Pseudoscience in audio - Milind N. Kunchur

[pkane2001]

34 minutes ago, semente said:

Which one, THD+N?

 

No, noise + various levels of nonlinearity + jitter

[fas42]

11 hours ago, pkane2001 said:

 

It's not a problem, Frank. I know what level distortion I can hear because I specifically tested for it. The problem is that you can imagine that you can hear levels of distortion below any human capability, and that is enough for you to believe that you do.

 

As semente is implying, you are relying on there being just the well defined measures of distortion at play; which means you can specifically test for that. Various types of electrical noise, and not the neat stuff bundled up with figures like THD+N, play a vital role in determining the subjective quality of playback; an excess of this degradation blurs the essential low level information, especially in digital replay, and causes the presentation to be lifeless, small, having little appeal as an exercise to listen to.

 

The impact of this is never measured, is swept under the carpet by those who wish to simplify the situation; but those who are keenly aware of these factors strive constantly to improve matters - anyone who is familiar with how far SQ can be evolved will never accept the severely compromised standard that most people live with ...

[pkane2001]

1 hour ago, fas42 said:

 

As semente is implying, you are relying on there being just the well defined measures of distortion at play; which means you can specifically test for that. Various types of electrical noise, and not the neat stuff bundled up with figures like THD+N, play a vital role in determining the subjective quality of playback; an excess of this degradation blurs the essential low level information, especially in digital replay, and causes the presentation to be lifeless, small, having little appeal as an exercise to listen to.

 

The impact of this is never measured, is swept under the carpet by those who wish to simplify the situation; but those who are keenly aware of these factors strive constantly to improve matters - anyone who is familiar with how far SQ can be evolved will never accept the severely compromised standard that most people live with ...

 

I think you know me well enough by now that you'd know I don't do things because they are easy :) In my testing, I generate a true nonlinearity, not just a simple harmonic distortion. In all respects it acts like the real thing, generating all the same THD and IMD products, aliasing, and anything else a true nonlinearity would generate with real music. Combine that with various types of noise at any desired level, correlated and uncorrelated jitter, and a configurable, band-limited feedback loop and you can pretty much simulate any distortion type and level. 

 

When you describe what you hear, you are describing what your brain interprets the sound to be. This is a far cry from the real thing, and is easy to demonstrate in a properly controlled, ears-only test. Until you eliminate the very high possibility that your perception is faulty, you'll never know the effect of any of your tweaks on sound... only on your imagination.

[fas42]

13 minutes ago, pkane2001 said:

 

I think you know me well enough by now that you'd know I don't do things because they are easy :) In my testing, I generate a true nonlinearity, not just a simple harmonic distortion. In all respects it acts like the real thing, generating all the same THD and IMD products, aliasing, and anything else a true nonlinearity would generate with real music. Combine that with various types of noise at any desired level, correlated and uncorrelated jitter, and a configurable, band-limited feedback loop and you can pretty much simulate any distortion type and level. 

 

Yep, no worries there! However, the closest you get to simulating noise issues is with the jitter component; my current system has an ongoing issue where non-optimal sitting of a CD in the player means too much jitter is passed to the speaker electronics; trivially easy to hear the degradation when not at its best. Which will be resolved by inserting a DIY signal regenerator into the link between player and speakers.

 

13 minutes ago, pkane2001 said:

 

When you describe what you hear, you are describing what your brain interprets the sound to be. This is a far cry from the real thing, and is easy to demonstrate in a properly controlled, ears-only test. Until you eliminate the very high possibility that your perception is faulty, you'll never know the effect of any of your tweaks on sound... only on your imagination.

 

I only have to open the CD drawer, slightly adjust the disk, and play again, to easily hear what too much noise does to the SQ - you might as well tell me I'm imagining the colour of blue, when I look at the sky. Other noise reduction measures just progress the same type of improvement.

