Our perception of audio

[fas42]

A trap one has to be careful not to fall into, is thinking that the quality of system playback is invariant, and that perceived differences must be due to one's hearing. I'm afraid all real system are constantly changing in quality, all the time - for myriads of reasons - though the changes may take place over longer periods of time. That the actual, measurable quality you hear at one time matches perfectly what it is on another occasion is a nonsense - I've spent years working on aspects of this " irritating behaviour" ...

An easy one that comes to mind: a friend I visit at times had sound which slowly deteriorated as we listened - over several minutes the sparkle just went out of the playback. He was totally unaware that this was happening, and I pointed it out, indicated what he should be listening for - and then he had no trouble picking it also. Long story short, he had a Naim integrated amp, with one of those "special" Alps blue volume pots - and the track was worn in the region where he normally set the gain. Adjusted system gain settings so that a different section of the pot track was used - problem went away! Which led to him ripping out that part, ordering a replacement, fitting it - and we were all happy again!

 

[tomelex]

When one has to ( is brave enough to ) confront this truism it's a sobering revelation but at the same time liberating for those who are able to accept it. I think most of us here have had such a expirance.

( seems I just plagiarised amir )

Great write up, thanks for the post

As a side note, a few years back I watched a food program that had blind taste tests, a good proportion of the folks involved could not reliably differentiate between chicken, beef and pork!

As a builder and experimenter in many things audio through the years, it is very sobering to swap in different components, sometimes quite expensive or should I say super high quality ones, and find out that you cant hear the difference and at that level cant really measure it either. You can say you have a better sounding widget because you know the component you put in has a lower self noise for example, but bless you if you could actually hear it......folks will not acknowledge that there is a hearing threshold and limits. Learning about electronics and our hearing abilities and self delusions liberated me a long time ago, like when I was in my very early twenties, when a group of us did all kinds of tests with audio gear, as we were all new electronics guys wanting to hone our electronics skills. And, we learned that theoretical and real world are two different things, the real world is messy and fuzzy and shades of gray, in many things, however, despite the theories, most things work according to them. Theory is an accepted explanation...not necessarily the real actual truth of course.

 

[fas42]

As a builder and experimenter in many things audio through the years, it is very sobering to swap in different components, sometimes quite expensive or should I say super high quality ones, and find out that you cant hear the difference and at that level cant really measure it either. You can say you have a better sounding widget because you know the component you put in has a lower self noise for example, but bless you if you could actually hear it......folks will not acknowledge that there is a hearing threshold and limits. Learning about electronics and our hearing abilities and self delusions liberated me a long time ago, like when I was in my very early twenties, when a group of us did all kinds of tests with audio gear, as we were all new electronics guys wanting to hone our electronics skills. And, we learned that theoretical and real world are two different things, the real world is messy and fuzzy and shades of gray, in many things, however, despite the theories, most things work according to them. Theory is an accepted explanation...not necessarily the real actual truth of course.

This probably tells one a lot about the difference in thinking between various people on this forum ... "swap in different components, sometimes quite expensive or should I say super high quality ones" is something I never did, ever. Apart from Van den Hul cables, 30 years ago - a oncer. I just looked at gear, saw what to me were weak links in the chain, strengthened those links - and the sound improved. Ah-ha!! There's a pattern here, I thought - and in one sense that's all I have ever done, since ...

 

[tomelex]

This probably tells one a lot about the difference in thinking between various people on this forum ... "swap in different components, sometimes quite expensive or should I say super high quality ones" is something I never did, ever. Apart from Van den Hul cables, 30 years ago - a oncer. I just looked at gear, saw what to me were weak links in the chain, strengthened those links - and the sound improved. Ah-ha!! There's a pattern here, I thought - and in one sense that's all I have ever done, since ...

