Those amazing audiophiles!
Arny Krueger
was transcribed from a SACD, and two were modified by means of added
distortion. They are *not* in any particular order:
Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
If you have a little time, you might want to listen to them and figure out
which one sounds the best to you, and report back.
Where "Those Amazing Audiophiles" come in is that these files were
auditioned by dozens of audiophiles on a forum someplace, and reported back
their results for a survey.
They preferred the *most distorted* file. Which one is it, and do you like
it the most, too?
Ethan Winer
> They preferred the *most distorted* file. Which one is it, and do you
like it the most, too? <
Why am I not surprised?
A lot of pros prefer the distortion analog tape adds. Hell, *I* like a
little added distortion sometimes on some tracks or an entire mix if the mix
tends toward sparse. I believe this is what some recording engineers call
"glue" to make their mixes gel a bit. But at least I know it's an effect,
not superior fidelity! So no toob amps for me, thanks. :->)
--Ethan
Bill
nothing except that it was a good sounding, mellow little piece. I'll
chime in when I can listen more seriously.
Bill Baka
videochas www.locoworks.com
clear and open, and three sounded muddy. So, what's the answer? Or is
"so what" the answer?
Arny Krueger
message
news:1163704630.6...@h48g2000cwc.googlegroups.com
> One and two sounded almost the same to me, one sounded
> (perhaps) more clear and open, and three sounded muddy.
Close but no cigar.
> So, what's the answer?
I thought I'd keep the right answers quiet until a few more people posted.
> Or is "so what" the answer?
Several of us who have tried it, got all exactly right. I did it by doing
blind comparisions in pairs using freeware downloaded from here:
www.pcabx.com . You might want to see if side-by-side comparisions with
instant switching helps you out.
Carey Carlan
news:0PGdnbIlltwQ6sHY...@comcast.com:
I find the differences to be very subtle.
FWIW, I dislike multitracked ensembles. The cues I would listen for
simply don't exist.
First attempt:
I think I hear just a bit more detail in the "fuzz" after the harpsichord
low notes in #1. #2 sounds most noisy, and #3 is in between. This is
sighted evaluation, not ABX.
Second attempt:
I think I hear more clarity in the violin strokes around 9-12 seconds in
in the third copy.
Third attempt:
I'm still leaning toward #3 as being the clean copy, but I'm not
convinced. There are some moments like 52-58 seconds in where there
simply sounds like less stuff fighting the sound in #3.
Mostly what I'm hearing is a lot of gooey reverb on the flute and an edgy
microphone on the harpsichord. I'm hearing room resonance messing with
the bass violin. Those have nothing to do with any distortion you added.
Harry Lavo
"Bill" <bb...@comcast.net> wrote in message
news:KR07h.3668$yE6...@newssvr14.news.prodigy.com...
> Arny Krueger wrote:
>> The following are links to three files, time-synched and level-matched.
>> One was transcribed from a SACD, and two were modified by means of added
>> distortion. They are *not* in any particular order:
>>
>> Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>> Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>> Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>>
>> If you have a little time, you might want to listen to them and figure
>> out which one sounds the best to you, and report back.
>>
>> Where "Those Amazing Audiophiles" come in is that these files were
>> auditioned by dozens of audiophiles on a forum someplace, and reported
>> back their results for a survey.
>>
>> They preferred the *most distorted* file. Which one is it, and do you
>> like it the most, too?
>>
You've been told by the person who supplied you with those files that this
statement is wrong....they preferred the "in-between" file with a
somedistortion added (as identified by you and confirmed by you), not the
high distortion file.
Peter Schneider
> The following are links to three files, time-synched and level-matched. One
> was transcribed from a SACD, and two were modified by means of added
> distortion. They are *not* in any particular order:
>
> Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
Track 01 is the most distorted. Especially at about 1 min when the
violin and the cello play together it sounds as if somebody had treated
the bows with sandpaper...
> Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
Track 02 sounds best to me. I believe it is the original.
> Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
I would rate Track 03 as second best. There is something slightly fishy
about the high frequencies; somehow the image falls apart a bit. It also
sounds a bit too "edgy" - the percussive harpsichord chords have a touch
of snare drum in them which I think sounds unnatural.
Thanks for the experiment! While I found it easy to spot the really
fouled up one, the differences between 02 and 03 are much more subtle to
my ears. I wouldn't be too surprised if 03 was the unmodified file.
Peter
Chevdo
says...
>
>
>
>Where "Those Amazing Audiophiles" come in is that these files were
>auditioned by dozens of audiophiles on a forum someplace, and reported back
>their results for a survey.
>
On a 'forum someplace'? Obviously you know all about this, including which
forum it took place on. You are deliberately hiding that particularly relevent
information, as is unfortunately standard modus operandi for you.
The files all sound the same to me, so obviously whatever differences there are
between them are slight. We already know that adding noise in some cases will
increase fidelity, that is what dither is for. So if the so-called distortion
added to one of these files was dither that the others hadn't been treated
with, after they were all reduced to a lower bitrate or had higher than final
bitrate processing applied at some point, then it would certainly make sense
that the one with 'distortion' would be the one with the highest fidelity.
Unless we know more about how these files were treated, (which is information
that would likely appear on the forum, the name of which you so conveniently
omitted, as it's unlikely they could get so many people to participate if they
were to only provide as much information as you have) then the entire excersize
is pointless.
Chevdo
says...
>
>"Arny Krueger" <ar...@hotpop.com> wrote in
>news:0PGdnbIlltwQ6sHY...@comcast.com:
>
>> The following are links to three files, time-synched and
>> level-matched. One was transcribed from a SACD, and two were modified
>> by means of added distortion. They are *not* in any particular order:
>>
>> Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>> Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>> Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>>
>> If you have a little time, you might want to listen to them and figure
>> out which one sounds the best to you, and report back.
>>
>> Where "Those Amazing Audiophiles" come in is that these files were
>> auditioned by dozens of audiophiles on a forum someplace, and reported
>> back their results for a survey.
>>
>> They preferred the *most distorted* file. Which one is it, and do you
>> like it the most, too?
>
>I find the differences to be very subtle.
>
>FWIW, I dislike multitracked ensembles. The cues I would listen for
>simply don't exist.
>
>First attempt:
>I think I hear just a bit more detail in the "fuzz" after the harpsichord
>low notes in #1. #2 sounds most noisy, and #3 is in between. This is
>sighted evaluation, not ABX.
>
If they aren't identical, which I suspect they might be, then I would say #1
has the highest fidelity, but that's little more than a guess.
Chevdo
pschneid...@fastmail.fm says...
>
>Arny Krueger schrieb:
>> The following are links to three files, time-synched and level-matched. One
>> was transcribed from a SACD, and two were modified by means of added
>> distortion. They are *not* in any particular order:
>>
>> Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>
>Track 01 is the most distorted. Especially at about 1 min when the
>violin and the cello play together it sounds as if somebody had treated
>the bows with sandpaper...
>
>> Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>
>Track 02 sounds best to me. I believe it is the original.