 

Once you've got a rig into a good enough state, it's remarkably easy to pick faulty playback - the latter, of course, has to be eliminated ... you're there, when no more flaws can be heard. Exactly the same as picking an MP3 version of some track, against the original - a casual listener won't hear it, but anyone who has their attention drawn to precisely what the lossy copy is getting wrong, will be able to pick it, every time.

 

 

[pkane2001]

16 hours ago, fas42 said:

Once you've got a rig into a good enough state, it's remarkably easy to pick faulty playback - the latter, of course, has to be eliminated ... you're there, when no more flaws can be heard. Exactly the same as picking an MP3 version of some track, against the original - a casual listener won't hear it, but anyone who has their attention drawn to precisely what the lossy copy is getting wrong, will be able to pick it, every time.

 

It's quite the opposite: once you know how to properly listen for faults, you realize that many things you thought were real were all in your head. Try picking the up the differences every time without knowing what device/track is playing when only very minor imperfections are there, especially those that are below known audibility levels. 

 

[fas42]

3 hours ago, pkane2001 said:

 

It's quite the opposite: once you know how to properly listen for faults, you realize that many things you thought were real were all in your head. Try picking the up the differences every time without knowing what device/track is playing when only very minor imperfections are there, especially those that are below known audibility levels. 

 

 

Meaning, audible differences between an original, and MP3 encoding, are all in your head?

 

Same style of mechanism as used by astronomers to pick changes in star fields is best - A and B can be rapidly switched into the vision, and the image glitch where a change has occurred is immediately picked up by the eye - I suggest, like perhaps you would, that someone merely stare at each one for long periods of time, to work out if there is a difference, .

 

Yes, of course human hearing adapts - that's why you can listen to an appallingly distorted copy of music you love, and still enjoy it. But that's not the point. What you are after is the elimination of flaws, any degradation of the sound, and so you use techniques that highlight that there is a problem. And you do this so that it no longer sounds like ho-hum hifi, but instead fully engages the emotions ...

[pkane2001]

25 minutes ago, fas42 said:

 

Meaning, audible differences between an original, and MP3 encoding, are all in your head?

 

Same style of mechanism as used by astronomers to pick changes in star fields is best - A and B can be rapidly switched into the vision, and the image glitch where a change has occurred is immediately picked up by the eye - I suggest, like perhaps you would, that someone merely stare at each one for long periods of time, to work out if there is a difference, .

 

Yes, of course human hearing adapts - that's why you can listen to an appallingly distorted copy of music you love, and still enjoy it. But that's not the point. What you are after is the elimination of flaws, any degradation of the sound, and so you use techniques that highlight that there is a problem. And you do this so that it no longer sounds like ho-hum hifi, but instead fully engages the emotions ...

 

MP3 was something you brought up. We were discussing your tweaks... unless you started making your own codecs???  ;)

I did spend some time in late 80's writing both, lossy and lossless audio compression algorithms. Listening to the different algorithms and experimenting made me really sensitive to a lot of the MP3 artifacts, especially as the format was initially released at low bit rates. Couldn't listen to MP3 for a number of years because of this.

 

You're changing the subject. Eye and ear are not the same, and the eye can be fooled just as easily. My emotions are usually engaged by the music I listen to, not by the equipment or the tweaks. But maybe that's different for you...

 

 

 

 

[fas42]

1 hour ago, pkane2001 said:

 

You're changing the subject. Eye and ear are not the same, and the eye can be fooled just as easily. My emotions are usually engaged by the music I listen to, not by the equipment or the tweaks. But maybe that's different for you...

 

 

The "uncanny valley" factor kicks in, for me. That is, obviously low quality playback is not an issue - the brain doesn't try and treat it as anything more than a cartoon representation of live music - all's well. But when ambitious replay gets it mighty close, but a niggling discrepancy is very clear, then it becomes disturbing - it's also the mosquito effect; once you aware it's "in the room", then you are constantly reminded of that fact ... and you can't talk yourself out of not noticing it.

