Yet the audibility of that level difference has been demonstrated repeatedly. Huh.
Let me just get this straight: someone who thought he could hear DAC differences tried again with a good level match and found that the differences disappeared, so now he believes that many DACs are transparent. Someone else did a similar thing with a poor level match (as he considers it relatively unimportant) and found that the differences were still audible, so he believes that most DACs sound different (and he can tell the difference between a DAC and vinyl!). Which of these should I believe?
None of them because you still know next to nothing about their tests....
Could you cite some publications with real data?
None of them!
Repeated quick A/B listening tests, blind or not, level matched or not, rely on human memory to recall the previous session, that in itself corrupts your results.
This is why, outside of a lab (and low class hifi shops) it becomes meaningless, unless the differences are indeed night'nday.
Only long sessions (weeks) with each unit can tell someone, how a piece of hifi 'sounds' to his ears.
Sometimes you think you heard something, but then you realize it isn't there.
Sometimes you miss something, but then you realize there is something there.
A human ear is not a precision lab measurement tool, certain things it is oblivious too, certain other, it focuses on, but ultimately the EARS are the boss!
If you don't hear any differences, fine, good for you, but to tell others they are imagining it, won't wash.
Even if some maybe.
0.2dB of level difference on the entire frequency band?
Was it done in an acoustically sealed capsule or someone's living room?
I am only asking because the background noise in an average living room can vary a few dB's from minute to minute, and to detect 0.2dB of level difference needs some real 'Golden Ears'
Perhaps you could points us to some article regarding this.
From my time in recording studios I've learned one thing for myself:
If differences do not show up in quick A/B comparisons they are likely not there. Longer term listening is useless to establish those since our ears are very adaptable and we get used to an items specific sound (different microphones compared to the real thing mostly in my case) very quickly.
Yes "focus" or "attention" or "awareness" plays a role.
No, i am not trying to argue in that way, it happens purely in your imagination.
If you stick to the things i´ve really written and leave your own assumptions aside, you´ll imho surely notice that.
You missed that i mentioned the experiments for "inattentional deafness", some examples i´ve cited several times in the past.
Psychophysics, cognitive psychology or behavorial science are doing these experiments for over 100 years and therefore we know a lot about distractors/confounders/bias mechanism, but most of it remains neglected in "normal" listening tests.
"Almost everyone listens to sound most of the time, so there is often an opinion that the evaluation of audio qualitymust be a trivial matter. This frequently leads to a serious underestimation of the magnitude of the task associated with formal evaluations of audio quality, which can lead to compromised evaluations and consequently the poor quality of results. Such a lack of good scientific practise is further emphasised when results are reported in journals or at international conferences and leads to a spread of scientific darkness instead of light."
(Bech/Zacharov; Perceptual Audio Evaluation–Theory, Method and Application)
(See my remark wrt inattentional deafness above)
Perception is seperated from the physiological reaction to an external stimulus.
And as we know already that "vision" behaves in that way and the "tactil sense" as well (think of the magician who takes your wrist clock) it would be more sursprising if our hearing sense would not work in a similar manner.
Although they differ in detail (resolution, bandwidth dynamic range) , the underlying mechanisms of processing the input are quite similar for our senses. For example Stevens Handbook of experimental psychology Vol. 1 offers a lot of information.
Agree totally in your case.
Recording studios goal is to appeal to majority, and to scrutinize too much is counter productive, also I believe it depends on the material too.
If an engineer is recording an orchestra, or Rita Ora.
Much akin to cooking for a chain restaurant, or cooking for home, your standards vary.
But here we are discussing, not cooking for a restaurant, but to choose one to dine at, day in day out.
Differences in well engineered DAC's are getting smaller and smaller, and there is that 'diminishing returns' curve.
They may be diminishing! but they are there to smaller and smaller extent as you go up the ladder.
I'm sorry, you did say 'in recording studios', didn't you?
Perhaps you could explain in detail, what conditions you are were talking about.
OK I get you better.
But aren't different mic.s made for different situations?
And don't you check these things on studio's monitor speakers amps?
Were those monitoring/listening equipments, chosen in a quick A/B tests, or were they chosen very carefully at length?
Wouldn't an engineer get used to his monitoring/recording equipment, so he could tell quickly if something is sounding right?
The engineers have been measuring, listening and re-measuring, coming up with new ways to measure things they didn't used to, all the time.
To disregard something in a quick A/B test is not scientific.
That's curious; I thought short-term auditory memory was the only way to enable results.KenTajalli said:
Back to preference again! I assume that in this context "meaningless" might mean 'full of meaning, but delivering a different result from my expectations'.
I think SY's point is not that 0.2dB level difference can be reliably heard as a level difference, but that it can often be heard even though commonly attributed to a quality difference.
No. Once again there is confusion. Cooking must be compared to the musical performance, not the accurate reproduction of that performance. Accurate sound reproduction - which is what this thread must be about, if it is to mean anything at all - is akin to a food technologist trying to reliably reproduce your meal in a factory so that every sample tastes exactly the same as the prototype you cooked at home. Success is marked not by tasters expressing a preference, but blind tasters not being able to distinguish between the prototype you cooked at home and a random sample from the factory.
In a lab, at designing point, yes. as an 'acid-test' of a final production piece of equipment, No.
Yes and No! some people argue, if I can not measure it in my lab, it does not exist, I am arguing, find a way of measuring it, if enough ears say it does. To think we already are measuring everything of relevance is .....!
Meaningless means, Joe Public does not care about such results, nor long term success of the equipment.
Since it can not be controlled (at home) asking someone if they conducted their level-matched A/B tests at 'home' at resolutions of 0.2dB or better, is not valid. read back and see how the 0.2db figure came to be discussed.
Cooking for a restaurant also means repeating your recipe!
What happens if a sizable group of 'tasters' say it tastes different?
it is being taken for granted that blind tests have irrefutably proven that DACs sound the same, have they?
Are you telling me, that all those high-end DAC manufacturers are simply fooling some of the people all of the time?
And more people and respectable manufacturers keep joining in?
Find a way of measuring it, if you fail, try again.
Last edited:
So you didn't really level match. That explains why you heard a difference between DACs. I've experienced that too when levels aren't matched even while listening to same DAC. Something for you to think about.
measuring the V of a test signal can be done as well as you can afford, matching even better since matching only requires a stable, repeatable measurement of sufficient resolution and not national lab traceable accuracy
in fact most soundcard's ADC are fine as long as you're comparing measurements from the same channel - R,L channels can have gain mismatch of few % ~ few x 0.1 dB
and even that can be locally calibrated for orders better matching valid for useful time between cals
measuring SPL is typically less certain - the easy assumption that not being pushed hard your speakers or headphones should put out the same sound fed the same V - as least for short term AB/X comparisons
in fact most soundcard's ADC are fine as long as you're comparing measurements from the same channel - R,L channels can have gain mismatch of few % ~ few x 0.1 dB
and even that can be locally calibrated for orders better matching valid for useful time between cals
measuring SPL is typically less certain - the easy assumption that not being pushed hard your speakers or headphones should put out the same sound fed the same V - as least for short term AB/X comparisons
Last edited: