OP
- Thread Starter
- #21
Maybe when I’m done distorting some good music files, I’ll do that
turns out it’s just as easy to add as it is to fix harmonic distortions.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Audibility of Harmonic distortion - a blind test proposal
- Thread starter pkane
- Start date
Maybe when I’m done distorting some good music files, I’ll do that
turns out it’s just as easy to add as it is to fix harmonic distortions.
MRC01
Major Contributor
My results are similar. Having done quite a bit of A/B/X testing over the years, I find that with ideally revealing sources (pure 2-tone, extremely well recorded vocals), and distortion applied in the upper mids to low treble where our hearing is most sensitive, I begin to hear distortion around -40 dB or 1%.
Note: distortion (HD or IM) is correlated with the signal. The thresholds of audibility for other kinds of noise not correlated with the signal can be higher or lower. For example, a few years ago I heard a low level pulse in recordings I made from a portable digital recorder. That noise happened to be at -46 dB. But this was a bit of an ideal case: it was a fixed frequency, nearly a pure tone, pulsing, and uncorrelated with the music, and in the range of 600-800 Hz. http://www.mclements.net/Mike/H4/H4-mod.html
Last edited:
Audibility of harmonic distortion depends on level, base frequency and spectral content. There is nothing like a single number as a harmonic distortion threshold. Below you can see that audibility of pure H2 starts at 0.1%. Higher harmonics are more audible than the H2.
Source: Coding High Quality Digital Audio, J. ROBERT STUART
AudioSceptic
Major Contributor
Yes, but what's the point? Who benefits from cheating?
The plot shown was with the sine tone. However, there is a next plot in the same paper
MRC01
Major Contributor
The second graph (difference tones) is interesting in that it suggests that for speakers to be transparent, bass distortion should be < 3% and tweeters should have less than 0.1% / -60 dB distortion. And that maximum hearing acuity varies with both loudness and frequency: in the low bass 80 dB SPL, mid-bass 70 dB, and midrange to treble 60 dB. That happens to be in the same range as a lot of acoustic music and should be a useful suggestion for setting critical listening levels. Also, since the energy of most acoustic instruments drops with frequency, this level of 80-70-60 from bass to treble could result from listening to a single instrument or ensemble. If it happens that the acoustic instruments people have developed over thousands of years, create SPLs that are at or near the range of our maximum hearing acuity at typical live listener distances, it is probably not coincidental.
Last edited:
- Joined
- Feb 23, 2016
- Messages
- 20,769
- Likes
- 37,633
https://pdfs.semanticscholar.org/0a0e/9492b1bd62f32a9b7baf9506918807c27de2.pdf
Here is the entire Stuart paper. Good work. This is what I refer to as Robert work. When he is trying to bamboozle you with MQA and similar it is Bob work. Opinions of Robert Stuart sometimes don't agree with what Bob Stuart is pushing in audiophile press and marketing.
As an example here is something from page 2:
Even among audio engineers, there has been considerable misunderstanding about digital audio, about the sampling theory, and about how PCM works at the functional level. Some of these misunderstandings persist even today. Top of the list of erroneous assertions are:
i. PCM cannot resolve detail smaller than the LSB (least-significant bit).
ii. PCM cannot resolve time more accurately than the sampling period.
This from the same fellow who pushes MQA to provide high resolution in order to get better time resolution and deblurring than is inherent to PCM.
Here is the entire Stuart paper. Good work. This is what I refer to as Robert work. When he is trying to bamboozle you with MQA and similar it is Bob work. Opinions of Robert Stuart sometimes don't agree with what Bob Stuart is pushing in audiophile press and marketing.
As an example here is something from page 2:
Even among audio engineers, there has been considerable misunderstanding about digital audio, about the sampling theory, and about how PCM works at the functional level. Some of these misunderstandings persist even today. Top of the list of erroneous assertions are:
i. PCM cannot resolve detail smaller than the LSB (least-significant bit).
ii. PCM cannot resolve time more accurately than the sampling period.
This from the same fellow who pushes MQA to provide high resolution in order to get better time resolution and deblurring than is inherent to PCM.
Last edited:
What year was that paper?
- Joined
- Feb 23, 2016
- Messages
- 20,769
- Likes
- 37,633
Thanks!
OP
That would be specifically second order harmonic distortion, I assume?Nice, warm simulated tube sound
AFAICS, the blind test was done already
https://archimago.blogspot.com/2020/01/internet-blind-test-is-high-harmonic.html
with pretty elaborate and very interesting conclusions
https://archimago.blogspot.com/2020/05/blind-test-results-part-i-is-high.html
https://archimago.blogspot.com/2020/06/blind-test-results-part-ii-is-high.html
https://archimago.blogspot.com/2020/06/blind-test-results-part-iii-is-high.html
or did you have something different in mind @pkane ?!
https://archimago.blogspot.com/2020/01/internet-blind-test-is-high-harmonic.html
with pretty elaborate and very interesting conclusions
https://archimago.blogspot.com/2020/05/blind-test-results-part-i-is-high.html
https://archimago.blogspot.com/2020/06/blind-test-results-part-ii-is-high.html
https://archimago.blogspot.com/2020/06/blind-test-results-part-iii-is-high.html
or did you have something different in mind @pkane ?!
OP
- Thread Starter
- #35
I’m not sure what you’re asking. Did I have something different in mind for doing what?
Oh, sorry. The thread title says "audibility-of-harmonic-distortion-a-blind-test-proposal". Was archimago's BT the one you wanted to have? Or you wanted (and maybe still want) a different kind of BT?
OP
- Thread Starter
- #37
Ah, yes. I originally proposed the test to @Archimago. We collaborated on it, although he did most of the heavy lifting. The results were interesting and enlightening, and I'm really glad we did it.
We did discuss a few other types of blind tests that would be a good follow-up for the future on distortion audibility and preferences. Doing such a test is a lot of work, and my hat is off to Archimago for doing it! All I did was provide the software and some feedback
- Joined
- Oct 25, 2019
- Messages
- 11,112
- Likes
- 14,776
Pfffft, thinly disguised attempt to shield your alter ego. Rumbled. May as well own up to being NWAVguy too.Ah, yes. I originally proposed the test to @Archimago. We collaborated on it, although he did most of the heavy lifting. The results were interesting and enlightening, and I'm really glad we did it.
We did discuss a few other types of blind tests that would be a good follow-up for the future on distortion audibility and preferences. Doing such a test is a lot of work, and my hat is off to Archimago for doing it! All I did was provide the software and some feedback
cool, hope to see some new tests soon.Ah, yes. I originally proposed the test to @Archimago. We collaborated on it, although he did most of the heavy lifting. The results were interesting and enlightening, and I'm really glad we did it.
We did discuss a few other types of blind tests that would be a good follow-up for the future on distortion audibility and preferences. Doing such a test is a lot of work, and my hat is off to Archimago for doing it! All I did was provide the software and some feedback
Archie does a really good job indeed and you shouldn't be so modest about your software (and work). If nothing else, it's a tremendous amount of work that you gave us all for free. Thank you again and good luck with it!
Similar threads