 

Just like sub-par MP3 ...

[pkane2001]

14 hours ago, fas42 said:

 

The "uncanny valley" factor kicks in, for me. That is, obviously low quality playback is not an issue - the brain doesn't try and treat it as anything more than a cartoon representation of live music - all's well. But when ambitious replay gets it mighty close, but a niggling discrepancy is very clear, then it becomes disturbing - it's also the mosquito effect; once you aware it's "in the room", then you are constantly reminded of that fact ... and you can't talk yourself out of not noticing it.

 

Just like sub-par MP3 ...

 

If you can imagine it, it must be real? Unfortunately that's how most fairy tales and conspiracy theories start.

[fas42]

7 hours ago, pkane2001 said:

 

If you can imagine it, it must be real? Unfortunately that's how most fairy tales and conspiracy theories start.

 

There has been plenty of research which demonstrates that the quality of the signals inside of us, what the nature of the electrical waveforms are like, is extremely poor. So, yes, we have to 'imagine' what it's really like, based on our previous experiences ... how can that work? Well, the ear/brain looks for differences; anything that doesn't fit into the expected pattern - and we change our 'mirage', to suit ... that's how we stumble forward in life ... .

 

Anything that works that way can go off the rails - very easily. Hence, "fairy tales and conspiracy theories" - it's just part and parcel of being human.

[pkane2001]

5 hours ago, fas42 said:

Hence, "fairy tales and conspiracy theories" - it's just part and parcel of being human.

 

True, but being human is also about trying to understand what is real and what is imagined. Thankfully, science has made some significant progress in this regard in the last 300 years or so.

 

[fas42]

1 hour ago, pkane2001 said:

 

True, but being human is also about trying to understand what is real and what is imagined. Thankfully, science has made some significant progress in this regard in the last 300 years or so.

 

 

The frontiers of human knowledge, at any point in time, are always a good source of examples where the real and imagined intertwine, are actually a blurred hodgepodge - say, dark matter, dark energy: could always turn out that there is nothing in it; neither exist: they turn out to be a crutch, because our understanding of "how everything works" has not evolved enough.

[pkane2001]

5 hours ago, fas42 said:

 

The frontiers of human knowledge, at any point in time, are always a good source of examples where the real and imagined intertwine, are actually a blurred hodgepodge - say, dark matter, dark energy: could always turn out that there is nothing in it; neither exist: they turn out to be a crutch, because our understanding of "how everything works" has not evolved enough.

 

Good thing that audio electronics is as far from the "frontier of human knowledge" as you can get! It's a well established discipline with a lot of known facts, results, and understanding and good engineering. Just because you can imagine that things are much more complicated doesn't make them so.

 

[botrytis]

On 8/16/2023 at 8:24 AM, Butsers said:

I am not talking numbers here, i am talking ears.
I never said i think numbers are ignorant, i think that people who only look at numbers and don't even bother to try and listen to some of the things they think they have an opinion about are ignorant.

 Ears can be fooled. It is what they use to sell snake oil.

[fas42]

11 hours ago, pkane2001 said:

 

Good thing that audio electronics is as far from the "frontier of human knowledge" as you can get! It's a well established discipline with a lot of known facts, results, and understanding and good engineering. Just because you can imagine that things are much more complicated doesn't make them so.

 

 

Things in audio are not complicated - but it's a field where "near enough is good enough". Which is why even in the latest audio shows every rig sounds different from the next ... "Accuracy??! Stuff accuracy!" Being true to the recording is generally regarded as equal to discovering how the universe works; an exercise that will approach an infinity of time to achieve ... much more exciting engaging in flavouring gambits, .

 

Removing the distortion signature of the playback chain is only tackled by a very small number of people, as a serious endeavour - if that's a well established discipline, well ...

