Do we crave distortion?

[pkane]

Thanks, I have missed to read this post. In Germany we say something like: two fools one thought

Haha! In English, the saying is a bit nicer to the two fools, and suggests that they have two great minds, but I like the German version better

‘Two great minds think a like’

 

[lashto]

The thing is that the electronics will see each harmonic as a fundamental,they not only they will add H2 to H2 but also H2 to H3 and so on.

agreed

I don't thing we want that.

why ?!
Your ear does exactly the same .. and you like it .. or even love it ..

 

[Sokel]

why ?!
Your ear does exactly the same .. and you like it .. or even love it ..

Because the actual instrument as I hear it live at my house is perfect by it's own,doesn't need any contribution.

I wish I had all the capability to reproduce it as is,as I wish the studio people had the same capability to deliver it to me exactly that way (dynamics,etc can be insane)

We settle for the lesser evil where we can but I'll never deliberately add more distortion,the opposite.

 

[lashto]

An alternative theory, for which I have no proof (nor care to find any), is that our brain compensates the ear's transfer function and adding additional distortion is less pleasing.

some extra thinking out loud ...
"our brain compensates the ear's transfer function": looks like you are introducing a hitherto unknown function of the brain. Anyone has proof of its existence? Otherwise, the hypothesis falls off the occam-scale.

The brain does not do extra/useless work. Most of what it does, provides some sort of evolutionary advantage. What would be the evo advantage of cancelling ear's TF?
None I can think of. Blame my lack of imagination

Another question: how does the brain know the ear's TF? Or how does the brain know that the ear changes the signal in any way? That would imply the brain having another mechanism to 'sense' the original signal, compare it with the ear's signal and say "gotcha, you tricky Ear"

The more I think about this "alternative", the less it sounds right...

 

[lashto]

Here my alternative to your alternative @DonH56 ...

I presume that a signal with ear-like-HD is interpreted by the brain as "pleasant harmonic tickling".
More ear-like-HD = more tickling = more pleasure.
Same as tickling, the range of "more" is quite small and it can easily cross into unpleasant.
Same as tickling, the perception is not the same for all people: some like it, some don't care, some hate it.
The engineers are known as busy, serious people with no time for "lay stuff" like tickling. It follows that they don't like tube-hd ... 'obviously'

Joking aside, my 'explanation' scores well on the occam-scale and makes (some) sense.

Spoiler: just more 'joking'

Assuming the above is true, here's a future dialog in an audio store:
Salesman: hello mr. amp buyer, may I ask an unusual question?
Amp Buyer: yes sure
Salesman: do you enjoy tickling?
Buyer (baffled): no, not really
Salesman (for himself): dang, cannot sell that expensive tube-amp with a juicy comission
Salesman (for the client): ok mister, let's find a nice, linear classD amp for you.
Buyer: wow mister, you read my mind

 

[lashto]

Thanks, I have missed to read this post. In Germany we say something like: two fools one thought

Haha! In English, the saying is a bit nicer to the two fools, and suggests that they have two great minds, but I like the German version better

‘Two great minds think a like’

good to see some love in this thread.
also +1 for the German version (especially since I am trying to improve my German)
And since I already answered pkane's graphs, that's all folks

 

[audiofooled]

Another question: how does the brain know the ear's TF? Or how does the brain know that the ear changes the signal in any way? That would imply the brain having another mechanism to 'sense' the original signal, compare it with the ear's signal and say "gotcha, you tricky Ear"

The shape of your pinnae is modifying the spectrum, giving you spectral cues of the sound elevation localization. Unlike dogs, humans do not require tilting their head in order to "cheat" by trying to resolve height information by creating vertical timing cues. Brain is calibrated to the shape of your pinnae and if you were to modify it by using some clay, your sound localization would be pretty much messed up. This is why binaural recordings don't work to some people, they are made for generic pinnae shape and for spectral cues, they are a hit and miss regarding your own "calibration file".

 

[lashto]

The shape of your pinnae is modifying the spectrum, giving you spectral cues of the sound elevation localization. Unlike dogs, humans do not require tilting their head in order to "cheat" by trying to resolve height information by creating vertical timing cues. Brain is calibrated to the shape of your pinnae and if you were to modify it by using some clay, your sound localization would be pretty much messed up. This is why binaural recordings don't work to some people, they are made for generic pinnae shape and for spectral cues, they are a hit and miss regarding your own "calibration file".

very nice ear-input, thank you !

