", as such, helps to boost more of the
>> original WAV's frequencies into a more audible range, surely that would
>> cause those frequencies to be preserved rather than discarded by a
>> psycho-acoustic lossy compression scheme, no?
>
>
> You still seem to be saying that you are willing to inflict damage caused
> by RMS normalization in order to have seperate files that play back in a
> fashion that prevents you from having to change the volume during the
> playback - while at the same time implying that there must be some
> underlying set of frequencies, that if boosted far enough, will somehow
> suddenly reveal themselves or enhance the remaining audio so that
> they can be preserved on a lossy format.
Perhaps he should be looking at something like ReplayGain?
http://replaygain.hydrogenaudio.org/
Ben Bradley
>Chris Johnson wrote:
>> Because if you were, you'd be able to experiment with this sort of
> > thing yourself. I have.
>
>I've been offered a couple of good suggestions for tests which I can
>perform using particular applications which are currently installed on
>my Linux box. I just haven't found time to conduct them yet.
I have the same problem, I read and post to Usenet (what better
thing to do when drinking coffee than read about making better
coffee?) when I should be doing other, (hopefully) more productive
things.
I'm off to do something else now...
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3F0046BF...@spamsucks.ionet.net...
> > As before in r.a.t, a difference in around 3dB in tthe maximum peak which
> > will make diddly-squat difference in each resultant MP3. No frequencies are
> > going to appear or dissappear with the two versions. If there is any
> > difference, it would be hugely swamped by the damage done primarily in the
> > MP3 coding in the first place.
> >
> > But what would I know ...
>
> Frequencies from the WAV will not be included in the resultant MP3 if
> their amplitudes are deemed to be "too quiet to be heard".
>
> Is this flawed reasoning?
Regardless of level, the data compression sceme will operate in the same
fashion, removing what *it* deems as inaudible or uneccessary frequencies.
My impression is (I'm not really into MP3 outside of selecting the layer Codec
and dialing the server line before live IBCs - or streaming audio at home), that
one would merely be providing a different 'base line' from which the encoder
will still be performing the same function.
I like the analogy that Sony uses: If the sound of a gunshot and the sound of a
pin dropping were to be recorded at the same time, then essentially the sound
made by the dropping pin could and would be eliminated.
It would appear that it matters not at what RMS level these sounds entered
the encoder.... because of what they are, the same action would still be taken.
However, you may indeed be saving some reverb tails or cymbal decay, etc.,
but at the cost of potential damage done by operations performed to the raw
.wav files.
--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com
>
> Myke
John LeBlanc
"Ben Bradley" <ben_nospa...@mindspring.com> wrote in message
news:3f006f66...@newsgroups.bellsouth.net...
> I'm off to do something else now...
Might I suggest a web page that approximates this thread?
http://orangecow.org/pythonet/pet-shop.html
John
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3F0047AF...@spamsucks.ionet.net...
> Roger W. Norman wrote:
> > Actually, were one to read the faqs, the mission statement is that this
> > group is for professionals who happen to be in audio production.
>
> My knowledge regarding MP3 production is directly related to a business
> project on which I am working. Just because you don't do this too
> doesn't mean that nobody else does.
>
> *Pretty arrogant I'd say!*
Don't foget the emoticons. ;-)
David Morgan (MAMS)
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message news:3F003815...@spamsucks.ionet.net...
> David Morgan (MAMS) wrote:
>
> >> I'm definitely here for the latter, though, sometimes my often too-bold
> >> personality causes others to incorrectly assume the former (until they
> >> see proof otherwise).
>
> > What if some of these people have been studying said "proof" since it's
> > inception and fail to agree with you?
>
> Allow me to restate my position, Mr. Morgan, as my previous word choice
> was flawed.
The printed medium is a tough one for clear communication and the
comprehension of 'intent'. I'll gladly stop being 'picky' which might
impede your progress.
> > You are going through all of this simply to place the audio
> > onto a temporary and an essentially 'disposable' medium !!!
>
> All Things Must Pass.
You take my decidedly blunt humor very well. :-)
> > Do you work ?? <g>
>
> I enjoy what I do, so, you tell me! <g>
Very hard then, I would imagine. When you like what you do, it's not
really as though it's work.
> > I give it (MP3) until 2008.
>
> All Things Muss Pass II.
LOL
> But what I need to know is whether
> or not my decision to amplify the level of that file saved any of the
> frequencies in that recording from being discarded by the ATH-based
> psychoacoustic filtering mechanisms of the MP3 lossy compression
> algorithm.
After all this exchange, I am really not qualified to reply with certainty.
It is my understanding that the encoder will perform the same function
regardless of level. (I stand prepared to be corrected).
> I am here in search of facts related to the stated mission of this
> thread as outlined in my initial post to this newsgroup. The
> unfortunate events which transpired in the other newsgroups are the
> result of my having prepared no mission statement prior to my
> engagements therein. Provided that the original purpose of this thread
> is maintained, I see no reason why the same unfortunate circumstances
> which occurred there should reoccur here.