Frank, every designer will try things to hear them. It is never a bad idea to put a better quality component in the system, however, not every one results in an audible or measurable improvement. For example, surely, you know that replacing all your carbon comp resistors with newer types with less noise and better long term stability will result in slightly lower noise floor, or depending on your circuit, an actual audible difference. Now, those are weak links, atleast in the old days. You do acknowledge that there are levels of quality of components for santas sake? Same with caps, grades of transistors and self nooise, fuse modulations due to high currents, and more and more, not the least is dealing with fields and their effects on nearby components (I am talking actual electronic parts here more than at system level). Swapping an interconnect can affect sound, however, without measurements, there is no way your hearing can decide you got more accurate to the source, so that part of your style is just pandering to your preferences...I know you proudly do not use measurements, that's a downfall but amateurs can do what they want in any case. High end though is filled with amateurs at the design end of things, AND with truly super skilled engineers too, make no mistake.

 

[fas42]

Of course there are different levels of quality, Tom - which is why we have, say, precision resistors . But it's a case of using them where they may make a difference - I did a DIY chip amp some years ago, used ordinary film resistors, except for the feedback ones - for these most important ones of all, I bought the ones with the best tempcos available, that didn't cost silly money. Now, do I know whether it made any difference to the sound compared to the use of cheap ones? No, I didn't try the execrcise of comparing - technically, this is where things were important, so that's where the money went.

Weak links are where switches are, and cables plugged in. Anything that depends solely on some weak pressure to maintain a good contact sticks out like a sore thumb - electrically, those places are a nightmare - so, I get rid of them.

 

[tomelex]

Of course there are different levels of quality, Tom - which is why we have, say, precision resistors . But it's a case of using them where they may make a difference - I did a DIY chip amp some years ago, used ordinary film resistors, except for the feedback ones - for these most important ones of all, I bought the ones with the best tempcos available, that didn't cost silly money. Now, do I know whether it made any difference to the sound compared to the use of cheap ones? No, I didn't try the execrcise of comparing - technically, this is where things were important, so that's where the money went.

Weak links are where switches are, and cables plugged in. Anything that depends solely on some weak pressure to maintain a good contact sticks out like a sore thumb - electrically, those places are a nightmare - so, I get rid of them.

Best temp co's. Thats one element/tolerance, whats another tolerance to consider for those two gain control resistors in your chip amp? You do rail on and on about it, perhaps you just were not being complete in your example.

Anyway, yes we are in the same boat about mechanical connections.

 

[fas42]

Tom, that's simply a technique for picking a better resistor - higher precision means better everything, it's a convenient way of isolating a higher quality part, from those readily available. Just to clarify, I haven't swapped those resistors on ready made components, so far - I was too busy fixing up more obvious issues!

Regarding measurements, I would happily use something that made it convenient to get numbers on "how good it is" - but I haven't come across anything yet that does that. I keep my eyes open as to what designers who design well regarded kit use, but they all say, "In the end, I just listen to what comes out" - QED ...

We disagree on the disappearing speaker thing - but that's what I use: how close am I to that quality in the reproduction?

 

[tomelex]

Tom, that's simply a technique for picking a better resistor - higher precision means better everything, it's a convenient way of isolating a higher quality part, from those readily available. Just to clarify, I haven't swapped those resistors on ready made components, so far - I was too busy fixing up more obvious issues!

Regarding measurements, I would happily use something that made it convenient to get numbers on "how good it is" - but I haven't come across anything yet that does that. I keep my eyes open as to what designers who design well regarded kit use, but they all say, "In the end, I just listen to what comes out" - QED ...

We disagree on the disappearing speaker thing - but that's what I use: how close am I to that quality in the reproduction?

My comment was about there are many parameters for precision resistors, and tempco is not the main thing I would be after in that spot. If you are swapping out stuff then its not just throw in a precision resistor. There is a lot of research to find exactly the best spec for each device in each spot. While there is no exact measurement for what happens in your ear brain interface or mine, can you tell me what particular parameter of a signal is not measurable? You have already done experiements that proved to yourself that there is a lower limit of hearing, you did that by measuring, by using an attenuator to make your intruding signal smaller and smaller until you could not hear it. We did those tests here on this site, I think most were at about 60db down limit.