"Brain is calibrated to the shape of your pinnae"
would that be the same as saying "the pinnae's shape calibrates your brain" !? i.e. the pinnae is somehow the 'active' part and the brain just 'reacts'.

also "change your pinnae and your brain is ~lost-in-space" ..
just for a moment or 'forever'? does re-calibration happen if you wear elf-ears for a year?

interesting stuff & I'm just thinking out loud. Not even sure what/why I am asking .. kinda trying to understand how spatial-perception may relate to harmonics/timbre-perception

 

[lashto]

Bell Labs did research in the 60's on what sounds better to humans. Now this was back when we didn't know about TIM even and they were transitioning from Tube "Line Amps" for the phone network. They testing indicated that Humans preferred 2nd harmonic with some 3rd harmonic at lower levels, with little/no higher order high harmonic distortion. I think this is why many have preferred tube amps. IIRC part of the effect is that it made the recording sound louder than the actual Level, 3rd harmonic made it sound more detailed. Saw this info 30+ years ago so just relating a recollection.

would be nice to have a study/link, pretty tired of googling..

The effects you describe do match other comments/studies/etc.
H2 is suposed to have an additive, glow-like effect: makes the fundamental louder, bigger, heavier, more dynamic, clearer .. on the negative side (i.e. too much of it): muddy, foggy, smeared, unclear.
H3 is suposed to be a bit of an opposite: makes the fundamental quieter and/or less dynamic, thinner, more sharp/precise/focused .. on the negative side: plasticky, hard, harsh

Both H2 & H3 are liked and the H4+ are progressively less liked. Particularly the H5+ odds are suposed to have all the bad qualities of H3 and more: metallic, piercing, painful

That's a lot of harmonic flowery speech ..
However, the progressively-less-liked part appears 'everywhere'. Matches well with masking, critical-bands, music-theory .. everything else.

 

[audiofooled]

very nice ear-input, thank you !

"Brain is calibrated to the shape of your pinnae"
would that be the same as saying "the pinnae's shape calibrates your brain" !? i.e. the pinnae is somehow the 'active' part and the brain just 'reacts'.

also "change your pinnae and your brain is ~lost-in-space" ..
just for a moment or 'forever'? does re-calibration happen if you wear elf-ears for a year?

interesting stuff & I'm just thinking out loud. Not even sure what/why I am asking .. kinda trying to understand how spatial-perception may relate to harmonics/timbre-perception

This video may answer these questions:

They say the brain would recalibrate on interpreting the input for the spectral ques, that is if you were wearing modified ears for a couple of weeks, but I don't know of any research that would confirm this. The area of hearing aids would perhaps dive into this...

 

[DonH56]

Here my alternative to your alternative @DonH56 ...

I presume that a signal with ear-like-HD is interpreted by the brain as "pleasant harmonic tickling".
More ear-like-HD = more tickling = more pleasure.
Same as tickling, the range of "more" is quite small and it can easily cross into unpleasant.
Same as tickling, the perception is not the same for all people: some like it, some don't care, some hate it.
The engineers are known as busy, serious people with no time for "lay stuff" like tickling. It follows that they don't like tube-hd ... 'obviously'

Joking aside, my 'explanation' scores well on the occam-scale and makes (some) sense.

Spoiler: just more 'joking'

Assuming the above is true, here's a future dialog in an audio store:
Salesman: hello mr. amp buyer, may I ask an unusual question?
Amp Buyer: yes sure
Salesman: do you enjoy tickling?
Buyer (baffled): no, not really
Salesman (for himself): dang, cannot sell that expensive tube-amp with a juicy comission
Salesman (for the client): ok mister, let's find a nice, linear classD amp for you.
Buyer: wow mister, you read my mind

I have never heard of "pleasant harmonic tickling". When I have created test files I have not liked added distortion but that is an insignificant sample of one. To me, if you add distortion, you are changing the source material from what they originally intended, so it is up to you to determine if you like the change or not. Your contention is that most people like the change, implying it would be better if the content producers simply added more distortion to the released recordings. Or audio equipment designers produced equipment with higher levels of distortion so as to sound better to most people.


I suppose that supports your implication of engineers being unpleasant snobs (my words). I have owned and liked a lot of tube equipment over the years, but never felt it was more accurate. Sometimes it sounded better, sometimes not. Whatever.

The studies I have are paper (printed) and buried. I took a quick look in my office but they are boxed elsewhere. No idea if they are on the web and I simply do not care; I have too many other things going on right now to spend a lot of time arguing on the internet. I am no longer an AES member. The most recent discussions I had indicate a low level of second-order distortion may be preferred by some listeners; higher orders generally not, but this is based upon a conversation with an engineer whom I respect and I do not have the references he cited at hand. Peter Aczel comes to mind; the rest were AES papers.