Wording. Your original post was a blaket statement addressing your
personal experiences, drawing your own conclusion there from, and
culminating by putting forth as fact that louder is better 'with lossy'.
It would be easy to misinterpret such an assertion as a challenge to
prove otherwise, rather than a search for confirmation. The printed
medium sucks sometimes.
> My posts in the other newsgroups are snapshots taken in the heat of
> battle following what I perceived to be a surprise attack. The first
> victim in every war is the truth.
It's surprising how many people have differing or even opposing viewpoints.
It's not necessarily an attack. Aggressive defense? <g>
> My purpose with this thread is to discuss and to learn the truth
> regarding whether or not adjusting the RMS level of a WAV file increases
> the likelihood that more of its frequencies will survive the ATH-based
> filtering processes employed by lossy compression schemes such as ATRAC
> and MP3.
Best of luck to you. I can help with uncompressed audio if needed, thus
my responses warning of potentially more damage than was worthy.
> The lack of definition of terms early on caused a *lot* of
> confusion, hence my desire here to first make certain that everyone
> (including I) was "straight on the terms". That was the only purpose
> for the "lecture" on demonstrating what the application I use can and
> will do with a WAV. The purpose of said demonstration was *not* to
> elighten others with regard to the concept of normalization itself.
>
> > My advice would be to ask questions, not to walk into the Dean's
> > office and tell him that you are about to begin teaching his classes.
>
> What are the questions I should ask?
I definitely left them above. ;-) Hang tough....
Bob Olhsson
<ma...@NOSPAm-a-m-s.com> wrote:
>It would appear that it matters not at what RMS level these sounds entered
>the encoder.... because of what they are, the same action would still be taken.
Exactly. All you accomplish by raising the level is to increase the
amount of ringing in the filters and, if you reduce the peak to average
ratio, decrease the amount of space available to hide the distortion
in. In the process you can simply cause more severe bandwidth
limitations as the codec lops off the high-end in order to meet the
desired bitrate.
Unfortunately a lot of people confuse the signal processing appropriate
for low-bit computer audio with that appropriate for low-bit lossy
coding. A pristine recording from a mike feed with no further signal
processing can sound stunningly good in an MP-3 encoding. As the
quality of the source drops, the quality of lossy-coded playback drops
faster.
--
Bob Olhsson Audio Mastery Recording Project Design and Consulting
Box 90412, Nashville TN 37209 Tracking, Mixing, Mastering, Audio for Picture
615.385.8051 FAX: 615.385.8196 Mix Evaluation and Quality Control
40 years of making people sound better than they ever imagined!
Peter Kaersaa
> > BTW the RAP 5 collection is a good source for testing various
>
Some time ago i did that, but lets see, CD 1 track 10 the esss
sounds from the choir. Listen to both the shape and the stereo
image of the choir.
Track 3 the piano, notice the clarity of the lower notes, and
the room in general.
Track 18 the guitar hand movement noise.
The ultimate test CD 2 track 1, listen to the background noise.
(!!! Don't blow up your speakers, the drum thing is loud !!!)
I encoded them with Lame, set at 160Kbit/best speed/quality.
Actually sounds fine at first but try to listen to the songs in
mp3 format a few time and then switch to the CD.
To me it's like Aahhh.. now I can relax.
Btw. The most noticeable way to hear mp3 artifacts are with
ocean waves or applause.
--
/ Peter Kaersaa
Mike Rivers
> > We don't do mp3 for a living.
>
> Well as a webmaster, I do!
That's something that you should have mentioned right up front, or
maybe you did and I missed it (and everyone else did too). Preparing
audio files for distribution over WWW is an art (and a little science)
in itself. I read an article a few years back about optimizing audio
for streaming distribution and it was quite involved. There were a
number of decisions to make, and there's no one method that sounds
good no matter how it's downloaded.
Not only is finding the right amount of dynamic range important, but
also finding the right bandwidth. The less you make the compression
algorithm do, the less it will change what you put into it. The trick
is to "degrade" what's going into the compression process in a way
that it's still acceptable before compression, and that the
compression won't make it significantly worse.
In other words, if YOU trim the frequency response on the low and
high ends, the algorithm won't have to do as much as if you feed it a
full bandwidth signal, so you'll be less surprised at what comes out.
Same with limited dynamic range. But to get the best sound at the
other end of the chain, you need to work both aspects, and you also
need to experiment with each song. No one formula works best for
everything.
--
I'm really Mike Rivers - (mri...@d-and-d.com)
Geoff Wood
> Roger W. Norman wrote:
>
> > You ARE posting to rec.audio.PRO and the reason you get this back is
simply
> > because it is PRO. We don't do mp3 for a living.
>
> Well as a webmaster, I do!
This isn't the contents of your 2100 CDs that you are putting up on the web,
by any cahnce ? What's the URL ?
geoff
Geoff Wood
> David Morgan (MAMS) wrote:
>
> >> I'm definitely here for the latter, though, sometimes my often too-bold
> >> personality causes others to incorrectly assume the former (until they
> >> see proof otherwise).