>
>>
Hmm judging from between :55-1:05 I agree with you. Seems like the only
part of the recording in which a difference is noticable. I didn't listen
that far along to make my initial assement, but #1 sounds most distorted
during those 10 seconds, #2 sounds least distorted, and #3 is in the middle.
Ben Bradley
>In article <0PGdnbIlltwQ6sHY...@comcast.com>, ar...@hotpop.com
>says...
>>
>>
>>
>>Where "Those Amazing Audiophiles" come in is that these files were
>>auditioned by dozens of audiophiles on a forum someplace, and reported back
>>their results for a survey.
>>
>
>On a 'forum someplace'? Obviously you know all about this, including which
>forum it took place on. You are deliberately hiding that particularly relevent
>information, as is unfortunately standard modus operandi for you.
>
>The files all sound the same to me,
Yeah, my ears suck too.
>so obviously whatever differences there are
>between them are slight.
OTOH, differences I've thought of as slight, or couldn't even hear
(for the first five or ten listenings) were said to be significant by
others here who make their living recording things. I've learned
enough not to scoff at what they say.
>We already know that adding noise in some cases will
>increase fidelity, that is what dither is for.
>So if the so-called distortion
>added to one of these files was dither that the others hadn't been treated
>with, after they were all reduced to a lower bitrate or had higher than final
>bitrate processing applied at some point, then it would certainly make sense
>that the one with 'distortion' would be the one with the highest fidelity.
Proper dithering LOWERS ...
Oh, forget it.
William Sommerwerck
And your point is...?
Fidelity is not about what one likes, but about what comes closest to the
original. What "sounds best" is of no interest. To me, anyway.
Ben Bradley
wrote:
>The following are links to three files, time-synched and level-matched. One
>was transcribed from a SACD, and two were modified by means of added
>distortion. They are *not* in any particular order:
>
>Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>
>If you have a little time, you might want to listen to them and figure out
>which one sounds the best to you, and report back.
Unfortunately, they do all sound the same to me, though the
playback system I listened to these with is far from the best I have.
But I did load them into Audacity and inverted one track, and when
played with either of the other two tracks, I can hear the
"distortion", immediately proving to myself that these tracks are not
the same. It's clearly not dither stuff as Chevo claims (Is Chevo the
new Phil?), as the difference signals 'play' at -20 to -40dB, well
above what a difference in dither/no dither would do. This difference
file only 'plays' on the louder parts (sounding like some really bad
crossover distortion), telling me the distortion is along the peaks of
the waveform.
APR
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:0PGdnbIlltwQ6sHY...@comcast.com...
system. I find it difficult to make any statement about which has the least
distortion, however, I would say the second is the cleanest??? or most pure
sound.
Julian
wrote:
>The following are links to three files, time-synched and level-matched. One
>was transcribed from a SACD, and two were modified by means of added
>distortion. They are *not* in any particular order:
>
>Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>
>If you have a little time, you might want to listen to them and figure out
>which one sounds the best to you, and report back.
I'll guess 2 is the original.
Julian
Arny Krueger
> The following are links to three files, time-synched and
Sharp readers know that I actually gave away the identiry of the web site
where the files were stored and discussed in the links, above.
The identity of the files is as follows:
> Track 01:
> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
*The most distorted
> Track 02:
> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
*The track with no added distortion
> Track 03:
> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
*The second-most distorted track.
Technical evidence based on tests done later on yesterday seems to confirm
these identifications.
I took my own advice this morning, and with google's help I came up with the
following link, which appears to be a report ( in dutch) by the creator of
the files:
http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/HoeHoorbaarIs5PrctTHDH2.pdf
On page 8 we find the following results for "most clean" from the forums
where the test was posted:
Hififorum: n = 11
Track 1 54.5%
Track 2 18.2%
Track 3 27.3%
Hear forum: n = 21
Track 1 38.1%
Track 2 33.3 %
Track 3 28.57%
Zelfbouwaudio forum n = 12
Track 1 56%
Track 2 0%
Track 3 33%
Drums
answer
in a blind test.
"William Sommerwerck" <grizzle...@comcast.net> wrote in message
news:HYGdnQXfQqb6rcDY...@comcast.com...
Carey Carlan
> The identity of the files is as follows:
>
>> Track 01:
>> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>
> *The most distorted
>
>> Track 02:
>> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>
> *The track with no added distortion
>
>> Track 03:
>> http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>
> *The second-most distorted track.
Knowing the correct answers I listened again and *still* can't hear any
reasonable differences. Does that mean that I'm deaf to 5% THD?
Scott Dorsey
>
>Knowing the correct answers I listened again and *still* can't hear any
>reasonable differences. Does that mean that I'm deaf to 5% THD?
Make a clipper with a couple zener diodes back to back, and put a 1 KC
tone through it... crank the level up until you can just barely hear
the distortion.
Now, measure the distortion on that tone.
You should find, if you're about average for engineers, that the 2% distortion
point is where you first hear clipping on 1KC tone, and it's also about the
same point where you can see flattopping on a scope.
Try it again with an 8KC tone and you can get a HUGE amount of flattopping
before you hear anything. The third harmonic at 24 KHz isn't audible itself
and you only hear the small amount of second harmonic (mostly a difference
product).
Distortion is a funny thing. And oh, yes... if you can't hear the difference
on your speakers, try it on different speakers. I find that I can hear all
kinds of nastiness on Sony MDR-V6es that are inaudible on Grados, for
instance... and low end distortion that is painfully obvious on my Maggies
is almost inaudible on the SLS speakers, while upper midrange distortion
issues that are audible on the SLS are hard to hear on the Maggies...
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Arny Krueger
news:Xns987F6EF3C...@140.99.99.130
Try ABXing them. Free comparators at www.pcabx.com.
> Does that mean that I'm deaf to 5% THD?
No. Second harmonic distortion is harder to hear than the higher orders.
It's all about masking.
Carey Carlan
>> Knowing the correct answers I listened again and *still*
>> can't hear any reasonable differences.
>
> Try ABXing them. Free comparators at www.pcabx.com.
>
>> Does that mean that I'm deaf to 5% THD?
>
> No. Second harmonic distortion is harder to hear than the higher orders.
> It's all about masking.
Do you know what specific distortion was added? Was it only/mostly 2nd
harmonic? Inquiring minds want to know.
Chevdo
tgio...@REMOVEnycap.rr.com says...
>
>If you changed the order you hear them I bet some would have a different
>answer
>in a blind test.
Personally I doubt that a double-blind test would result in any useful data
whatsoever, except perhaps if that ten seconds around the one minute mark was
the focus.
Chevdo
ben_nospa...@frontiernet.net says...
>
>
>>so obviously whatever differences there are
>>between them are slight.
>
> OTOH, differences I've thought of as slight, or couldn't even hear
>(for the first five or ten listenings) were said to be significant by
>others here who make their living recording things. I've learned
>enough not to scoff at what they say.
>
And without examining the results of double-blind testing, how do you know that
what they are telling you is an accurate representation of reality and not a
figment of their delusion? Sounds to me like you are easily socially
subordinated.
Chevdo
ben_nospa...@frontiernet.net says...