Psychoacoustics is not my field of study so I am working from memory of what those in the field have said in past conferences and papers. Since you have already decided your position, and I am just not all that interested in debating whether and how much added distortion humans may or may not like. My career has focused on eliminating, or at least reducing, distortion so I am quite biased against the concept that it is good. In audio or anything else. You say I am wrong, and so it may be, but I will leave that to you.

 

[UliBru]

some extra thinking out loud ...
"our brain compensates the ear's transfer function": looks like you are introducing a hitherto unknown function of the brain. Anyone has proof of its existence? Otherwise, the hypothesis falls off the occam-scale.

The brain does not do extra/useless work. Most of what it does, provides some sort of evolutionary advantage. What would be the evo advantage of cancelling ear's TF?
None I can think of. Blame my lack of imagination

Another question: how does the brain know the ear's TF? Or how does the brain know that the ear changes the signal in any way? That would imply the brain having another mechanism to 'sense' the original signal, compare it with the ear's signal and say "gotcha, you tricky Ear"

The more I think about this "alternative", the less it sounds right...

The brain gets information from outside. In our case the incoming soundwave is filtered by the transfer function of the ear. But the brain just recognizes the incoming signal and it learns from the beginning (birth) what the incoming signal will mean.

There is a simple example and you can test it easily by yourself. Forgive me that the example is about how we see and not how we hear.
You are just reading my post and if I ask you if you have also recognized your nose during reading you will tell me by high chance that you have not recognized it. Close one eye and you can easily see the according side of your nose. Close the opposite eye and you recognize the other side of your nose. So the brain is perfect to cancel the view of the nose when you look with both eyes. Now you can test this: put some white color at the tip of your nose and experience how your sight is changing. The color will disturbe your viewing. After a while you will notice that your brain has accomodated to the new situation and you will see as usual. Take the color away and again the brain is confused until it gets accomodated again.

I'm pretty sure that if you glue something into your auricles your sound perception will change. And after a while you will again accomodate to that and you can listen to music as before.

IIRC I have read about cochlear implants where the patient just hears noise and scratches in the beginning. It takes its time but then the patient can start to learn what all the noise is meaning and he can start to hear again. As implants never work as perfect as our ears the final result is not perfect as before the hearing loss. But anyway it helps. And it proves that our brain is learning. This includes to filter out something that is not important for the incoming information.

And so we should be used to ignore the 2nd harmonics of our ears.

 

[lashto]

I have never heard of "pleasant harmonic tickling".

me neither

When I have created test files I have not liked added distortion but that is an insignificant sample of one. To me, if you add distortion, you are changing the source material from what they originally intended, so it is up to you to determine if you like the change or not. Your contention is that most people like the change, implying it would be better if the content producers simply added more distortion to the released recordings.

Producers do it already, for a long time. That effect is quite proven I would say.

Or audio equipment designers produced equipment with higher levels of distortion so as to sound better to most people.

Nelson Pass does ~that. According to some of his blog entries, he specifically tunes for H2 and "moderate amounts of good distortion" and "minimal feedback".
Did read an Agostino interview a while ago where he explained how he tunes for pleasant H2/H3.
Many seem to do that kind of "tuning"

I suppose that supports your implication of engineers being unpleasant snobs (my words). I have owned and liked a lot of tube equipment over the years, but never felt it was more accurate. Sometimes it sounded better, sometimes not. Whatever.

The studies I have are paper (printed) and buried. I took a quick look in my office but they are boxed elsewhere. No idea if they are on the web and I simply do not care; I have too many other things going on right now to spend a lot of time arguing on the internet. I am no longer an AES member. The most recent discussions I had indicate a low level of second-order distortion may be preferred by some listeners; higher orders generally not, but this is based upon a conversation with an engineer whom I respect and I do not have the references he cited at hand. Peter Aczel comes to mind; the rest were AES papers.

Psychoacoustics is not my field of study so I am working from memory of what those in the field have said in past conferences and papers. Since you have already decided your position, and I am just not all that interested in debating whether and how much added distortion humans may or may not like. My career has focused on eliminating, or at least reducing, distortion so I am quite biased against the concept that it is good. In audio or anything else. You say I am wrong, and so it may be, but I will leave that to you.

I am not exactly decided ... or sure.
Just thinking that a HD-to-preferences score/metric should be doable "today"

anyway, quite busy too ... many thanks for the reply.

