“Human ears are better than measurements!” / “Humans can hear things we can’t even measure!”

[Barrelhouse Solly]

There is no difference. If you can hear something, you can measure it. If you like a certain characteristic, you can measure that too. Precisely because the instruments we use (including ADCs) are better than human hearing by orders of magnitude.

You can then analyse what "character" you like and what it looks like in the measurements. Be thorough, eliminate your unreliable wetware between your ears as good as possible, and the conclusions will even enable you to buy blindly - or rather deafly - to a degree. It really works once you figured out what your taste technically, physically means.

Exactly.

 

[solderdude]

@solderdude I would not be so strict as to say that amplifiers don't "sound" and only output electrical signals. Yes, that is technically correct. But I think that everybody understands the reality that in the end, we listen to amps through speakers or headphones.

In the end we listen to speakers in a room or headphones on one's head. Sure they are driven by amplifiers but you don't listen to amplifiers.
Amplifiers can (and some do) change the output response audibly. This can be caused by linear distortion (which can even be level dependent) or caused by a (too high for the load) output resistance or from various forms of non-linear distortion or power output may be insufficient above certain levels.
Usually these 'amplifiers' are purpose built to modify sound in certain way(s).
In most cases of half decent amplifiers the distortion is too low to be audible and frequency response is well outside the audible range.
In that case you don't listen to amplifiers nor can you 'hear' an amplifier.

If they produce the same electrical signal which is fed to these transducers, they will produce the same sound under otherwise unchanged boundary conditions (speakers, room, environmental noise, and so on). Therefore, I'm OK to say that amps can and often do "sound identical".

Agreed that this is so. Ampifiers don't 'sound' though nor (should) modify it.
My point is that even if 2 amplifiers measure/perform the same under an identical load a person could still be convinced they 'sound' differently because of many reasons.
This is what perception can do.

I think some of us have used say tone control/EQ/effects and while making slight adjustment of say bass, treble or some effect heard (perceived) a different 'sound' only to find out later the tone control or effect was not enabled and thus nothing changed but still heard a different 'sound'.
Or, play music during the day and play the same music later in the evening and while technically everything remained the same can still hear a different 'sound'.

So I am not disagreeing with you but am of the opinion that using the word 'sound' can be tricky as that is perception related.

 

[MediumRare]

ADCs designed and used for measurements are typically as good or better than ADCs used for recordings in the relevant measurable dimensions (SNR, THD, linearity, flatness of FR).

The idea is that they can definitely fully capture and analyze the characteristics of the device under test (e.g. an amplifier) and any potential audio sample / recording that device might play for the test. So nothing stays hidden from them - there can't be any magic property we miss when measuring, because our measurement instruments are at least as good as those used for the music recording itself.

That will, of course, not be true for all combinations of all ADCs ever used for measurements and recordings. But an Audio Precision analyzer or a Cosmos ADCiso is designed to be used to test and evaluate other ADCs or DACs - so it must outperform them in the relevant metrics, otherwise you couldn't use such an analyzer to judge or further improve the tested device.

AFAIK the measurement device merely needs to meet or exceed the capabilities of the source (e.g. a CD) or even ideal human hearing. The source production equipment is irrelevant.

 

[RandomEar]

AFAIK the measurement device merely needs to meet or exceed the capabilities of the source (e.g. a CD) or even ideal human hearing. The source production equipment is irrelevant.

Yes, the worst element of the chain should be the determinator. But with HighRes audio, extremely performant DACs and all the unrelated nonsense less informed people may come up with, I think it simplifies the argument a lot if we can conclude that our measurement devices are superior to any element of the chain. Be it in the recording studio, on the transport medium / disc or in the playback system.

 

[RandomEar]

[...]
My point is that even if 2 amplifiers measure/perform the same under an identical load a person could still be convinced they 'sound' differently because of many reasons.
This is what perception can do.
[...]
So I am not disagreeing with you but am of the opinion that using the word 'sound' can be tricky as that is perception related.

Yes, that is correct. I used "sound" as in real, actual signal (electrical or acoustic, which you could argue is imprecise again) and not as a perceived subjective quality. That would be better termed "impression of ..." in the context of this thread. But it is true that the word sound is often used ambiguously in the context of HiFi, so people will argue that it "sounds different" to them. Which can be true even though there is no actual difference in the sound being reproduced. Ahhhhh

 

[MediumRare]

Yes, the worst element of the chain should be the determinator. But with HighRes audio, extremely performant DACs and all the unrelated nonsense less informed people may come up with, I think it simplifies the argument a lot if we can conclude that our measurement devices are superior to any element of the chain. Be it in the recording studio, on the transport medium / disc or in the playback system.