>
>It's clearly not dither stuff as Chevo claims (Is Chevo
I made no such claim and my name isn't Chevo, retard.
Chevdo
says...
Listen between :55-1:05, that's the only portion in which the differences are
noticable. In this test those who are patient are far more likely to come up
with the correct answers than those who aren't, since the only portion in which
the differences are noticable is near the very end of the files. I wasn't
patient enough to find the differences until I read Peter Schneider's post, in
which he pointed out that the differences are noticable around a minute into
the program. I checked, and he's right.
(note however, that I make that claim based on my belief that all humans
without significant ear damage differ in ability to hear only slightly, and
that belief is based on the best evidence available to date).
Chevdo
>In article <lnfql25ohfe7gs143...@4ax.com>,
>ben_nospa...@frontiernet.net says...
>
>It's clearly not dither stuff as Chevo claims (Is Chevo
BTW, you clearly don't know the difference between a claim and an if->then
statement. And without a grasp of fundementals like that, your input into this
and any other discussion requiring dialectic can only be muddled.
Arny Krueger
news:zHQ7h.17492$C94.9994@edtnps82
> In article <eUF7h.26866$xw1....@twister.nyroc.rr.com>,
> tgio...@REMOVEnycap.rr.com says...
>>
>> If you changed the order you hear them I bet some would
>> have a different answer
>> in a blind test.
>
> Personally I doubt that a double-blind test would result
> in any useful data whatsoever, except perhaps if that ten
> seconds around the one minute mark was the focus.
I used DBTs to obtain correct identification of the files. I also provided a
link to software that people can use to do their own DBTs with these files,
or any other files that one might happen to find.
Arny Krueger
> "Arny Krueger" <ar...@hotpop.com> wrote in
> news:4amdnV_m1OLC_MLY...@comcast.com:
>
>>> Knowing the correct answers I listened again and *still*
>>> can't hear any reasonable differences.
>>
>> Try ABXing them. Free comparators at www.pcabx.com.
>>
>>> Does that mean that I'm deaf to 5% THD?
>>
>> No. Second harmonic distortion is harder to hear than
>> the higher orders. It's all about masking.
>
> Do you know what specific distortion was added?
Second harmonic
> Was it only/mostly 2nd harmonic?
AFAIK, only 2nd because that's what the (Dutch) technical report seems to
say.
My own technical analysis pointed to even order distortion and lots of it.
The tip-off was that some positive peaks were elongated. Negative peaks
weren't so obviouisly changed.
My hearing suggested second.
Arny Krueger
> In article
> <0PGdnbIlltwQ6sHY...@comcast.com>,
> ar...@hotpop.com says...
>> Where "Those Amazing Audiophiles" come in is that these
>> files were auditioned by dozens of audiophiles on a
>> forum someplace, and reported back their results for a
>> survey.
> On a 'forum someplace'? Obviously you know all about
> this, including which forum it took place on.
Actually, the identity of the forum was revealed by the links I provided. I
later on used this information to provide full disclosure about the files.
>You are deliberately hiding that particularly relevent
> information, as is unfortunately standard modus operandi
> for you.
This is paranoid blather. If I had identified the forum where this came
from, it would destroy any blindness that the test had for RAP participants.
As soon as it was appropriate, I provided a link to the fullest and most
detailed report about the files that I could find.
> The files all sound the same to me, so obviously whatever
> differences there are between them are slight.
FWIW, we're talking increments of 5% THD here. Rather gross by some people's
standards.
> We
> already know that adding noise in some cases will
> increase fidelity, that is what dither is for.
This test wasn't about adding noise, it was about adding second harmonic
distortion, and fairly gross amounts of it.
<snip remaining speculation about dither>
> Unless we know more about how these files were treated,
> (which is information that would likely appear on the
> forum, the name of which you so conveniently omitted, as
> it's unlikely they could get so many people to
> participate if they were to only provide as much
> information as you have) then the entire excersize is
> pointless.
I was interested in providing the people on RAP who wanted to play, with a
opportunity for a blind test. For those who want a sighted test, they can
read the report by the person who created the files, based on the link I
provided in a clearly identified post.
OK, the report is in Dutch but anybody else's Dutch may be better than mine.
Julian
wrote:
>I was interested in providing the people on RAP who wanted to play, with a
>opportunity for a blind test. For those who want a sighted test, they can
>read the report by the person who created the files, based on the link I
>provided in a clearly identified post.
Arny,
What's the answer? Or did I miss the answer in a clearly identified
link somewhere? :-)
Julian
Arny Krueger
news:52c0m2prgqvri275h...@4ax.com
There is a post on this thread that says "don't read this..." in its title
line. It spills the beans.
Bob Quintal
news:6RQ7h.17509$C94.9248@edtnps82:
> In article <Xns987F6EF3C...@140.99.99.130>,
> gul...@hotmail.com says...
>>
>>"Arny Krueger" <ar...@hotpop.com> wrote in
>>news:v6adnYgxu4GsVsDY...@comcast.com:
>>
>>> The identity of the files is as follows:
>>>
>>>> Track 01:
>>>>
http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>>>
>>> *The most distorted
>>>
>>>> Track 02:
>>>>
http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>>>
>>> *The track with no added distortion
>>>
>>>> Track 03:
>>>>
http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>>>
>>> *The second-most distorted track.
>>
>>Knowing the correct answers I listened again and *still* can't
>>hear any reasonable differences. Does that mean that I'm deaf
>>to 5% THD?
>
> Listen between :55-1:05, that's the only portion in which the
> differences are noticable.
Not if you have good hearing. The distortion in the flute at
around 0:14 to 0:17 in sample 1 was immediately grating to my
ears. and again at 0:25~0:30. the section at 0:55 to 1:05 is, as
you mentioned, a red cape to a bull.
Distortion in sample three is less noticeable than that of
sample 1, but because of the blatant distortion in the first
sample, it was easy to hear the lesser distortion in sample 2
and the lack of it in sample two.
I wonder, if samples 1 and three had been inversed, would I have
had such an easy time of it?. .
In this test those who are patient
> are far more likely to come up with the correct answers than
> those who aren't, since the only portion in which the
> differences are noticable is near the very end of the files.
> I wasn't patient enough to find the differences until I read
> Peter Schneider's post, in which he pointed out that the
> differences are noticable around a minute into the program. I
> checked, and he's right.
>
> (note however, that I make that claim based on my belief that
> all humans without significant ear damage differ in ability to
> hear only slightly, and that belief is based on the best
> evidence available to date).
>
I trend to think that everybody suffers some hearing degradation
from birth, but some more than others. There is also the issue
of processing aural information after the ear has transformed
the air pressure variations to nerve impulses. I doubt that
that's constant.
--
Bob Quintal
PA is y I've altered my email address.
--
Posted via a free Usenet account from http://www.teranews.com
Richard Smol
Arny Krueger wrote:
> The following are links to three files, time-synched and level-matched. One
> was transcribed from a SACD, and two were modified by means of added
> distortion. They are *not* in any particular order:
>
> Track 01: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
> Track 02: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
> Track 03: http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>
> If you have a little time, you might want to listen to them and figure out
> which one sounds the best to you, and report back.
