AES Paper Digest: Do Audio Op-amps Sound Different?

[SIY]

There is an enormous amount of literature on blind tasting evaluations of wine. If you have an interest in it, start with Ann Noble's papers- she was the goddess of scientific sensory evaluation and its relationship with commerce. Likewise, staying on the UC Davis track, Amerine's "Wines: Their Sensory Evaluation," while rather old, is a great introduction to how this is done professionally.

If audio reviewers had to pass the same sort of tests that professional wine evaluators and certification holders (e.g., WCET, MW, MS), there would be a lot fewer audio reviewers!

 

[Krunok]

If you question the use of the ADC in the mentioned test setup, because it might mask differences between the DUTs, you´ve to suspect that all the other parts of the reproduction environment could provide the same (or similar) kind of masking . So it´s an inherent problem of the perceptual sensory test approach for audio.

I fully agree that all the other parts of the playback chain are participating in masking, I merely wanted to stress that loudspeakers participate by far the most.

 

[oivavoi]

I fully agree that all the other parts of the playback chain are participating in masking, I merely wanted to stress that loudspeakers participate by far the most.

Obviously. For the house I'm moving into, where the plan is to have two systems in different rooms with different speaker directivities (which suits different kinds of music IME/IMO), the ratio of investment on speakers/transducers vs electronics is about 15 to 1, if I'm not mistaken.

 

[Krunok]

There is an enormous amount of literature on blind tasting evaluations of wine. If you have an interest in it, start with Ann Noble's papers- she was the goddess of scientific sensory evaluation and its relationship with commerce. Likewise, staying on the UC Davis track, Amerine's "Wines: Their Sensory Evaluation," while rather old, is a great introduction to how this is done professionally.

If audio reviewers had to pass the same sort of tests that professional wine evaluators and certification holders (e.g., WCET, MW, MS), there would be a lot fewer audio reviewers!

Oh, I'm familiar with her whine wheel but I still suggest you post a scientifically based blind test as I have never really seen one. But please do it in a separate topic so that we avoid the wrath of the Moderator!

 

[SIY]

As a retired accredited ISO 9000 series QMS Senior Auditor I am well versed on the subject of 'quality'. The technical definition of quality is that something meets its specification. If its specification is low-level and the 'product' complies then it meets its quality criteria.

I must have been unclear. I was speaking of "quality" in the context of wine, not "quality" as a general concept (Though I don't have your background, I set up and ran my company's ISO9001 QMS for some years, so have some familiarity with quality in the general sense, as opposed to the specifics of wine quality). As an example, a UC Davis-trained winemaker would find any detectable hint of brettanomyces as a quality flaw, whereas a lot of wine drinkers like me find it an often-interesting aroma note in certain wines and a positive for quality in those wines. There is, of course, no specification for that!

 

[SIY]

Oh, I'm familiar with her whine wheel but I still suggest you post a scientifically based blind test as I have never really seen one. But please do it in a separate topic so that we avoid the wrath of the Moderator!

Pick up any random issue of American Journal of Enology and Viticulture.

 

[Wombat]

Which misses imo the point that i tried to address (while strangely supporting it nevertheless).

If you question the use of the ADC in the mentioned test setup, because it might mask differences between the DUTs, you´ve to suspect that all the other parts of the reproduction environment could provide the same (or similar) kind of masking . So it´s an inherent problem of the perceptual sensory test approach for audio.

Therefore we should always be hesitant/careful (when)/ to draw far reaching categorical conclusions from test results.

Edit: Nothing wrong with (even) strong beliefs provided that you keep in mind that it is a belief.
"Much less" leaves room in a sufficiently way, although from my experience it depends if "much less" also might mean "i can´t stand it" .

Are you saying that the miniscule distortion levels of well designed signal processors can be distinguished, by listening, from the inherent loudspeaker distortions?

 

[oivavoi]

Are you saying that the miniscule distortion levels of well designed signal processors can be distinguished, by listening, from the inherent loudspeaker distortions?

Earl Geddes has apparently done blind tests which shows that we are much more sensitive to distortion from electronics than "mechanical" distortion in transducers. Not sure exactly why though. I'm also not sure whether these tests have been published, or whether they have been replicated.

 

[Wombat]

Pick up any random issue of American Journal of Enology and Viticulture.

I could pick up a journal of the homeopathy trade and find opinions that support the practise. Some credible supporting information would be more pertinent.

 

[Wombat]

Earl Geddes has apparently done blind tests which shows that we are much more sensitive to distortion from electronics than "mechanical" distortion in transducers. Not sure exactly why though. I'm also not sure whether these tests have been published, or whether they have been replicated.

Why mention them then?

This sort of response is problematic on a science-based forum.

 

[Krunok]

Earl Geddes has apparently done blind tests which shows that we are much more sensitive to distortion from electronics than "mechanical" distortion in transducers.

Sorry, but we are assuming he didn't, unless you can prove otherwise by showing us his actual work. Let us not repeat ourselves..

 

[Krunok]

Pick up any random issue of American Journal of Enology and Viticulture.

I highly suspect they print scientific papers there..

 

[Krunok]

Are you saying that the miniscule distortion levels of well designed signal processors can be distinguished, by listening, from the inherent loudspeaker distortions?

Finally! THAT is the question for which we're looking an answer here!

 

[Jakob1863]

I highly suspect they print scientific papers there..