The folks on this site know the difference between measurements (the right ones at the right place) and ear/brain preferences. I do not think anyone here schooled in electronics will say that there is some mysterious parameter of a signal or signals that can not be measured.

Your idea that noise is the issue, can you explain then how far down in db the noise is in your system vs the signal amplitude, you claim that when we get a signal the noise increases enough with amplification but yet you have not measured that noise increase with signal increase or do you have links to someone who has done this. Measurments and limits of hearing, they are important. An increase of random noise needs to be pretty high level to be heard against a huge db difference to the signal. You already have ambient noise of atleast 30db in your home. Is analog playback noise the same as digital playback noise?

 

[fas42]

Thanks for those thoughts, Tom.

With regard to measurable parameters, I would never say it's not measurable, not at all - but I would say that the current methods haven't developed beyond what they were decades ago, and this needs to change. Countless times I have come across suggestions for more useful measurement techniques, in comments, papers, research material - the one thing that binds them is the fact that they are never taken up; they're always completely ignored.

The biggest problem is low level noise/distortion that correlates with the signal, in terms of strength and character. Adding an intruding signal which is constant, and non complex, and checking for the audibility of that is not good enough, because our ear/brain hearing system is more sophisticated than that. As mentioned a few times, my best guesstimate is that it's in the 60 to 80dB down range - as a subjective reference marker, it's like hearing tape noise, and even the best units produce noise which is easy to hear, if the volume is turned up.

But the key thing is that tape noise is constant, it does not modulate with the signal strength - hence can be subjectively discarded. The modulated noise which is the real problem has a different character, normally, between analogue and digital - the analogue version is 'nicer' and hence easier to "live with", but digital artifacts typically have a subjectively deadening quality to them, which is why there has been this ongoing resistance to CDs, etc.

 

[amirm]

But the key thing is that tape noise is constant, it does not modulate with the signal strength - hence can be subjectively discarded.

The most prominent version of this is in lossy audio compression. Yet vast majority of audiophiles fail to hear them there even though it is orders of magnitude higher than whatever other situation you are imagining. I am happy to put such test in front of you and see how you do. If you fail, what does it say about your hypothesis?

And this is all hypothesis right? Since you say you can't measure it, you have never worked backward to 1:1 associate whatever you are hearing with correlated noise.

 

[watchnerd]

In addition to the problems of placebo effect / confirmation bias that plague our perception of audio, when run into the diminishing returns engendered not just by real engineering challenges, but by hedonic adaptation.

To quote form Psychology Today (wow, never thought I'd ever write those words...):

"Psychologists call this phenomenon hedonic adaptation—the idea is that no matter how good something makes us feel (or, for the record, how bad), most of the time we drift back to where we started, emotionally-speaking."

The most joy I've ever had listening to music hasn't changed to be higher as life goes on -- the peaks of happiness are the same, whether it was 16 year old me listening to cassettes on a boom box or the multi-kilobucks system I have now.

 

[tomelex]

Thanks for those thoughts, Tom.

With regard to measurable parameters, I would never say it's not measurable, not at all - but I would say that the current methods haven't developed beyond what they were decades ago, and this needs to change. Countless times I have come across suggestions for more useful measurement techniques, in comments, papers, research material - the one thing that binds them is the fact that they are never taken up; they're always completely ignored.

The biggest problem is low level noise/distortion that correlates with the signal, in terms of strength and character. Adding an intruding signal which is constant, and non complex, and checking for the audibility of that is not good enough, because our ear/brain hearing system is more sophisticated than that. As mentioned a few times, my best guesstimate is that it's in the 60 to 80dB down range - as a subjective reference marker, it's like hearing tape noise, and even the best units produce noise which is easy to hear, if the volume is turned up.