>
> > What if some of these people have been studying said "proof" since it's
> > inception and fail to agree with you?
>
> Allow me to restate my position, Mr. Morgan, as my previous word choice
> was flawed.
>
> without flaw.
You have be told, and directed to verification of the truth, However this
does not sit well with your previously-held misconceptions of 'the truth',
and although you have modified your stance slightly wrt a few aspects, you
steadfastly refuse to absorb the knowledge you have been made a ware of.
Switching newsgroups won't change the 'truth'. You are just as off the wall
here as you are in rec.audio.tech, though the people here have less
tolerance for stupidity than in other places, as you will doubtlessly find.
But not from me.
geoff
Geoff Wood
"Lord Hasenpfeffer" <my...@spamsucks.ionet.net> wrote in message
> To my (admittedly limited) knowledge, I am unaware of any instance when
> I have reduced the dynamic range of a recording by bumping up its RMS
> level a few notches. I only tamper with RMS levels whenever they fail
> to "meet spec" according my current purpose or taste. Many of the WAVs
> I've ripped are already louder than I would ever consider making them be
> if left to my own devices.
Have yo not yet read and understood the intsructions for yor favorite Linux
'Normalise' application ?
geoff
Bob Olhsson
wrote:
> The trick
>is to "degrade" what's going into the compression process in a way
>that it's still acceptable before compression, and that the
>compression won't make it significantly worse.
I read that too but my experience has been the opposite. This WAS true
of the very earliest streaming applications but it doesn't seem to be
true of the better codecs.
dmi...@spamblock.demon.co.uk
> dmi...@spamblock.demon.co.uk wrote:
> > I would be somewhat surprised that any current CD was not NORMALISED
> > to somewhere around (just below) 0dbfs.
> Yes, as that has apparently been "all the rage" for several years
> running now. Actually, *abuse* of the 0dBFS threshold is really "all
> the rage" these days. I personally am only interested in "normalizing"
> older, quieter CDs mastered during the stone and middle ages before
> encoding them as MP3s so that my final files sound somewhat "modern" to
> my ears and not all tinny and weak when juxtaposed with MP3s encoded
> from more recently mastered CDs.
I would be somewhat surprised to learn that any mainstream commercial
release has ever gone out without being normalised (using the strict
definition of the term) to somewhere within 1db of FS. Given this, it
sounds like you are actually compressing or limiting to raise the RMS
level.
> In no way am I interested in brutally forcing any RMS levels to go
> "through the roof". Anywhere from strictly normalized to "slightly hot"
> levels (depending on genre) are all that I seek.
Well, if you are going for 'slightly hot' then you must be limiting or
clipping somewhere to keep within the actual hard limit of 0dbFS, so
irrespective of what the tool calls itself this cannot be said to be
normalising. I would also observe that a track which is quieter will
tend to sound 'tinny & weak' when compared to a louder track even without
being coded to MP3 first, it is just how humans work. I am sure that lots
of people here have stories of finding the 'extra something' at the end of
a long mix session by tweaking the control room monitor volume up a db or
two for the benefit of a client.
Many times I've
> encountered WAVs from CDs that are, according to my standards, mastered
> too loudly. When this is the case, I _do_ simply "rip-and-encode" the
> unmodified, original WAVs (contrary to what dipshit said of me earlier
> when he said I seek to indiscriminately normalize every CD I own). If
> that were the case, I'd end up making half of what's in my library
> *quieter* than it is on the original CDs, not louder! :)
Making them quieter does not undo the compression used to make them
boring to start with!
> If normalization is _needed_ for a "mix-CD" sporting tracks by various
> artists from various sources, I will "normalize" all the tracks
> individually to bring them to a common loudness, across-the-board.
Well, this will get all the tracks to the same peak level, but they are
unlikely to all sound the same volume, as to a first approx, we hear RMS
not peak. What you probably want to do is (being careful that no
intermediate file clips (32 bit signed int may be a good intermediate
format for this)), is to equalise the average RMS levels (which will make
the peak level end up all over the place), then normalize the entire
collection to put the peak level into range. Your choice of averaging
function matters if you want good results.
> However, depsite its name, the "normalize" application I use is capable
> of doing more than just "textbook normalization" if I tell it to do so.
> This has been a major stumbling block for me when I've previously
> attempted to discuss its behaviour amongst others who've not used nor
> even heard of it before.
Do you have a URL for a tarball?
"Normalize" can be made to limit the peaks
> when instructed by the user to boost an RMS level beyond the textbook
> normalization level. One of the guys over in the other newsgroup
> suggested that we call this "limitizing" because there is no other,
> readily available, predefined textbook term to describe it.
Sounds like fast attack hard knee limiting (Possibly with compression) to
me? Have you seen that compressor Dyson wrote when he was at free BSD?
It is excellent for this sort of thing.