[pkane2001]

2 hours ago, fas42 said:

Removing the distortion signature of the playback chain is only tackled by a very small number of people, as a serious endeavour - if that's a well established discipline, well ...

 

Hmm, it is. Distortion is well defined. Sources of distortion are well known. Distortion measurements and mitigations are well documented. The only thing missing from this wealth of knowledge is your method of detecting distortion: by listening with the full knowledge of which device is playing. That's one thing that is very well known to cause false detections in all humans, and for this reason isn't used by any peer-reviewed/published scientists, unless they specifically want to demonstrate how useless this approach really is. Mark my words, as long as you continue to use this method, you'll never find the tweak solution to the distortion problem, simply because you are not hearing distortion. You're imagining it. 

[fas42]

Paul, I give you full credit for working so hard at pushing your agenda ... "people will find it impossible to hear distortion, unless at least one measurement can 'verify' it; then, it will trivially obvious to anyone listening ..."

 

Of course, I known the 'solution' for over 35 years ... the in between times have been a steady evolution of understanding, and building upon methods, to get answers faster, for them to be more robust, and to handle more scenarios.

 

Hearing distortion? So simple - get any rig, put on a 'difficult' recording, for that setup; wind up the volume, and then put your ear near the treble/midrange drivers of one speaker - the level of crud is monstrous, you won't be able to stand this for more than a few seconds ... this, boys and girls, is ...  wait for it ... gasp!!, distortion . Solve the distortion issues by tweaking ... the crud disappears ... simples ...

[pkane2001]

2 hours ago, fas42 said:

Hearing distortion? So simple - get any rig, put on a 'difficult' recording, for that setup; wind up the volume, and then put your ear near the treble/midrange drivers of one speaker - the level of crud is monstrous, you won't be able to stand this for more than a few seconds ... this, boys and girls, is ...  wait for it ... gasp!!, distortion . Solve the distortion issues by tweaking ... the crud disappears ... simples ...

 

Curious, if I bypass the binding posts and solder speaker wire directly to the speaker, will the "monstrous crud and distortion" be ameliorated? I distinctly recall someone making that suggestion as the first step in improving my otherwise excellent-sounding system ;)

 

[fas42]

20 hours ago, pkane2001 said:

 

Curious, if I bypass the binding posts and solder speaker wire directly to the speaker, will the "monstrous crud and distortion" be ameliorated? I distinctly recall someone making that suggestion as the first step in improving my otherwise excellent-sounding system ;)

 

 

And, as always, it's a depends ... .

 

As in, it depends on whether that is the most significant weakness in your system - yes, I would certainly do some tests to verify the integrity of that link. In spite of my active speakers being ridiculously cheap, the level of quality is quite remarkable, such as beautiful soldering inside the box, of the leads to the drivers; the umbilical cable between the two cabinets has good quality connectors, which will have much better contact integrity characteristics than the typical binding posts - plus, I have added damping around the connection area to reduce the impact of vibration. So far, this link has not proved to be as significant as other areas ... .

 

I get "grubby sound" from my setup, currently, if jitter in the optical link is too high - so, work is in progress to sort that ... you know, you could call it, step by step surgery, .

[Jud]

21 hours ago, pkane2001 said:

 

Curious, if I bypass the binding posts and solder speaker wire directly to the speaker, will the "monstrous crud and distortion" be ameliorated? I distinctly recall someone making that suggestion as the first step in improving my otherwise excellent-sounding system ;)

 

 

This is particularly interesting since solder is not fabulously conductive - it's used because it makes contacts last, not because it makes them electrically better.

[fas42]

50 minutes ago, Jud said:

 

This is particularly interesting since solder is not fabulously conductive - it's used because it makes contacts last, not because it makes them electrically better.

 

Yes. What solder does is improve long term integrity; not the absolute qualities of a link, at the time that it's created. Best connection would be a single strand of silver, followed by copper as material for the same dimensions, between the endpoints within the separate components. Which won't happen, because it's impractical.