They all sound awful to me, since the flute is quite out of tune. Very
annoying...
RS
Richard Smol
Arny Krueger wrote:
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:0PGdnbIlltwQ6sHY...@comcast.com
> > Track 01:
So you just found out why mastering engineers use exciters on a mix.
Good for you!
RS
Julian
wrote:
>> Do you know what specific distortion was added?
>
>Second harmonic
>
>> Was it only/mostly 2nd harmonic?
>
>AFAIK, only 2nd because that's what the (Dutch) technical report seems to
>say.
2nd order harmonic distortion only is not a fair test, IMO. 2nd
harmonic is the most "pleasing" of all distortions because being the
octave it is always perfectly in tune with the music. IMO if they
used 3rd or better still 5th harmonic people would be more likely to
identify the clam at far lower levels of THD.
Julian
Arny Krueger
> On Sat, 18 Nov 2006 23:49:40 -0500, "Arny Krueger"
> <ar...@hotpop.com> wrote:
>
>
>>> Do you know what specific distortion was added?
>>
>> Second harmonic
>>
>>> Was it only/mostly 2nd harmonic?
>>
>> AFAIK, only 2nd because that's what the (Dutch)
>> technical report seems to say.
>
> 2nd order harmonic distortion only is not a fair test,
> IMO.
Well, it is what it is. For example, it is a tougher test than similar
percentages of sixth order.
> 2nd harmonic is the most "pleasing" of all
> distortions because being the octave it is always
> perfectly in tune with the music.
Unfortunately, that's only half the story. If you distort a complex signal
(music qualifies!) you get *both* harmonics and IM. The IM is much more
likely to be aharmonic.
> IMO if they used 3rd
> or better still 5th harmonic people would be more likely
> to identify the clam at far lower levels of THD.
This is true, but the explanation of *why* may be controversial. I tend to
shy away from explanations that presume preference, because preference is
far from being a uniform effect.
Higher harmonics are less likely to be masked, and masking is a property of
hearing that is far less variable from person to person, than preference.
There are at least two reasons why higher harmonics are less likely to be
masked:
(1) Frequencies that are further away from the fundamental are simply less
likely to be masked by the ear.
(2) High order harmonics tend to be less intense, so the maskers tend to be
smaller.
Scott Dorsey
>>>> Was it only/mostly 2nd harmonic?
>>>
>>> AFAIK, only 2nd because that's what the (Dutch)
>>> technical report seems to say.
>>
>> 2nd order harmonic distortion only is not a fair test,
>> IMO.
>
>Well, it is what it is. For example, it is a tougher test than similar
>percentages of sixth order.
Yes, and the second harmonic may tend to make the sound more "mellow" and
perhaps might actually make it sound more balanced if there are already
some higher order even distortion products. Adding some low order even
stuff can make the higher order stuff less audible and the overall sound
more pleasing.
>> 2nd harmonic is the most "pleasing" of all
>> distortions because being the octave it is always
>> perfectly in tune with the music.
>
>Unfortunately, that's only half the story. If you distort a complex signal
>(music qualifies!) you get *both* harmonics and IM. The IM is much more
>likely to be aharmonic.
Probably, but how did they distort the original signal in this test?
I'll say that adding midrange IM tends to add more of a sense of ensemble...
instruments tend to blend into one another a bit. Lots of folks think this
is a good sound when they are trying to turn discrete tracks into a coherent
mix. But, on an accurate recording, it can cause lack of definition.
>> IMO if they used 3rd
>> or better still 5th harmonic people would be more likely
>> to identify the clam at far lower levels of THD.
>
>This is true, but the explanation of *why* may be controversial. I tend to
>shy away from explanations that presume preference, because preference is
>far from being a uniform effect.
I think the psychoacoustic research on audibility of harmonics is pretty
extensive already and so I don't see any of this explanation being controversial
in any way.
>Higher harmonics are less likely to be masked, and masking is a property of
>hearing that is far less variable from person to person, than preference.
Yes. And, adding lower order harmonics can make higher order ones less audible.
>There are at least two reasons why higher harmonics are less likely to be
>masked:
>
>(1) Frequencies that are further away from the fundamental are simply less
>likely to be masked by the ear.
>
>(2) High order harmonics tend to be less intense, so the maskers tend to be
>smaller.
Right. This isn't in any way surprising. All that stuff was pretty well
covered years ago.
Arny Krueger
news:ejqlqh$mun$1...@panix2.panix.com
> Arny Krueger <ar...@hotpop.com> wrote:
I don't read Dutch so well...
However, it appears that Matlab was used to add the right fraction of x
squared
> I'll say that adding midrange IM tends to add more of a
> sense of ensemble... instruments tend to blend into one
> another a bit.
Please tell me how to get IM without harmonics or vice-versa, without
bandpass filtering. ;-)
> ots of folks think this
> is a good sound when they are trying to turn discrete
> tracks into a coherent mix. But, on an accurate
> recording, it can cause lack of definition.
Agreed.
>>> IMO if they used 3rd
>>> or better still 5th harmonic people would be more likely
>>> to identify the clam at far lower levels of THD.
>> This is true, but the explanation of *why* may be
>> controversial. I tend to shy away from explanations that
>> presume preference, because preference is far from being
>> a uniform effect.
> I think the psychoacoustic research on audibility of
> harmonics is pretty extensive already and so I don't see
> any of this explanation being controversial in any way.
Not everybody seems to believe in psychoacoustics. ;-)
>> Higher harmonics are less likely to be masked, and
>> masking is a property of hearing that is far less
>> variable from person to person, than preference.
> Yes. And, adding lower order harmonics can make higher
> order ones less audible.
Right, because they mask the upper ones.
>> There are at least two reasons why higher harmonics are
>> less likely to be masked:
>> (1) Frequencies that are further away from the
>> fundamental are simply less likely to be masked by the
>> ear.
>> (2) High order harmonics tend to be less intense, so the
>> maskers tend to be smaller.
>
> Right. This isn't in any way surprising. All that stuff
> was pretty well covered years ago.
Then it is a tutorial for those who are not yet hip to the current findings
of psychoacoustics.
Julian
>Yes, and the second harmonic may tend to make the sound more "mellow" and
>perhaps might actually make it sound more balanced if there are already
>some higher order even distortion products. Adding some low order even
>stuff can make the higher order stuff less audible and the overall sound
>more pleasing.
My point exactly. Doesn't the Aphex Aural Exciter actually add second
order harmonics on purpose because people tend to like the way they
sound?
>>> 2nd harmonic is the most "pleasing" of all
>>> distortions because being the octave it is always
>>> perfectly in tune with the music.
>>
>>Unfortunately, that's only half the story. If you distort a complex signal
>>(music qualifies!) you get *both* harmonics and IM. The IM is much more
>>likely to be aharmonic.
Do you know how much IM was produced in this study?
>>> IMO if they used 3rd
>>> or better still 5th harmonic people would be more likely
>>> to identify the clam at far lower levels of THD.