> > Further the phycoacoustic model IIRC derives 'too quiet' threshold
> > from the RMS level of the signal, not an absolute threshold (IE at
> > some frequency it may be 30db below curent RMS level, NOT 40db below
> > 0dbFS).
> Mmm-HMMM.... Now *that's* something of which I was not previously aware.
> If that's true then my hypothesis for boosting amps to save freqs may
> indeed be fatally flawed.
> Are you sure you're not thinking of frequency masking when you say this?
Yes, to a first approx, frequency masking weights towards energy in
nearby bands when calculating thresholds, what I am thinking of looks
at overall energy.
> > It is worth noting that whatever you do to the input data, it is only
> > possible to fit a fixed amount of information into the output
> > stream,
> Now, I do understand that, however, I haven't contemplated it very much.
It is worth working thru the implications of this as it really brings home
the absence of a free lunch.
> > if you force the encoder to include more frequency data (higher
> > resolution or more bands active, then time or ampletude resolution
> > MUST necassarily suffer).
> Excellent information. By time resolution you mean in that the file
> would have to be made to play slower in order to accommodate the
> increased amount of data?
No, not slower (which would imply a higher effective bit rate), but
that the information about when something happens may have to be stored
less precisely.
> Also, would not such effects be greater at lower bitrates than at higher
> ones?
All the compromises apply more at lower bitrates!
> And if so, would this not mean that my hypothesis would actually become
> more appropriately applied towards higher bitrate MP3s than at lower ones?
> Because I'm guessing here, by reversing your logic, that a large enough
> bitrate could eventually be employed which would cause the encoder to
> either "pad the file with zeroes" or store the additional data depending
> on the normalized status of the WAV being encoded.
Depends on if your hypothesis holds at all, I am yet to be convinced
that anything beyond psycological effects are at play here (Louder is
usaually perceved as better).
Consider that a 'perfect' data compression tool would simply store the
gain used in the normalisation once (after all it does not change during
playback), thus normalising has little effect on the amount of
*information* in the .wav file.
> And if that's true, what bitrate may I be talking about?
There are a few lossless wav file compression tools around, some of them
are even reasonably good, find one then see how much it can reduce the
size of a typical wave file. This will give you some idea of how much
data is actually redundant in a wav file and of how much *information*
is required to represent that file.
It will be program dependent, a file containing a single 1Khz tone can
be losslessly represented in very few bits, a thrash metal gig will take
rather more (but why anyone would bother....).
Regards, Dan.
--
** The email address *IS* valid, do NOT remove the spamblock
And on the evening of the first day the lord said...........
.... LX 1, GO!; and there was light.
Chris Johnson
Bob Olhsson <o...@hyperback.com> wrote:
> Unfortunately a lot of people confuse the signal processing appropriate
> for low-bit computer audio with that appropriate for low-bit lossy
> coding. A pristine recording from a mike feed with no further signal
> processing can sound stunningly good in an MP-3 encoding. As the
> quality of the source drops, the quality of lossy-coded playback drops
> faster.
I've also found that to be true- and interestingly, have found
advantages to using high-res sources or fancy noise-shaped dithered 16
bit- though really that will work better with encoders which are
discarding information near Nyquist, because again it's a tradeoff-
unless the encoder disregards stuff over 20K (which seems common), it'll
waste information trying to encode the zone where error information is
dumped, namely very high-frequency noise.
In a way this is an argument for not doing any processing, which
would give the low-amplitude MFSL version the edge.
I think what's really going on here is this: Myke has distinct
preferences for a particular RMS/peak ratio. This is not unusual at all
as far as I'm concerned- I like about 15 db of peaks myself, Myke wants
only 10. There's a distinct sound associated with both extremes- well
mixed audio with very high crest factor is liable to sound particularly
open and stark (digital mixing can be prone to this effect if you're not
using a lot of compression) and very low crest factor sounds congested
and dense.
Super-high crest factor can sound almost magical and unreal, not even
physically there, more like an imagined fantasy of what music might be
like, easily disrupted by unwanted ambient noise.
Super-low crest factor can sound as solid and dense as a brick,
dominating over any other sound, crude and unvarying.
Dark Side happens to be very high on the crest factor, even higher
than I like it, so it's understandable that Myke doesn't like it the way
it was recorded, and wants to change it.
In a way this is a glimpse of the future- one day, you'll pick out a
recording that you think sounds nice (say, "Low Spark of High Heeled
Boys", with its high crest factor and extended top-end) and your trusty
iMusicBox computerized gizmo will, on the fly, tinker with every single
track on your random playlist to 'map' roughly the same frequency
balance and dynamic behavior onto everything you hear. The hard part is
putting crest factor back into stuff that's been peak limited, but this
isn't impossible, it's just impossible to do well. Something like Dark
Side of the Moon in your playlist, being a very well recorded piece with
excessive fidelity and crest factor, will turn out perfectly suited to
this new world, because it will be malleable into any possible desired
'sound' and will sound its best whatever you do to it.