 

Cold welding would be the next best method, for a join - crimping done with no flaws; which is not trivial. Again, impractical. Good physical contact, with solder stabilising the join - stopping movement, making the join gas tight - is a "good enough" solution.

 

Best audio, IME, comes about from improving the worst links; not making by some links brilliant, and ignoring "the devil's in the details" considerations ...

[pkane2001]

12 hours ago, Jud said:

 

This is particularly interesting since solder is not fabulously conductive - it's used because it makes contacts last, not because it makes them electrically better.

 

That's the problem with imagination, Jud :) Just because there's some tiny possibility of a miniscule conduction difference that has an insignificant effect on the large voltages and currents fed into a passive speaker, one can't automatically assume that the result will be "grubby sound" or "monstrous distortion" levels, like our friend Frank here does on a regular basis :) 

 

 

[fas42]

Nothing like a bit of Straw Man - so easily makes the uncomfortable stuff go away ... .

 

The problem isn't conduction - it's noise. In part caused by non-linear resistances, and other material behaviours that allow unwanted electrical activity to occur. Once one has a grasp on the mechanisms at play, it's quite easy to adjust some area of a system to increase the chance of such noise - and then reverse that change. The hard bit is when you can hear the damage being done to the SQ, and you then try to track down, or guess, the origins of the interference - so much of this is caused by a lack of understanding, or just laziness, of audio gear manufacturers. A good example is the latest bit of kit bought by audio friend up the road, a retro piece by Ruark: full of modern circuitry ... but it still sounded like something from 50 years ago, some of the time. Savage internal "sorting out" of all the bits of silliness, in the implementation, has lifted it considerably out of the doldrums - now capable, of Big Sound ...

 

Unfortunately, "bits of silliness" exist everywhere, in audio setups - just believing that they can't possibly exist won't make them go away ... you either have to buy gear which was designed by people with their heads screwed on - the D&D 8C speakers are a very good recent example - or fix the issues, somehow.

[botrytis]

Lp

1 hour ago, fas42 said:

Nothing like a bit of Straw Man - so easily makes the uncomfortable stuff go away ... .

 

The problem isn't conduction - it's noise. In part caused by non-linear resistances, and other material behaviours that allow unwanted electrical activity to occur. Once one has a grasp on the mechanisms at play, it's quite easy to adjust some area of a system to increase the chance of such noise - and then reverse that change. The hard bit is when you can hear the damage being done to the SQ, and you then try to track down, or guess, the origins of the interference - so much of this is caused by a lack of understanding, or just laziness, of audio gear manufacturers. A good example is the latest bit of kit bought by audio friend up the road, a retro piece by Ruark: full of modern circuitry ... but it still sounded like something from 50 years ago, some of the time. Savage internal "sorting out" of all the bits of silliness, in the implementation, has lifted it considerably out of the doldrums - now capable, of Big Sound ...

 

Unfortunately, "bits of silliness" exist everywhere, in audio setups - just believing that they can't possibly exist won't make them go away ... you either have to buy gear which was designed by people with their heads screwed on - the D&D 8C speakers are a very good recent example - or fix the issues, somehow.

What utter nonsense.

[fas42]

At times I just have to laugh at the 'nonsense thinking' that audiophiles have trapped themselves in, ... just came back from being out, and switched on my actives, after days of not being used- and am right now using this to warm them up,

 

A 1951, mono, live recording - audience noises are OTT, feet shuffling on the stage, props moving, constant rumbling; would be rejected out of hand these days - after about 10 minutes to settle down, magic ... I'm just gonna have to listen to the end, at least, of the first CD ...

 

Right now, listening from the other end of the house - to the faraway concert stage, down where the system is playing ... you see, that's what getting a setup optimised gives you - an illusion, which is special every time. And that quality is so, so easily lost - the slightest misstep can take it all away.