>>
>>This is true, but the explanation of *why* may be controversial. I tend to
>>shy away from explanations that presume preference, because preference is
>>far from being a uniform effect.
You are absolutely correct to do so when making scientific
investigations. Thank You for your diligence in this regard. I'm
just sitting here BS'ing however :-). I'd really like to see the
exact same test done with odd order harmonics. I predict the the
exact opposite result. People will NOT prefer the one with odd order
harmonics.
>>(1) Frequencies that are further away from the fundamental are simply less
>>likely to be masked by the ear.
I'll bet that octaves - 4th, 8th 16th order harmonics are far more
pleasing to the ear than 5th, 9th, 17th.
>>(2) High order harmonics tend to be less intense, so the maskers tend to be
>>smaller.
Correct and you'd expect the threshold to be proportionally lower, but
I still predict research would show most odd harmonics are
proportionally more disturbing than even order.
Julian
Ben Bradley
I'm not sure I'd use the word 'pleasing', or even 'consonant', but
they definitely sound different.
Or you may be right, as a Hammond organ does not include the 7th
harmonic on the drawbars. Musically, it's perhaps the most dissonant
harmonic.
Arny Krueger
>> Yes, and the second harmonic may tend to make the sound
>> more "mellow" and perhaps might actually make it sound
>> more balanced if there are already some higher order
>> even distortion products. Adding some low order even
>> stuff can make the higher order stuff less audible and
>> the overall sound more pleasing.
>
> My point exactly. Doesn't the Aphex Aural Exciter
> actually add second order harmonics on purpose because
> people tend to like the way they sound?
AFAIK the Aphex process is not based on a mulitiplier, but instead it is
based on a half-wave rectifier. That means that it adds a certain collection
of even harmonics, not just second harmonic.
>>>> 2nd harmonic is the most "pleasing" of all
>>>> distortions because being the octave it is always
>>>> perfectly in tune with the music.
>>>
>>> Unfortunately, that's only half the story. If you
>>> distort a complex signal (music qualifies!) you get
>>> *both* harmonics and IM. The IM is much more likely to
>>> be aharmonic.
> Do you know how much IM was produced in this study?
I know how to make 5% nonlinear distortion, and I could punch an audio
rightmark through it.
>>>> IMO if they used 3rd
>>>> or better still 5th harmonic people would be more
>>>> likely to identify the clam at far lower levels of THD.
>>>
>>> This is true, but the explanation of *why* may be
>>> controversial. I tend to shy away from explanations
>>> that presume preference, because preference is far from
>>> being a uniform effect.
>
> You are absolutely correct to do so when making scientific
> investigations. Thank You for your diligence in this
> regard. I'm just sitting here BS'ing however :-).
> I'd really like to see the exact same test done with odd
> order harmonics.
Answer one is that a similar test was done for orders 2-5 here:
http://www.pcabx.com/technical/nonlinear/
Answer two is that to make it an apples-to-apples comparison, masking should
be compensated for.
> I predict the the exact opposite
> result. People will NOT prefer the one with odd order
> harmonics.
My point is that you can't separate the fact that 3rd is the first odd
order, and its higher than second.
>>> (1) Frequencies that are further away from the
>>> fundamental are simply less likely to be masked by the
>>> ear.
> I'll bet that octaves - 4th, 8th 16th order harmonics are
> far more pleasing to the ear than 5th, 9th, 17th.
IME 4th is nastier than 3rd, if you see my point.
Go to the pcabx site and listen for yourself!
>>> (2) High order harmonics tend to be less intense, so
>>> the maskers tend to be smaller.
> Correct and you'd expect the threshold to be
> proportionally lower, but I still predict research would
> show most odd harmonics are proportionally more
> disturbing than even order.
Listen for yourself...
Ben Bradley
wrote:
>"Scott Dorsey" <klu...@panix.com> wrote in message
>news:ejqlqh$mun$1...@panix2.panix.com
>
>> Arny Krueger <ar...@hotpop.com> wrote:
>
(sorry, who wrote what was lost and I'm to lazy to recreate it)
>>>> 2nd harmonic is the most "pleasing" of all
>>>> distortions because being the octave it is always
>>>> perfectly in tune with the music.
>
>>> Unfortunately, that's only half the story. If you
>>> distort a complex signal (music qualifies!) you get
>>> *both* harmonics and IM. The IM is much more likely to
>>> be aharmonic.
>
>> Probably, but how did they distort the original signal in
>> this test?
>
>I don't read Dutch so well...
I copy/pasted the first page of that pdf into Babelfish and did
Dutch->English. It gave the expected Yoda-like syntax (always amusing
to read), but I thought it was understandable enough. Perhaps someone
could do that with the whole paper and put it up on a website...
>However, it appears that Matlab was used to add the right fraction of x
>squared
>
>> I'll say that adding midrange IM tends to add more of a
>> sense of ensemble... instruments tend to blend into one
>> another a bit.
>
>Please tell me how to get IM without harmonics or vice-versa, without
>bandpass filtering. ;-)
I was about to say FFT->processing->inverse FFT, but of course FFT
IS bandpass filtering. And the transforms and processing are liable to
create other undesired effects. But it could be interesting - has
anyone ever tried this?
>> ots of folks think this
>> is a good sound when they are trying to turn discrete
>> tracks into a coherent mix. But, on an accurate
>> recording, it can cause lack of definition.
>
>Agreed.
>
>>>> IMO if they used 3rd
>>>> or better still 5th harmonic people would be more likely
>>>> to identify the clam at far lower levels of THD.
>
>>> This is true, but the explanation of *why* may be
>>> controversial. I tend to shy away from explanations that
>>> presume preference, because preference is far from being
>>> a uniform effect.
>
>> I think the psychoacoustic research on audibility of
>> harmonics is pretty extensive already and so I don't see
>> any of this explanation being controversial in any way.
>
>Not everybody seems to believe in psychoacoustics. ;-)
In other words, on RAO everything is controversial!
Scott Dorsey
>> sense of ensemble... instruments tend to blend into one
>> another a bit.
>
>Please tell me how to get IM without harmonics or vice-versa, without
>bandpass filtering. ;-)
You use bandpass filtering! That's what testing is all about, isolating
different effects so you know what causes what audible change.
>> I think the psychoacoustic research on audibility of
>> harmonics is pretty extensive already and so I don't see
>> any of this explanation being controversial in any way.
>
>Not everybody seems to believe in psychoacoustics. ;-)
I recommend psychoacoustics, and not just because I have a degree in
the subject. But by the same token, I also believe that audibility
of all kinds of things aren't very well established yet. Still,
harmonic distortion effects are pretty well modelled these days.
Scott Dorsey
>>Yes, and the second harmonic may tend to make the sound more "mellow" and
>>perhaps might actually make it sound more balanced if there are already
>>some higher order even distortion products. Adding some low order even
>>stuff can make the higher order stuff less audible and the overall sound
>>more pleasing.
>
>My point exactly. Doesn't the Aphex Aural Exciter actually add second
>order harmonics on purpose because people tend to like the way they
>sound?