Which is of course what 'mastering' is about, even if people don't
always use it for that purpose anymore. Done properly, it's the craft of
delivering recorded music in such a way that it's able to adapt to
whatever playing conditions you impose on it- and we have yet to see the
full extent of what that will mean. Boomboxes with EQs or fake reverb
presets are barely scratching the surface of what you can do.
Chris Johnson
Chris Johnson
Lord Hasenpfeffer <my...@spamsucks.ionet.net> wrote:
> In no way am I interested in brutally forcing any RMS levels to go
> "through the roof". Anywhere from strictly normalized to "slightly hot"
> encountered WAVs from CDs that are, according to my standards, mastered
> too loudly. When this is the case, I _do_ simply "rip-and-encode" the
> unmodified, original WAVs (contrary to what dipshit said of me earlier
> when he said I seek to indiscriminately normalize every CD I own). If
> that were the case, I'd end up making half of what's in my library
> *quieter* than it is on the original CDs, not louder! :)
Oh, this _so_ makes me want to get in touch with whoever's writing
Myke's software and set them up with a nice sine-based transfer function
for leaning out those over-loud CDs. It wouldn't take much, and I've
been using that in my own software for eons, it's just that when I
record stuff I _don't_ squash it so there's no point to trying to
unsquash it, so I just don't use that functionality. Don't listen to
modern squashed music either- praps I should.
All you need to do is a simple function that operates like a peak
expander with essentially no 'knee': slowest possible transition between
FS samples (untouched) and near-zero samples (attenuated somewhat).
Doesn't even have to incur any latency. Though admittedly doing this
properly is math-intensive compared with peak limiting: but since CDs
only have 65535 data values you can just make a lookup table, no problem
there.
I'm sure someone else will hit on this at some point. More than happy
to help out if asked. I code GPL so working with open-source guys suits
me.
Chris Johnson
Chris Johnson
Erwin Timmerman <erw...@stack.no> wrote:
> No. With "more dynamic" he means "more dynamic". Which means, very soft
> parts as well as very loud parts. The RAP5 collection is a collection of
> songs recorded by recording engineers on this newsgroup. Most of them
> despise the loud louder loudest war that is going on in commercial releases,
> and won't contribute to it. The way their recordings sound, shows it. Really
> nice to listen to.
For what it's worth my entry is relevant to this crest factor
discussion. The version of "Take A Number" on the RAP5 collection has
essentially a scratch vocal, and was processed using phase-reversed
doctored sidechain compression to crank up the crest factor VERY high,
in fact higher than I ever use anymore.
I was trying to show what the opposite extreme could sound like-
think I more or less succeeded within the limitations of what I had to
work with. I believe that in this case, many people would agree with
Myke that this version of "Take A Number" would make a better mp3 if it
was peak limited and its crest factor was reduced- again, because it was
intentionally exaggerated wildly for use as an example.
The current version happens to be louder and denser, with a different
snarly-punkish lead vocal that's about 32 times louder :)
Chris Johnson
Chris Johnson
"David Morgan \(MAMS\)" <ma...@NOSPAm-a-m-s.com> wrote:
> Average RMS amplitude is determined by the entire file content *selected*.
> Some sections of the song may be higher, some parts lower... it is not a
> constant, therefore the term "average".
>
> You would use "Peak" normalization to accomplish the goal you describe
> above. Why? There could be specific segments of the program that
> are actually much higher in RMS than the "average" would indicate. It
> is these sections that would suffer the most traumatic damage.
This is one hell of a good point, particularly for someone like me
who hacks on DSP software a lot. When I read this whole topic it's as a
closet sympathizer with Myke: why shouldn't you be able to seamlessly
cast all possible music into any given 'mold' that you happen to like?
I write batch-processing MacOS 8 audio mastering software (no, not
OSX- no, not PC...) and one of the tricks I've done well with is
metering. I render the song visually across sort of a 'strip chart',
with peaks represented as black speckles, and RMS loudness over time as
a gray region that goes higher on the chart and is rendered darker, the
closer it is to -0dB RMS.
As such, I've seen exactly what David is describing, _visually_.
Bass notes produce very high RMS when they occur. Certain vocal
sounds, like sustained vowel sounds well-recorded, can produce very high
RMS. The funny thing is, when music is very dynamic, these
characteristics can express themselves and the RMS is all over the map
and varying wildly, but when music is squashed and peak-limited, not
only does most of the RMS come up, but these moments of high RMS come
down. You don't see brief bursts of supersuperhigh RMS loudness, instead
you just get an unvarying blare. Bass in particular loses its capacity
to be unexpectedly loud and present, and the vocal resonances also stop
being any different from the background sound, and sort of merge with
the music.
Again, this is not based on listening opinions (though that IS what
happens), I'm reading this stuff off a computer-generated, calibrated
chart of RMS/peak over time. There most certainly are isolated sounds
that produce unusually high RMS, and that is most certainly lost when
you peak limit.