Don´t want to be offensive, but you can´t have it both ways. If assuming or mentioning of things without "proof" isn´t "allowed" you simply can´t post the statement cited above.

 

[Jakob1863]

Are you saying that the miniscule distortion levels of well designed signal processors can be distinguished, by listening, from the inherent loudspeaker distortions?

Afair i didn´t post an assertion about the reason for my perception, just mentioned that i - despite the fact that in most cases loudspeaker and room distortion are much larger - i´m able to differentiate (most likely) between to DUTs although both are of good quality, which means both have measured numbers well below the known thresholds of hearing.

 

[Krunok]

Don´t want to be offensive, but you can´t have it both ways. If assuming or mentioning of things without "proof" isn´t "allowed" you simply can´t post the statement cited above.

Hahaha - no offense taken, I'm merely waiting for @SIY to post a link as he was the one who mentioned it. I think we should all be able to post a link to an article we're quoting otherwise it's hearsay..

 

[Krunok]

Afair i didn´t post an assertion about the reason for my perception, just mentioned that i - despite the fact that in most cases loudspeaker and room distortion are much larger - i´m able to differentiate (most likely) between to DUTs although both are of good quality, which means both have measured numbers well below the known thresholds of hearing.

"Most likely"? Are you able or are you not able? How did you actually tested yourself so that you got the "most likely able" as a result?

 

[oivavoi]

Why mention them then?

This sort of response is problematic on a science-based forum.

Sorry, but we are assuming he didn't, unless you can prove otherwise by showing us his actual work. Let us not repeat ourselves..

This kind of response is kinda tedious. Applied psychoacoustics with regards to sound reproduction is a small, small, actually tiny scientific field, compared to most other fields. The number of replications of any of the existing studies are close to zero (with a couple of exceptions). Are you seriously claiming that every single post on this forum should have a reference to a peer-reviewed publication? In that case, the forum would die down very quickly.

Also, this kind of debate is actually how science is done. This is how scientists communicate between themselves. They mention possibilities and hypotheses, and take it from there (and ultimately try to measure and confirm/disconfirm). It is not against the spirit of science to say things one is not 100 sure about. Quite on the contrary, as long as one is clear about what the epistemic status of the claim is (in this case I was clear that I was not sure about the epistemic status of the claim).

A short google search led me to this interview though, where dr. Geddes says exactly what I said he said:

Our studies indicated that distortion in a loudspeaker is not likely to be a major factor as long as the loudspeaker is operated within its design limits. This is because low orders of distortion (2nd, 3rd, etc.) are not highly audible because of masking. Loudspeakers can essentially only exhibit low orders of distortion because the higher orders require large accelerations, i.e. large forces. Loudspeakers do exhibit very large amounts of low order distortion but not high orders of distortion (6th, 7th, etc.) – as long as they are not overdriven or have design flaws like Rub and Buzz – but the low orders are simply not audible. Using a well-made driver within its design limits lets one completely ignore the issue of nonlinear distortion in a loudspeaker.

This is not true at all for electronics. Electronics can generate very high orders of nonlinearity such as crossover distortion or clipping. One must be very careful in electronics design to prevent these higher orders from occurring especially at low levels. The problem is that this problem is never evaluated for electronics (well certainly never shown) and I doubt that it is even tested very often. What they do show is THD as a function of level but fail to note if the low level result is for crossover distortion or noise. If it is crossover distortion (which is very high order at very low levels) then it is highly audible even at fractions of a percent. That’s the problem with THD, it just does not show what we need to know.

Searching a bit more, I found this AES convention paper written by Geddes and his wife: http://p3dal.com/Files/articles/AuditoryPerceptionNonlinearDistortion.pdf

But alas! It is only a convention paper, it has not been through peer review! So what does your rulebook say in this case, am I allowed to post it or not?

 

[Wombat]

Afair i didn´t post an assertion about the reason for my perception, just mentioned that i - despite the fact that in most cases loudspeaker and room distortion are much larger - i´m able to differentiate (most likely) between to DUTs although both are of good quality, which means both have measured numbers well below the known thresholds of hearing.

(most likely) ?

 

[Arnold Krueger]

Earl Geddes has apparently done blind tests which shows that we are much more sensitive to distortion from electronics than "mechanical" distortion in transducers. Not sure exactly why though. I'm also not sure whether these tests have been published, or whether they have been replicated.

Earl found that one form of nonlinear distortion, crossover distortion in poorly designed class ab amplifiers that experience bias failure and slip into class b might evidence this problem.

This type of distortion is shown as item B in Figure 1 in the paper you referenced. FWIW I think the paper is just fine, but Earl is a friend.

One thing that many might need to understand is that Earl is primarily interested in loudspeakers, and is among those of us who see amplifiers (and DACs) as an often well-solved problem.

OTOH many of us including Earl see audible nonlinear distortion in speakers as something that will be with us for a long time, and therefore it is very intersting.

This problem is old news because it was experienced and documented long before electronic amplification was even possible. It reasserted itself in first generation SS amplifiers. By the third generation SS power amp biasing circuits in mainstream gear was pretty consistent and reliable, even in low priced amps.

I don't think it has been mentioned but sticking due to a rubbing voice coil might be more common today than bias failure in amplifiers. Same basic distortion. With the resuscitation of R2R converters, those boutique DACS are also susceptible.