No, it adds _very high order_ even harmonics. Which is a very different
thing, and much more audible than second harmonic distortion.
>Thank You for your diligence in this regard. I'm
>just sitting here BS'ing however :-). I'd really like to see the
>exact same test done with odd order harmonics. I predict the the
>exact opposite result. People will NOT prefer the one with odd order
>harmonics.
Depends. Third order harmonics they might like. Seventh... they won't.
But, if given the choice between seventh and eighth, lots of people will
pick the seventh over the eighth.
>>>(1) Frequencies that are further away from the fundamental are simply less
>>>likely to be masked by the ear.
>
>I'll bet that octaves - 4th, 8th 16th order harmonics are far more
>pleasing to the ear than 5th, 9th, 17th.
For the most part the opposite is true... those high order even ones
tend to make for that "1970s solid state" sound. It's possible that I am
particularly sensitive to that having lived through the era where everything
really did sound that way all the time.
>>>(2) High order harmonics tend to be less intense, so the maskers tend to be
>>>smaller.
>
>Correct and you'd expect the threshold to be proportionally lower, but
>I still predict research would show most odd harmonics are
>proportionally more disturbing than even order.
Depends on the order.
Julian
wrote:
>AFAIK the Aphex process is not based on a multiplier, but instead it is
>based on a half-wave rectifier. That means that it adds a certain collection
>of even harmonics, not just second harmonic.
Thanks for clarifying that.
>> I'd really like to see the exact same test done with odd
>> order harmonics.
>
>Answer one is that a similar test was done for orders 2-5 here:
>
>http://www.pcabx.com/technical/nonlinear/
OK. have now listened.
>Answer two is that to make it an apples-to-apples comparison, masking should
>be compensated for.
What do you mean by masking?
>My point is that you can't separate the fact that 3rd is the first odd
>order, and its higher than second.
Is true.
>IME 4th is nastier than 3rd, if you see my point.
After listening I would say I come to the opposite conclusion. Why do
you think it is nastier?
Julian
Chris Hornbeck
wrote:
>Please tell me how to get IM without harmonics or vice-versa, without
>bandpass filtering. ;-)
Doppler!
Chris Hornbeck
"I hate to see that evening Sun go down"
Chevdo
says...
>
>"Chevdo" <ch...@dont.org> wrote in message
>news:zHQ7h.17492$C94.9994@edtnps82
>> In article <eUF7h.26866$xw1....@twister.nyroc.rr.com>,
>> tgio...@REMOVEnycap.rr.com says...
>>>
>>> If you changed the order you hear them I bet some would
>>> have a different answer
>>> in a blind test.
>>
>> Personally I doubt that a double-blind test would result
>> in any useful data whatsoever, except perhaps if that ten
>> seconds around the one minute mark was the focus.
>
>I used DBTs to obtain correct identification of the files. I also provided a
>link to software that people can use to do their own DBTs with these files,
>or any other files that one might happen to find.
Oh really. Please describe exactly how the test you did took place. I don't
believe you know what double-blind means. For one thing, it would be virtually
impossible for a single person to conduct a double-blind test. Besides, this
is the first we've heard of you doing any testing, and what relevence would
your testing have anyway? The forum in which this event took place certainly
made no claims that it was to be administered double-blind, and this being your
first mention of your own so-called 'double blind' testing of the files, you
haven't even mentioned what the results of your test. Did you guess right, or
not? And don't forget to describe exactly how you performed a 'double-blind'
test on yourself...
Chevdo
says...
>
>ch...@dont.org (Chevdo) wrote in
>news:6RQ7h.17509$C94.9248@edtnps82:
>
>> In article <Xns987F6EF3C...@140.99.99.130>,
>> gul...@hotmail.com says...
>>>
>>>"Arny Krueger" <ar...@hotpop.com> wrote in
>>>news:v6adnYgxu4GsVsDY...@comcast.com:
>>>
>>>> The identity of the files is as follows:
>>>>
>>>>> Track 01:
>>>>>
>http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/1.wav
>>>>
>>>> *The most distorted
>>>>
>>>>> Track 02:
>>>>>
>http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/2.wav
>>>>
>>>> *The track with no added distortion
>>>>
>>>>> Track 03:
>>>>>
>http://amorgignitamorem.nl/Audio/5ProcentTHDexperiment/3.wav
>>>>
>>>> *The second-most distorted track.
>>>
>>>Knowing the correct answers I listened again and *still* can't
>>>hear any reasonable differences. Does that mean that I'm deaf
>>>to 5% THD?
>>
>> Listen between :55-1:05, that's the only portion in which the
>> differences are noticable.
>
>Not if you have good hearing.
Your response ignores my disclaimer 'note' at the bottom of the post you are
responding to, therefore, as a response it is disingenuous and self-serving.
Perhaps you should have made that statement on its own, rather than as a
response to what I had said.
However, if there is indeed a spot between :55-1:05 that is distorted enough
that each track can be distinguished from another, as is my claim, then anyone
could listen to that portion, determine the distortion order of the files, and
then claim that their exceptionally 'good ears' allowed them to distinguish the
difference at an earlier point in the program. If you did in fact determine
correctly the difference in the files from an earlier portion of the program,
it certainly is a shame that you didn't mention it in this thread before the
correct answer, as having been determined from the segment between :55-1:05,
had been posted, because that would vindicate your alleged 'good ears'. In
fact, if I were you I probably wouldn't even mention it, since I'd realize that
I missed the opportunity for vindication, and that mentioning it too late
would only make me look like a deceiver, and a very petty one at that.
Arny Krueger
> In article
> <LJednSA-8qTofMLY...@comcast.com>,
> ar...@hotpop.com says...
>>
>> "Chevdo" <ch...@dont.org> wrote in message
>> news:zHQ7h.17492$C94.9994@edtnps82
>>> In article <eUF7h.26866$xw1....@twister.nyroc.rr.com>,
>>> tgio...@REMOVEnycap.rr.com says...
>>>>
>>>> If you changed the order you hear them I bet some would
>>>> have a different answer
>>>> in a blind test.
>>>
>>> Personally I doubt that a double-blind test would result
>>> in any useful data whatsoever, except perhaps if that
>>> ten seconds around the one minute mark was the focus.
>>
>> I used DBTs to obtain correct identification of the
>> files. I also provided a link to software that people
>> can use to do their own DBTs with these files, or any
>> other files that one might happen to find.
>
> Oh really. Please describe exactly how the test you did
> took place.
I started a program that is an automated double-blind test administrator,
loaded a pair of wave files to compare, and then compared them using the
program.
There are a number of such programs available via links from the
www.pcabx.com home page. For these tests I chose this one:
http://www.kikeg.arrakis.es/winabx/
> I don't believe you know what double-blind means.
Chevado, I *know* you blow hard.
> For one thing, it would be virtually impossible
> for a single person to conduct a double-blind test.
Even with the assistance of appropriate software to manage the test for me?
> Besides, this is the first we've heard of you doing any
> testing, and what relevence would your testing have
> anyway?
Abject ignorance of relevant facts noted.