Chris Johnson
Chris Johnson
Erwin Timmerman <erw...@stack.no> wrote:
> I still have to wonder about the "threshold of human hearing" they speak
> about, because with playback systems that have a volume knob, it just doesn't
> apply. If the CD is so damn soft that you can't hear a thing, turn up the
> volume. Given that fact, what should I, if I were defining the model, pick as
> the absolute threshold level?? It would just be a wild guess. Do they pick the
> threshold level relative to the loudest peak? If that were the case,
> normalizing a wave wouldn't do anything for getting better encoding. The
> threshold level would be increased just as much as the quiet waveforms get
> louder. Do they pick it relative to the RMS level of a song? If so,
> "normalizing" it your way would actually only INCREASE the level under which
> data was dropped, so it would actually hurt the encoding result. Do they pick
> it relative to the loudness of the moment? That way the result wouldn't change
> because with every low volume part the threshold would decrease as well. All
> the more reason just to duck into the source code and find out for real. Or
> maybe a message on
> http://board.mp3-tech.org/w-agora.php3?site=agora&bn=agora_mp3techorg would
> help.
FWIW, ATH is a lookup table based on Fletcher-Munson loudness
thresholds and relative to digital full scale, not to anything else.
So it's basically a hard, multiband noise gate, at a higher threshold
for very low and very high frequencies.
My take on this is pretty much that if ATH had anything to do with
it, changing the ATH level would be just as good. I hacked a version of
a LAME shared library to be able to do that- I think the functionality
eventually made it into the original encoder. I guess notlame is more of
a just-run-it project and doesn't give users that level of control, but
if it does, problem over.
Unless, of course, Myke (hey, is this Cyco Myko of Suicidal
Tendencies?) simply doesn't like over-high crest factors, which is
totally understandable. He doesn't like overly LOW crest factors either,
and has said so. I'd say the guy is just sensitive to crest factor.
Given what I already know, if I was doing mastering for him I could give
him stuff he loved every time, just by zeroing in on the crest factor he
likes best.
That's useful to know, that people do have preference about this even
if they don't know what it is they're hearing. I could have told you
that, but Myke becomes a real-world example backing me up :D
Chris Johnson
Lord Hasenpfeffer
Emoticons cannot describe... ;-)
Myke
Lord Hasenpfeffer
<snip>
> First of all, everything is dependent on the psychoacoustic model applied.
Without knowledge of everything, however, I was afraid to give voice to
that even though it was my initial assumption since it is such an
integral part of MiniDisc recording with which I am more experienced
than with MP3.
> As you're a coder yourself, it shouldn't be too hard to derive the
> psychoacoustic model applied in this particular program.
My strength as a coder lies in PHP web-applications development. My
experience with C++ is limited to that which I have done only with
Borland C++ Builder 5 which does not apply.
> I still have to wonder about the "threshold of human hearing" they speak
> about, because with playback systems that have a volume knob, it just doesn't
> apply.
Unless I am mistaken, frequencies discarded during encoding remain
absent during playback regardless of volume. This is central to what I
am seeking to avoid if it is indeed possible - hence my hypothesis.
> If the CD is so damn soft that you can't hear a thing, turn up the
> volume.
This is true for CD audio, however, turning up the volume of an MP3 will
not restore the frequencies which have been lost during the encoding
process. I've stated many times already that this discussion with
relation to CD audio is practically irrelevant. Practically I say
because I am aware of the anal-retentive aspects of increased amplifier
noise, etc., which come into play and will be undoubtedly be cited if I
do not. <g?>
> Given that fact, what should I, if I were defining the model, pick as
> the absolute threshold level?? It would just be a wild guess. Do they
> pick the threshold level relative to the loudest peak? If that were
> the case, normalizing a wave wouldn't do anything for getting
> better encoding. The threshold level would be increased just as
> much as the quiet waveforms get louder.
Correct. If this is the case then the goal of disproving my hypothesis
will have been achieved.
> Do they pick it relative to the RMS level of a song?
I don't know. By coming here I was hoping to come into contact with
somebody who already knew the answers to questions like these which
would, in turn, eliminate so much opportunity for flaming and
discussing. This is where contacting the developers of the codecs will
obviously be required, since most in this forum obviously populate their
professional daily lives pursuing other goals.
> If so, "normalizing" it your way would actually only INCREASE the level under
> which data was dropped, so it would actually hurt the encoding result.
Correct, AFAICS.
> Do they pick it relative to the loudness of the moment?
Obtaining the answers to questions like this is undoubtedly the
*fastest* track to proving my hypothesis as being either true or false
and actually serve to further clarify its goal. The tests which have
been suggested are excellent and useful in the practical sense, however,
answers to these questions are the only way I see of being able to
obtain proof in the theoretical.
Mike Rivers
> > The trick
> >is to "degrade" what's going into the compression process in a way
> >that it's still acceptable before compression
> I read that too but my experience has been the opposite. This WAS true
> of the very earliest streaming applications but it doesn't seem to be
> true of the better codecs.