 

[lashto]

The brain gets information from outside. In our case the incoming soundwave is filtered by the transfer function of the ear. But the brain just recognizes the incoming signal and it learns from the beginning (birth) what the incoming signal will mean.

There is a simple example and you can test it easily by yourself. Forgive me that the example is about how we see and not how we hear.
You are just reading my post and if I ask you if you have also recognized your nose during reading you will tell me by high chance that you have not recognized it. Close one eye and you can easily see the according side of your nose. Close the opposite eye and you recognize the other side of your nose. So the brain is perfect to cancel the view of the nose when you look with both eyes. Now you can test this: put some white color at the tip of your nose and experience how your sight is changing. The color will disturbe your viewing. After a while you will notice that your brain has accomodated to the new situation and you will see as usual. Take the color away and again the brain is confused until it gets accomodated again.

I'm pretty sure that if you glue something into your auricles your sound perception will change. And after a while you will again accomodate to that and you can listen to music as before.

IIRC I have read about cochlear implants where the patient just hears noise and scratches in the beginning. It takes its time but then the patient can start to learn what all the noise is meaning and he can start to hear again. As implants never work as perfect as our ears the final result is not perfect as before the hearing loss. But anyway it helps. And it proves that our brain is learning. This includes to filter out something that is not important for the incoming information.

The explanations/experiments are clear enough, do not even need to do them to believe

And so we should be used to ignore the 2nd harmonics of our ears.

but .. why would the brain do that? What is the gain/advantage for that extra effort?
Guess the hypothesis can be tested if someone builds an auditive implant with zero distortion (pretty tough job already) and connects it directly to the brain/nerves, bypassing the whole ear. All possible in theory.

I asked a few times "do you think that your ears sound bad?" Some may have thought it was a funny Q. Or rhetorical. It wasn't.
Considerig the strong distortion-hate in the engineering department, here's a possible/probable future.
Some engineer will look at his ear's HD and he'll say: wow, truckloads of HD in there ... I can do better ... I can 'fix' this. And he'll build the distortion-free audio-implant above.
And then, he/others may figure out that what they hear is not anymore "what the artist intended" ... and another round of 'audiophile wars' with start

 

[lashto]

I'd say the road-ahead looks good enough for a harmonic-to-preference score:

1. figure out the ideal-HD-spectrum
2. compute a sort of "Harmonic Affinity Score" as the diff beween DUT-HD and the ideal-HD

No 1 should be all in the studies I posted. I seen a few attempts at defining it but IMO none has gone far enough, none has included all aspects. I also have some extra fringe ideas .. which may not interest anyone

No 2 should be an easier follow-up, at least for a first/naive formula.
And Geddes already has a formula. The NewHampshire study has one too. Both look a bit 'short' to me but who knows ... they can be tested now (using ASR's measurement data)

.. and I'll take a break here. Most probably for the rest of 2023.
Merry Christmas and Happy New Year everyone!

 

[Sokel]

I'd say the road-ahead looks good enough for a harmonic-to-preference score:

1. figure out the ideal-HD-spectrum
2. compute a sort of "Harmonic Affinity Score" as the diff beween DUT-HD and the ideal-HD

No 1 should be all in the studies I posted. I seen a few attempts at defining it but IMO none has gone far enough, none has included all aspects. I also have some extra fringe ideas .. which may not interest anyone

No 2 should be an easier follow-up, at least for a first/naive formula.
And Geddes already has a formula. The NewHampshire study has one too. Both look a bit 'short' to me but who knows ... they can be tested now (using ASR's measurement data)

.. and I'll take a break here. Most probably for the rest of 2023.
Merry Christmas and Happy New Year everyone!

Maybe you have missed my post,but do you prefer the live or the recorded-reproduced sound of the instrument?

(let's assume that someone plays it low and easy,to eliminate the dynamic thing)
Cause if you do,you'll be the first I know and then I will be able to understand that there are people who prefer crippled/distorted/whatever version of the real and that may need some further digging if there are others like you.

(and Happy New Year)

 

[kevinh]

would be nice to have a study/link, pretty tired of googling..

The effects you describe do match other comments/studies/etc.
H2 is suposed to have an additive, glow-like effect: makes the fundamental louder, bigger, heavier, more dynamic, clearer .. on the negative side (i.e. too much of it): muddy, foggy, smeared, unclear.
H3 is suposed to be a bit of an opposite: makes the fundamental quieter and/or less dynamic, thinner, more sharp/precise/focused .. on the negative side: plasticky, hard, harsh

Both H2 & H3 are liked and the H4+ are progressively less liked. Particularly the H5+ odds are suposed to have all the bad qualities of H3 and more: metallic, piercing, painful

That's a lot of harmonic flowery speech ..
However, the progressively-less-liked part appears 'everywhere'. Matches well with masking, critical-bands, music-theory .. everything else.