I’m confident that’s merely speculative so no need to stretch your point that far. Occam’s razor.

 

[JPA]

In full agreement with the OP.

Live music, in person - always the gold standard though.

Agreed, but I would qualify that by saying unamplified live music.

 

[Ropeburn]

In full agreement with the OP.

Live music, in person - always the gold standard though.

That's weirdly limited. There's a lot of music never played live but simply recorded. There's also a buttload of live concerts and venues that sound seriously bad. And that's supposed to be the gold standard?

 

[ahofer]

I see a lot of discussion above of what constitutes a complete set of measurements, but I don’t think it follows that listening reveals things that measurements can’t reveal, rather the listening might reveal that the right measurement has not been taken.

Perhaps. But none of that suggests that a loudspeaker with poor measurable distortion, non-linearities, or badly uneven dispersion is worth spending any of one’s precious listening time. And it still seems that a conventional measurement suite such as @amirm performs is completely sufficient for electronics.

 

[DWPress]

There's a lot of music never played live but simply recorded. There's also a buttload of live concerts and venues that sound seriously bad.

Ok, I'll simplify: Live music heard with your own ears have not gone through an ADC.

I'll agree there are many bad venues but there are exceptional ones as well. Just sit around a fire pit or quiet back yard and listen to a singer/songwriter or small group and forget about walls of sound reinforcement in a stadium show....

 

[Ropeburn]

Ok, I'll simplify: Live music heard with your own ears have not gone through an ADC.

I'll agree there are many bad venues but there are exceptional ones as well. Just sit around a fire pit or quiet back yard and listen to a singer/songwriter or small group and forget about walls of sound reinforcement in a stadium show....

But DACs are okay? Of digital instruments, which have been staples for decades? Or ADC-DAC chains of digital live mixers and effects, which have been staples for decades? Digital transmission inside the venue, which has been... you get the idea.

Are you talking about exclusively pure analog live music, only that is your precious gold standard? That would be so far removed from live music reality and such a small subset, it's not even funny.

 

[NTK]

Ok, I'll simplify: Live music heard with your own ears have not gone through an ADC.

I'll agree there are many bad venues but there are exceptional ones as well. Just sit around a fire pit or quiet back yard and listen to a singer/songwriter or small group and forget about walls of sound reinforcement in a stadium show....

However, our "memory" of the live event isn't entirely trustworthy. David Griesinger described his experience when comparing reproduction of the binaural recording (made using probe microphones inserted into his ear canals to record the soundwaves as close to his eardrums as possible) to his memory of the event.

 

[Waveform Fidelity]

The importance of achieving the “flattest” frequency response as possible to get a more realistic representation of speech sounds is questionable.

The reality is that most microphones for recording is rather chosen for their characteristics more that “flatness”, and it’s not uncommon that the microphone is chosen by the vocalists based on how they perceive the sound of their voice ”in their head”. The loudspeakers in the studio used for mixing the speech weren’t necessarily very flat in the frequency response, and most of the time, there are some extreme equalization done to the vocal track to ”cut through” in the mix, and not being drowned out by the rest of the instrumentation.

All the above mentioned things makes the frequency ”deviation window” quite large for what will be perceived as a realistic and natural-sounding vocal track, and a loudspeaker with a flat +/-0.5dB frequency response will not necessarily sound more realistic than a loudspeaker with +/-3dB deviations in the response, which is quite common with most loudspeakers.

Yes, I completely agree. When listening to the TV sounds, as one example, one can tell that certain microphones that are selected have a presence hump, or in the case of close-miked lapel microphones how the studio did a bad job at equalizing for the specific body position, but for those broadcasters (TV news) that actually do try to get it correct, the speech sound are very realistic.

Unless we use our speakers for sound re-enforcement we do not have reference to the original speech. When we have parties in our back garden I use my measurement mic and loudspeaker for sound re-enforcement and thats an amazing reference measurement for loudspeaker neutrality… and this is where we might still disagree, flat frequency response must by definition, sound more natural. I am surprised why HIFI speaker manaufacturer’s don’t use such a test. Maybe, manufacturers like KEF when referring to “voicing” perhaps this is what they are actually doing? Its what Wharefdale did decades ago!

 

[MattHooper]

Tangential note:

I often marvel at the audiophile vinyl fans and the incredibly elaborate, ridiculously expensive turntables that they claim open up ever new frontiers of "extracting more from the groove".