But the key thing is that tape noise is constant, it does not modulate with the signal strength - hence can be subjectively discarded. The modulated noise which is the real problem has a different character, normally, between analogue and digital - the analogue version is 'nicer' and hence easier to "live with", but digital artifacts typically have a subjectively deadening quality to them, which is why there has been this ongoing resistance to CDs, etc.

Frank, for an electrical engineer you do not seem to have kept up with things from the last 35 or 40 years mate. (I know you have, but have you paid attention to it?) You have heard of frequency domain analysis and you do know that THD PLUS NOISE is exactly what it says. Just take Nelson Pass, he has numerous patents for eliminating all forms of distortion from op amps, and these patents are like 30 years old now, killing distortions down to nothing. He knows what sells, and its harmonic spray with a slight emphasis on second harmonics, so while he tunes by ear, he knows that making a HALCRO will not sell to the brainwashed wealth clientale who prefer to stay in audio phile cookoo land and keep the churn going. Any real advancement in audio is easily measurable, and while you can decry the standard measurements, real audio designers do a whole lot of different tests, advanced tests, to get what they want. Its the industry you need blame for the same old specs, just like you need to blame the artists for allowing their work to be compressed to hell. Don't blame objectivists who know something about audio science as far as transducers and electronics go, that somehow they and many of us here are lost as to how to move a simple audio signal around an electronic circuit.

If you want to believe that noise is the answer, just do a THD plus NOISE test with your sound card, on your amplifier, figure out what it is at no signal and volume all the way up and running a tone at full power and see what you get, its easy really. And I know you have a sound card, and I know you can do spectrum analysis with it, well, it is the harmonic spray that causes amps to sound different, its known. Really, Amir has presented something for you to do in above post.

If the biggest problem is as you say, then don't add a non complex and constant intruding tone, add as complex a signal as you want and see how far down it will go and what the effects are at high and low levels etc, do the test, show us, put it up here for us to hear with our ears.

 

[fas42]

In addition to the problems of placebo effect / confirmation bias that plague our perception of audio, when run into the diminishing returns engendered not just by real engineering challenges, but by hedonic adaptation.

To quote form Psychology Today (wow, never thought I'd ever write those words...):

"Psychologists call this phenomenon hedonic adaptation—the idea is that no matter how good something makes us feel (or, for the record, how bad), most of the time we drift back to where we started, emotionally-speaking."

The most joy I've ever had listening to music hasn't changed to be higher as life goes on -- the peaks of happiness are the same, whether it was 16 year old me listening to cassettes on a boom box or the multi-kilobucks system I have now.

People need to be reminded again where I got on board this current journey. The sort of quality of sound that's currently being discussed on WBF, by Steve Williams, Leif S, mullard88, etc, happened to me 30 years ago - no one else at the time, those decades ago, was talking about such experiences - they were too busy declaring that digital sound was the Work of the Devil! Again, I saw colour TV, while everyone was watching B&W - I can't unsee that, just because the concept disturbs people ...

 

[Blumlein 88]

Thanks for those thoughts, Tom.

With regard to measurable parameters, I would never say it's not measurable, not at all - but I would say that the current methods haven't developed beyond what they were decades ago, and this needs to change. Countless times I have come across suggestions for more useful measurement techniques, in comments, papers, research material - the one thing that binds them is the fact that they are never taken up; they're always completely ignored.

The biggest problem is low level noise/distortion that correlates with the signal, in terms of strength and character. Adding an intruding signal which is constant, and non complex, and checking for the audibility of that is not good enough, because our ear/brain hearing system is more sophisticated than that. As mentioned a few times, my best guesstimate is that it's in the 60 to 80dB down range - as a subjective reference marker, it's like hearing tape noise, and even the best units produce noise which is easy to hear, if the volume is turned up.