This was long before the Internet or digital storage of files. Became aware of it in the early 70's in some long lost audio book I had at the time. Only things I remember from 50 years ago were meeting my wife and some concerts (GD Winterland 73 and 74 especially).

 

[MattHooper]

When I'm listenning to a guy playing a Stradivarius, no, I'm not gonna add distorsion.
Of course not.
Not open for discussion on the subject.

Because the actual instrument as I hear it live at my house is perfect by it's own,doesn't need any contribution.

I wish I had all the capability to reproduce it as is,as I wish the studio people had the same capability to deliver it to me exactly that way (dynamics,etc can be insane)

We settle for the lesser evil where we can but I'll never deliberately add more distortion,the opposite.

I view most if not all reproduced sound as distorted to begin with. From the colorations of microphones on down, what comes out of a sound system, be it a sax, voice, violin, drums etc, all sound diminished - diminished in sonic/harmonic complexity/timbre, and in richness, presence, and other qualities.

I therefore have no qualms about adding anything that, to my ears, gives something back in the direction that I find "more natural" sounding.

So, take the sound of a violin. The assumption seems to be that if you take a "good" recording of a violin, adding distortion could only make it sound worse. But, again, to my ears the recording has already made it sound worse. When I listen to a REAL solo violin what strikes me first is how big and rich it sounds - even played in it's highest register, the sound is thicker, richer than the wiry, thin facsimile I hear from reproduced sound. And since that's one of the main impressions that sticks with me, if a distortion seems to "enrichen" the sound of a recording, it can move it more in the direction my brain takes as "more natural, more like the real thing." And that is what I seem to get using my old tube amplification. Whenever I compare more neutral, accurate solid state amplification to the tubes amps in my system, I hear a trade-off: the solid state sounds a teeny bit more clean and precise, but the tube amplification sounds more "relaxed, rich, full, slightly thicker." And since those are just the type of qualities I hear in real life violins, for me that recreates just a bit more of the gestalt of hearing the real thing. It's just enough to not take away from the recorded character of the instrument - it'll still be a strad or whatever, but the recording just sounds a teeny bit more in the direction of "real/natural" to my ears. It's subtle, but to me significant.

 

[Deleted member 21219]

This information may be pertinent to this thread.


I once heard an opinion:
1) The inner ear creates harmonic distortion.
2) The degree of distortion rises with greater SPL.
3) The monotonically-rising distortion of tube gear mimics this rise with level, creating an illusion in the brain of "dynamic" sound that would not otherwise exist.

I'm not saying that I believe this, and I'm not saying that I don't believe this. I don't understand it well enough to say.

However, here are scientific papers on the subject:

Intracochlear distortion products are broadly generated by outer hair cells but their contributions to otoacoustic emissions are spatially restricted - Scientific Reports

Detection of low-level sounds by the mammalian cochlea requires electromechanical feedback from outer hair cells (OHCs). This feedback arises due to the electromotile response of OHCs, which is driven by the modulation of their receptor potential caused by the stimulation of mechano-sensitive...

www.nature.com

The origin of mechanical harmonic distortion within the organ of Corti in living gerbil cochleae - Communications Biology

He and Ren measured sub-nanometer vibrations of the reticular lamina from the apical ends of the outer hair cells in living gerbil cochleae using a custom-built low-coherence interferometer. Their results indicate that mechanical harmonics are generated by the motile outer hair cells and can...

www.nature.com

https://journals.physiology.org/doi/full/10.1152/physrev.00029.2012

Distortion in those good vibrations

The features that make inner ear hair cells so sensitive to vibrations may also be responsible for the introduction of surprisingly large distortions.

www.ncbi.nlm.nih.gov

Wouldn't this mechanism would be important to the individual when SPL levels are low, but unimportant or even undesirable when SPL levels are high? That would coincide with the opinion that I heard, and would be consistent with the way that I noticed people react to recorded music.

Jim

 

[Tangband]

You are still mixing production with reproduction.

Production: effects - artistic process. If done well it sells.

Reproduction: effects - bad.

Besides if H2 is sooo great why don’t producers are adding even more of it and save us the trouble adding it randomly during reproduction. It would sell even better according to you.

Can you show some measurements to back up this statement ? Or is it all anecdotes ?