I guess progress is good. It allows people who don't understand the
problem to not make mistakes as big as they used to.
Lord Hasenpfeffer
> FWIW, ATH is a lookup table based on Fletcher-Munson loudness
> thresholds and relative to digital full scale, not to anything else.
OK, it has a name. That's good. I was not aware of that. Gives me
something with which to Google. Thanks.
> So it's basically a hard, multiband noise gate, at a higher threshold
> for very low and very high frequencies.
This sounds very good to me on a purely emotional level because while
the goal is to either prove or disprove my hypothesis, my *hope* is that
it is proven true. But I cannot allow my hopes to cloud my judgement.
Once the hypothesis is confirmed or denied and/or modified as necessary
to accommodate variables involving VBR and things like that, I can get
back to sending my levels to the moon. Just kidding. Actually I'd like
to spend more time examining just how much limiting, etc. I am causing
to occur in my pop/rock WAVs. Even though I'd prefer not to discuss it
within the context of this thread, it *is* a matter of great interest
and concern to me. I believe some either are or have tried to equate my
headstrong desire to stay on-topic with wanton disregard for all other
topics. Such is not the case. To everthing there is a season.
> Unless, of course, Myke (hey, is this Cyco Myko of Suicidal
> Tendencies?)
Yeah, I'm the one who penned the line "tire tracks all across her head"
some years back! :)
Richard Kuschel
>Lord Hasenpfeffer wrote:
>
>> Does anybody in this NG have experience with attempting to create
>> higher-quality MP3s or do they all just rip-and-encode like everyone
>> else, taking whatever they can get from _unmodified_ WAVs extracted via
>> DAE?
>
>All MP3 encoders are not created equal. You might find that some
>encoders sound better than others for a given .wav., but generally
>speaking, louder is better regardless of what's being evaluated.
>
>When I make MP3s of my material to send via email or website, I just use
>the program material "as is." I don't recall any serious evaluations of
>MP3 encoders or encoding levels here on RAP.
>
>AT
>
>
>
I've heard some MP3's that were made with variable bit rateat the high quality
setting and vas really surprised at how good they sounded.
There was a little deterioation of reverb tails , but that waas about it. No
glaring artifcts.
Richard H. Kuschel
"I canna change the law of physics."-----Scotty
Lord Hasenpfeffer
>>> We don't do mp3 for a living.
>> Well as a webmaster, I do!
> That's something that you should have mentioned right up front, or
> maybe you did and I missed it (and everyone else did too).
Actually, no, I had not mentioned it previously because I don't
understand why anyone would care to know.
> Preparing audio files for distribution over WWW is an art
> (and a little science) in itself.
Yes it is. One that I attempting to master. This is one reason why I
didn't care so much about the things I was saying with regard to the
MFSL CD edition of Pink Floyd's "Dark Side Of The Moon". I'm not trying
to master a superior edition of that CD. I'm trying to learn how to
make "better" MP3s than I knew how to make during 2001 and most of 2002.
I first noticed about 8 months ago that by "normalizing" the RMS levels
of my WAVs to -10dBFS prior to encoding them, I was able to create files
which to my ears sounded dramatically better that those I'd made before.
I understand the simple concept that most people equate louder with
better, however, given the ATH-based aspects of lossy compression
schemes, I feel essentially compelled to question the possibility that
*just maybe* there's more to this particular picture than simple "louder
is better" mentality. If there is not, then fine; life goes on. But if
there is, I want to be able to capitalize on any advantages I may be
able to discover from both others' knowledge on the subject and by
conducting my own tests.
At this particular time, whether or not I've ignorantly and brutally
slaughtered a Pink Floyd classic cow (A.H.M.?!) isn't particularly high
on my list of concerns. That doesn't mean I think I'm superior to
masters in the field. It simply means right now, that such discussions
are beyond the scope of what I'm actually trying to learn and to
accomplish. There'll be plenty of time to learn everything else I
currently don't know and/or fully understand about the sanctity of the
original dynamic range tomorrow. As for today, I don't have a mental
folder open and available to store that kind of information so, garbage
in, garbage out.
> I read an article a few years back about optimizing audio
> for streaming distribution and it was quite involved.
Source???
> There were a number of decisions to make, and there's no one method
> that sounds good no matter how it's downloaded.
I believe it! :)
> Not only is finding the right amount of dynamic range important, but
> also finding the right bandwidth. The less you make the compression
> algorithm do, the less it will change what you put into it. The trick
> is to "degrade" what's going into the compression process in a way
> that it's still acceptable before compression, and that the
> compression won't make it significantly worse.
Very interesting. Noted. Thanks.
> In other words, if YOU trim the frequency response on the low and
> high ends, the algorithm won't have to do as much as if you feed it a
> full bandwidth signal, so you'll be less surprised at what comes out.
I can do that with sox. Kewl. Of course, everything sounds like a
telephone then but oh well. The soundcards and speakers used by the
computers of most people I know these days sound about that bad if not
worse anyway.