Here is an early JAES paper related to double blind testing of audio
equipment and developments:
Clark, David L., "High-Resolution Subjective Testing Using a Double-Blind
Comparator", Journal of the Audio Engineering Society, Vol. 30 No. 5, May
1982, pp. 330-338.
It mentions me by name. This is fairly well-known to RAP regulars.
More recently but hardly new, the www.pcabx.com web site is about 6 years
old, and also well-known to be mine. Where have you been keeping your head
in the sand, Chevado?
> The forum in which this event took place
> certainly made no claims that it was to be administered
> double-blind,
That would be their problem.
> and this being your first mention of your
> own so-called 'double blind' testing of the files,
There are dozens of people who frequent this forum who are the least bit
surprized by this. Chevado, You may be the only one who *is* surprised!
> you haven't even mentioned what the results of your test.
Positive identifications, of course.
> Did you guess right, or not?
I didn't guess, my findings were reliable.
> And don't forget to
> describe exactly how you performed a 'double-blind' test
> on yourself...
Been there, done that, many times.
Arny Krueger
> Arny Krueger <ar...@hotpop.com> wrote:
>> "Scott Dorsey" <klu...@panix.com> wrote in message
>>
>>> I'll say that adding midrange IM tends to add more of a
>>> sense of ensemble... instruments tend to blend into one
>>> another a bit.
>>
>> Please tell me how to get IM without harmonics or
>> vice-versa, without bandpass filtering. ;-)
>
> You use bandpass filtering! That's what testing is all
> about, isolating different effects so you know what
> causes what audible change.
Let me clarify - I'm talking about generating IM and/or the generation of IM
in regular audio production equipment.
>>> I think the psychoacoustic research on audibility of
>>> harmonics is pretty extensive already and so I don't see
>>> any of this explanation being controversial in any way.
>> Not everybody seems to believe in psychoacoustics. ;-)
> I recommend psychoacoustics, and not just because I have
> a degree in the subject. But by the same token, I also
> believe that audibility
> of all kinds of things aren't very well established yet.
In practice, "all kinds of things" tend to be specific and highly varied.
That makes things more complex.
> Still, harmonic distortion effects are pretty well
> modelled these days.
Totally agreed.
Arny Krueger
> On Sun, 19 Nov 2006 20:21:50 -0500, "Arny Krueger"
> <ar...@hotpop.com> wrote:
>
>> AFAIK the Aphex process is not based on a multiplier,
>> but instead it is based on a half-wave rectifier. That
>> means that it adds a certain collection of even
>> harmonics, not just second harmonic.
> Thanks for clarifying that.
>>> I'd really like to see the exact same test done with
>>> odd order harmonics.
>> Answer one is that a similar test was done for orders
>> 2-5 here:
>> http://www.pcabx.com/technical/nonlinear/
> OK. have now listened.
Well, what do you think?
>> Answer two is that to make it an apples-to-apples
>> comparison, masking should be compensated for.
> What do you mean by masking?
More specifically, I mean concurrent masking or frequency masking. In
general, masking is a property of the ear that makes the ear insensitive to
softer sounds that are near to louder sounds that are also similar. For
example, if I play a 1 KHz tone, it will prevent anybody from hearing a
concurrent 999 Hz tone that is also softer. The softer tone with similar
pitch is masked by the louder tone.
There is also temporal masking, where a loud sound will keep anybody from
hearing a softer sound that immediately follows it.
http://www.embedded.com/showArticle.jhtml?articleID=164300682
"As an example, when we listen to a strong tone of a particular frequency,
we tend to be insensitive to the presence or absence of weaker sounds in the
nearby frequencies. This is known as frequency masking and is illustrated in
Figure 1. The tone at frequency f1 is "masked" (and therefore inaudible) by
the loud tone at frequency f0. The tone at frequency f2 is audible however
since it lies outside the "masking region" of frequency f0.
"The frequency masking phenomenon is exploited extensively in audio coding.
Each frame (which is collection of samples such as 1024 sample) of audio is
first transformed to the frequency domain, thereby decomposing it into a
collection of tones at various frequencies. The signal is then analyzed to
determine which tones would be irrelevant because of masking by nearby
louder tones. By coding only the tones that are audible to the human ear,
tremendous compression ratios can be achieved.
>> My point is that you can't separate the fact that 3rd is
>> the first odd order, and its higher than second.
> Is true.
>> IME 4th is nastier than 3rd, if you see my point.
> After listening I would say I come to the opposite
> conclusion. Why do you think it is nastier?
It's just how things sound to my ear. Hey we both looked at the same
evidence and have different preferences. No harm, no foul, right?
Arny Krueger
> However, if there is indeed a spot between :55-1:05 that
> is distorted enough that each track can be distinguished
> from another, as is my claim, then anyone could listen to
> that portion, determine the distortion order of the
> files, and then claim that their exceptionally 'good
> ears' allowed them to distinguish the difference at an
> earlier point in the program.
Or, they could use the comparison software that I keep mentioning, and do
comparisons up and down the whole sample easily because the software keeps
the two samples in synch, and quickly discover the same thing, were it to be
true.
Then they would be rewarded for using modern tools and being a little
dilligent.
Since the same tools facilitate doing a DBT, they would have pretty fair
confirmation of their judgement.
Carey Carlan
>> Do you know what specific distortion was added?
>
> Second harmonic
>
>> Was it only/mostly 2nd harmonic?
>
> AFAIK, only 2nd because that's what the (Dutch) technical report seems
> to say.
>
> My own technical analysis pointed to even order distortion and lots of
> it. The tip-off was that some positive peaks were elongated. Negative
> peaks weren't so obviouisly changed.
That's why I didn't hear it. I sometimes add second order harmonics on
purpose to recordings that need a bit of brightening. My ear doesn't hear
that as distortion.
Scott Dorsey
>"Scott Dorsey" <klu...@panix.com> wrote in message
>> Arny Krueger <ar...@hotpop.com> wrote:
>>> "Scott Dorsey" <klu...@panix.com> wrote in message
>>>
>>>> I'll say that adding midrange IM tends to add more of a
>>>> sense of ensemble... instruments tend to blend into one
>>>> another a bit.
>>>
>>> Please tell me how to get IM without harmonics or
>>> vice-versa, without bandpass filtering. ;-)
>>
>> You use bandpass filtering! That's what testing is all
>> about, isolating different effects so you know what
>> causes what audible change.
>
>Let me clarify - I'm talking about generating IM and/or the generation of IM
>in regular audio production equipment.
You can't really, but with analogue tape you can do all kinds of things
that create IM in varying types and degrees. That's part of what makes
the analogue machine such a powerful tool... you can set the bias point to
a whole range of different places, all of which are "correct" in that they
have low but nonzero distortion, and equalize the response flat using the
emphasis controls. All of those places produce different distortion
spectra and all of them produce different kinds and degrees of IMD. Most
of the harmonic stuff is a non-issue since it's mostly low-order, but the
IMD can be a valuable tool, and you can control it surprisingly well.
>In practice, "all kinds of things" tend to be specific and highly varied.