I have seen a Neumann cutting lathe in action, and it is a much, much simpler machine than any of these audio jewellery contraptions (though it is heavy and expensive), and the cutting head is moved tangentially across the lacquer.

Funnily enough almost all audiophile tables don't have a tangential arm and thus produce tracking distortion.

:Gulp:

 

[RexrothPigeon]

But DACs are okay? Of digital instruments, which have been staples for decades? Or ADC-DAC chains of digital live mixers and effects, which have been staples for decades? Digital transmission inside the venue, which has been... you get the idea.

Are you talking about exclusively pure analog live music, only that is your precious gold standard? That would be so far removed from live music reality and such a small subset, it's not even funny.

No true Scotsman, basically

 

[rjp]

I see a lot of discussion above of what constitutes a complete set of measurements, but I don’t think it follows that listening reveals things that measurements can’t reveal, rather the listening might reveal that the right measurement has not been taken.

Well said. I agree, and I think this is what most people really mean when they post that their ears hear differences that measurements do not support. They are not really claiming their ears are *better* than measuring devices, but more likely that they are hearing something that has not been measured. The average consumer doesn't have the technical background to realize that this is what they meant to say, so they end up getting ridiculed for claiming to have "ears better than measuring devices", when actually they had a valid point if we took the time to listen and understand. I believe the burden is on us, who understand the technical underpinnings, to show why the measurements we rely on are adequate to support any claims of sonic equality. It would make a great sticky to have a thread like that if we do not already.

 

[DWPress]

Are you talking about exclusively pure analog live music, only that is your precious gold standard?

As I said - I'm not arguing against this thread at all. I completely embrace the fact I have no golden ears or arcane knowledge of hypothetical auditory hallucinations. I've played live, recorded since analog tape days and have done plenty of engineering and mixing too. I've been using pro gear and signal processing for well over a decade in the work.

I guess I'm fortunate enough to have lots of musician friends. I'll also hazard to add that I would rather listen to them acoustically or in a good sounding small venue than listen to one of their albums. Go listen to some music....

However, our "memory" of the live event isn't entirely trustworthy.

I would also argue that part of our memories around this sort of thing are probably highly psychological associations with the experience. Not just auditory but the rest of the senses record there impressions of the experience as well. Did Griesinger enjoy the music when he was recording it, was it a good mix at the sound booth and in his position while recording?

He'd have to stand really really still to get a decent recording, there's a reason tapers have dedicated sections for static mic rigs if the venue allows it.

 

[Hayabusa]

If they produce the same electrical signal which is fed to these transducers

Indeed, but you have to measure the signal at the transducer terminals to be 100% sure the signal is still the same

 

[solderdude]

And it still seems that a conventional measurement suite such as @amirm performs is completely sufficient for electronics.

For checking some basic specs .. yes.
When one wants to know about all relevant measurements that are interesting to fully characterize performance not all required measurements are always done.

This, however, is quite understandable as doing all possible measurements under all kinds of circumstances /conditions as well as doing certain EMC measurements are very time consuming (and expensive) and so is reporting all of that.

It would be the equivalent of doing extensive blind listening tests with various individuals and writing a paper about that.
To make reviews interesting enough and readable one always has to make compromises of what and how to report.

 

[Waveform Fidelity]

Sorry I feel I should make an obvious point

A SPL converted by a microphone to a signal that then feeds a perfect loudspeaker that then produces the identical SPL (as measured by a perfect microphone) doesn’t confirm that measurements detect everything that the ears hear. Yes we all agree the perfect recording-replay chain has that limitation, but many of us have been discussing live sound, and yet still loudspeakers don’t sound like the live instrument either (although they can be pretty close)

What’s missing is the inability of the loudspeaker to replicate the “acoustic size of the source” This is why the recordings of say a grand piano cannot be used to create the same sound over a loudspeaker, and its true for simple sound re-enforcement too.

However, if the industry had the motivation and/or financial means (which it doesn’t) it could embark on some serious R&D in this area in the future to create multi-beam speaker phased array with arbitrary beam patterns and pointing such that speakers (not the way we think of them today) could emulate the acoustic sizes of each of the musical instruments being played. Grand pianos have broad beams, flutes have small beams ,etc

Now before anyone flames me for a wacky idea - think of near real time AI extracting the identity of each instrument in a musical sequence, and computing the phased array coefficients to dynamically change the beam patterns of the large MIMO loudspeaker in near real time. We already do an equivalent in the wireless industry in 5GNR mmW including pointing the beams at different spatial locations. Hek all we really need (I exaggerate) are a few wireless 5GNR antenna engineers to start thinking about this problem in the acoustics world!