I blame the generally lo-fi state of "computer speakers" these days for
basically throwing the public's world of audio backwards by about 35-40
years to when AM transistor radios ruled the roost. I suppose by
trimming the highs and lows prior to encoding, this wonderfulness is
addressed.
> Same with limited dynamic range.
Makes sense.
> But to get the best sound at the other end of the chain, you need
> to work both aspects, and you also need to experiment with each song.
> No one formula works best for everything.
I have learned as a result of having worked and played extensively with
AM radio -> MiniDisc -> MP3 recordings that mono MP3s (1) consume only
half as much storage and bandwidth as stereo MP3s and (2) possess the
same sample rate of stereo MP3s encoded at twice their bitrates. In
other words, stereo 128KBps MP3s and mono 64KBps MP3s both employ
44.1KHz sample rates. Therefore, if stereo is not required, switch to
mono and cut the bitrate in half. In other words, mono preserves the
fidelity (aside from the obvious difference) and doubles one's sending
and receiving capacity with no additional consumption of resources in
terms of storage and bandwidth.
Furthermore, it is also my experience-based belief that, "normalizing"
mono AM radio WAVs (to RMS -12dBFS) which have been transferred from
MiniDisc masters *definitely* improves the perceived quality of the
resultant MP3s. But is that just because louder is better in general?
Or because louder is better with lossy? Or both? (I believe both.)
Lord Hasenpfeffer
> This isn't the contents of your 2100 CDs that you are putting up on the web,
> by any cahnce ? What's the URL ?
*Ha!* :-)
Lord Hasenpfeffer
>
>> The trick
>> is to "degrade" what's going into the compression process in a way
>> that it's still acceptable before compression, and that the
>> compression won't make it significantly worse.
>
> I read that too but my experience has been the opposite. This WAS true
> of the very earliest streaming applications but it doesn't seem to be
> true of the better codecs.
I'm glad to read that.
WBRW
> release has ever gone out without being normalised (using the strict
> definition of the term) to somewhere within 1db of FS.
How about this:
Ace Of Base / The Sign (Arista, 1993)
Track 1
Filename F:\All_That_She_Wants.wav
Peak level 60.8 %
Track 2
Filename F:\Don't_Turn_Around.wav
Peak level 77.6 %
Track 3
Filename F:\Young_And_Proud.wav
Peak level 78.6 %
Track 4
Filename F:\The_Sign.wav
Peak level 56.2 %
Track 5
Filename F:\Living_In_Danger.wav
Peak level 81.7 %
Track 6
Filename F:\Dancer_In_A_Daydream.wav
Peak level 71.0 %
Track 7
Filename F:\Wheel_Of_Fortune.wav
Peak level 88.5 %
Track 8
Filename F:\Waiting_For_Magic_(Total_Remix_7").wav
Peak level 80.3 %
Track 9
Filename F:\Happy_Nation.wav
Peak level 81.1 %
Track 10
Filename F:\Voulez-Vous_Danser.wav
Peak level 76.5 %
Track 11
Filename F:\My_Mind_(Mindless_Mix).wav
Peak level 89.1 %
Track 12
Filename F:\All_That_She_Wants_(Banghra_Version).wav
Peak level 56.6 %
"The Sign" was Ace Of Base's biggest U.S. hit ever, reaching #1 on the
Billboard chart in 1994 -- yet, on the album, it actually has the
LOWEST peak level of all the tracks, only hitting 56.2% of full scale,
or -5.0 dB!
And with regard to MP3 compression, an MP3 file can be LOSSLESSLY
NORMALIZED on a frame-by-frame basis in 1.5 dB increments. The actual
compressed data is NOT changed -- only an ancillary "loudness
scale-factor". You can even LOSSLESSLY add fade-ins and fade-outs to
MP3 files, by changing the scale-factor on a frame-by-frame basis.
(The "MP3Trim" utility can do that.) So, if you're concerned about
this whole issue, simply pre-normalize the incoming WAV file by a
multiple of 1.5 dB, and then use a utility like "MP3GAIN" to
losslessly normalize it back down to the originally intended level.
In fact, you should be doing this anyway with any MP3 files you might
download, as today's over-compressed pop music often drives MP3
decoders into extreme amounts of clipping unless the level of the MP3
file is reduced to a "safe" value. (The MP3 ENCODING is fine -- it's
just the DECODER that adds its own extra clipping upon playback.)
I've seen downloaded MP3s where the file had to be knocked down by 6
dB, just to get it below the level of clipping during playback!
Lord Hasenpfeffer
> Have yo not yet read and understood the intsructions for yor favorite Linux
> 'Normalise' application ?
Yeah, here's one thing that piqued my curiosity.
--peak Adjust using peak levels instead of RMS levels. Each file will
be adjusted so that its maximum sample is at full scale. This
just gives a file the maximum volume possible without clipping;
no normalization is done.
What was that last line again?
"no normalization is done."
That's what I thought it said.
OK,