But the key thing is that tape noise is constant, it does not modulate with the signal strength - hence can be subjectively discarded. The modulated noise which is the real problem has a different character, normally, between analogue and digital - the analogue version is 'nicer' and hence easier to "live with", but digital artifacts typically have a subjectively deadening quality to them, which is why there has been this ongoing resistance to CDs, etc.

As tomelex wrote there is THD+N. There are measurements where that is done in stepped fashion with signal level varying from low to high to see if the noise floor is modulated with signal level. Those can also be stepped with multiple tones.

If this won't measure what you think you are hearing then what will? What sorts of more useful measurement techniques have you come across that are never taken up?

 

[Blumlein 88]

People need to be reminded again where I got on board this current journey. The sort of quality of sound that's currently being discussed on WBF, by Steve Williams, Leif S, mullard88, etc, happened to me 30 years ago - no one else at the time, those decades ago, was talking about such experiences - they were too busy declaring that digital sound was the Work of the Devil! Again, I saw colour TV, while everyone was watching B&W - I can't unsee that, just because the concept disturbs people ...

Well what needs to be determined is whether or not you are seeing color TV while watching a B&W set?

 

[watchnerd]

People need to be reminded again where I got on board this current journey. The sort of quality of sound that's currently being discussed on WBF, by Steve Williams, Leif S, mullard88, etc, happened to me 30 years ago - no one else at the time, those decades ago, was talking about such experiences - they were too busy declaring that digital sound was the Work of the Devil! Again, I saw colour TV, while everyone was watching B&W - I can't unsee that, just because the concept disturbs people ...

I have no idea who those WBFers are. Or see any reasons I should care what creative prose they choose to post somewhere.

I don't spend time there.

 

[fas42]

He knows what sells, and its harmonic spray with a slight emphasis on second harmonics, so while he tunes by ear, he knows that making a HALCRO will not sell to the brainwashed wealth clientale who prefer to stay in audio phile cookoo land and keep the churn going

A story about Halcro - we had a super high end show in Sydney about 15 years ago - and a demo system used the very best combo of Halcro. From vinyl, the sound was exceptional, probably the best LP sound I've heard; from digital, via the best dCS setup available, the sound was awful! Vivaldi, via a CD I owned, it had everything wrong about it, all the things that people who point gleefully to the "defects of digital" would have applauded!

If you want to believe that noise is the answer, just do a THD plus NOISE test with your sound card, on your amplifier, figure out what it is at no signal and volume all the way up and running a tone at full power and see what you get, its easy really. And I know you have a sound card, and I know you can do spectrum analysis with it, well, it is the harmonic spray that causes amps to sound different, its known. Really, Amir has presented something for you to do in above post.

If the biggest problem is as you say, then don't add a non complex and constant intruding tone, add as complex a signal as you want and see how far down it will go and what the effects are at high and low levels etc, do the test, show us, put it up here for us to hear with our ears.

If the pressure goes on, you might tempt me to create such a test - you have been warned!

 

[fas42]

Well what needs to be determined is whether or not you are seeing color TV while watching a B&W set?

Well, how it works is that the set is really colour, and always has been - but you probably are familiar with poor TV transmission, where a colour signal drops back to B&W presentation, because that's the best it can do, at that moment. What I got back then was the set snapping into colour, for a while, and then lapsing into B&W, most of the time - what I had to do was work about how to improve the transmission so that the set went colour, and stayed there
</Analogy>

 

[amirm]

People need to be reminded again where I got on board this current journey. The sort of quality of sound that's currently being discussed on WBF, by Steve Williams, Leif S, mullard88, etc, happened to me 30 years ago - no one else at the time, those decades ago, was talking about such experiences - they were too busy declaring that digital sound was the Work of the Devil! Again, I saw colour TV, while everyone was watching B&W - I can't unsee that, just because the concept disturbs people ...

You're da' man Frank....

Which stock should I buy Frank?

 

[fas42]

Ummmm ... https://finance.yahoo.com/quote/FAS?p=FAS