>That makes things more complex.
That's why audio is fun.
Chevdo
says...
>
>"Chevdo" <ch...@dont.org> wrote in message
>news:Ubc8h.6417$rB6.5678@clgrps13
>
>> However, if there is indeed a spot between :55-1:05 that
>> is distorted enough that each track can be distinguished
>> from another, as is my claim, then anyone could listen to
>> that portion, determine the distortion order of the
>> files, and then claim that their exceptionally 'good
>> ears' allowed them to distinguish the difference at an
>> earlier point in the program.
>
>Or, they could use the comparison software that I keep mentioning, and do
>comparisons up and down the whole sample easily because the software keeps
>the two samples in synch, and quickly discover the same thing, were it to be
>true.
>
Yes Arny, your software would be useful for informal listening tests like the
one we're talking about. But it does not magically make people hear what they
could not hear without the software.
>Since the same tools facilitate doing a DBT, they would have pretty fair
>confirmation of their judgement.
The only thing they facilitate is being able to flip between the tracks without
having a couple seconds of silence in between. Yes, this is helpful, but it is
not as magical as you seem to be suggesting by harping on it incessantly. It's
not even required for any listening tests, it simply allows for the possibility
of reduced error by short-term memory loss. And by 'allows for the
possibility', I mean that using it won't necessarily reduce that error at all,
but it might.
So, what's your point?
Chevdo
says...
>
>I started a program that is an automated double-blind test administrator,
>loaded a pair of wave files to compare, and then compared them using the
>program.
That is not a double-blind test, that is a single-blind test.
>
>> For one thing, it would be virtually impossible
>> for a single person to conduct a double-blind test.
>
>Even with the assistance of appropriate software to manage the test for me?
If you are managing the software, it is not a double-blind test, it is a
single-blind test..
As for the rest of your response, yes, you are a dunce, any anyone who isn't
who has read your posts knows that, too.
Chevdo
>
>In article <LIqdnTbTEu9WFfzY...@comcast.com>, ar...@hotpop.com
>says...
>>
>>I started a program that is an automated double-blind test administrator,
>>loaded a pair of wave files to compare, and then compared them using the
>>program.
>
>That is not a double-blind test, that is a single-blind test.
>
>>
>>> For one thing, it would be virtually impossible
>>> for a single person to conduct a double-blind test.
>>
>>Even with the assistance of appropriate software to manage the test for me?
>
>If you are managing the software, it is not a double-blind test, it is a
>single-blind test..
>
In fact, technically it's not even single-blind, since you knew what you were
supposed to be listening for (one file being the most distorted, another
file being in the middle, and another being the least distorted) before you
began the listening test. If you really want to do blind testing, get a few
friends together, read up on what blind testing means, since you clearly don't
know, and do your tests according to actual blind protocols. Claiming that
what you've been doing is 'blind testing' just makes you look foolish.
Arny Krueger
> In article
> <LIqdnTbTEu9WFfzY...@comcast.com>,
> ar...@hotpop.com says...
>>
>> I started a program that is an automated double-blind
>> test administrator, loaded a pair of wave files to
>> compare, and then compared them using the program.
>
> That is not a double-blind test, that is a single-blind
> test.
Please explain.
>>> For one thing, it would be virtually impossible
>>> for a single person to conduct a double-blind test.
>>
>> Even with the assistance of appropriate software to
>> manage the test for me?
>
> If you are managing the software, it is not a
> double-blind test, it is a single-blind test..
Never bothered to find out how the software works, now did you?
> As for the rest of your response, yes, you are a dunce,
> any anyone who isn't who has read your posts knows that,
> too.
Flattery will get you nowhere with me, Chevvy Baby.
Arny Krueger
> In fact, technically it's not even single-blind, since
> you knew what you were supposed to be listening for (one
> file being the most distorted, another
> file being in the middle, and another being the least
> distorted) before you began the listening test.
Thanks for sharing, Chevvy Baby. You really don't know how the software
works, do you?
> If you
> really want to do blind testing, get a few friends
> together, read up on what blind testing means, since you
> clearly don't know, and do your tests according to actual
> blind protocols.
Well, we squeaked this implementation of DBT past the JAES review board as
part of this article:
Clark, David L., "High-Resolution Subjective Testing Using a Double-Blind
Comparator", Journal of the Audio Engineering Society, Vol. 30 No. 5, May
1982, pp. 330-338.
Maybe you know more about the definition of DBT than the whole Audio
Engineering Society review board?
> Claiming that what you've been doing is 'blind testing' just makes you
> look foolish.
Trust me Chevvy, they aren't laughing with you, they're laughing at you.
Arny Krueger
> In article
> <H72dneb3qZnmEfzY...@comcast.com>,
> ar...@hotpop.com says...
>>
>> "Chevdo" <ch...@dont.org> wrote in message
>> news:Ubc8h.6417$rB6.5678@clgrps13
>>
>>> However, if there is indeed a spot between :55-1:05 that
>>> is distorted enough that each track can be distinguished
>>> from another, as is my claim, then anyone could listen
>>> to that portion, determine the distortion order of the
>>> files, and then claim that their exceptionally 'good
>>> ears' allowed them to distinguish the difference at an
>>> earlier point in the program.
>>
>> Or, they could use the comparison software that I keep
>> mentioning, and do comparisons up and down the whole
>> sample easily because the software keeps the two samples
>> in synch, and quickly discover the same thing, were it
>> to be true.
> Yes Arny, your software would be useful for informal
> listening tests like the one we're talking about. But it
> does not magically make people hear what they could not
> hear without the software.
No magic - time synchronization is well known to have the potential to
dramatically improve hearing acuity and eliminate a well-known source of
false positives.
>> Since the same tools facilitate doing a DBT, they would
>> have pretty fair confirmation of their judgement.
> The only thing they facilitate is being able to flip
> between the tracks without having a couple seconds of
> silence in between.
If you want to educate yourself a little Chevvy, there is a switchover delay
time control on the PCABX comparitor. See what adding a few seconds delay
does to your results, for a difference that is on the subtle side.
> Yes, this is helpful, but it is not
> as magical as you seem to be suggesting by harping on it
> incessantly.
Have you done your homework on this matter?
> It's not even required for any listening
> tests, it simply allows for the possibility of reduced
> error by short-term memory loss.
So you do know that its not magic.
But you missed the part where anybody with a little insight can correctly
identify identical sounds that are a little out-of-synch.
> And by 'allows for the
> possibility', I mean that using it won't necessarily
> reduce that error at all, but it might.
All one can do is do whatever reasonable things that can be done to help the
listener score better.
> So, what's your point?
Clear evidence that you're missing the point, even when you clearly know the
point? ;-)
Julian
wrote:
>"Chevdo" <ch...@dont.org> wrote in message
>news:8et8h.21231$C94.5752@edtnps82
>> Claiming that what you've been doing is 'blind testing' just makes you
>> look foolish.
>
>Trust me Chevvy, they aren't laughing with you, they're laughing at you.
I've already killed this idiot so I only see responses from others,
but it reminds me why he's in my kill filter.
Julian