Do we crave distortion?

[kevinh]

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.

 

[lashto]

> So .. you just need to prove that those graphs you are showing do actually sound 'bad'

Do I?

Of course you do. Unless you have a DBT, lines on a graph are just ... lines on a graph. May sound better, may sound worse ... or may not sound at all.

All I need to show is that harmonics produced by musical instruments are different than IMD produced by electronics.

Very good, we know they look different. And who says they sound bad/worse?
Last I checked, studios use lots of electronics. Their suposedly "badly distorted by electronics" product sells for billions every year.

Whether you like the extra distortion or not is not my concern, although I've now built two separate tools for you to listen and evaluate this for yourself.

I did. I like your tools. Also like some of the 'distortion' they produce.

> IIRC, an assertion like "THD never comes alone, more THD = more IMD" was posted as a fact/argument in this thread. Numbers show it to be quite wrong: some kinds of THD add loads of IMD, some add very little (maybe close to none!?).

Your discussion of THD vs IMD is very wrong, as has already been pointed out. Stop arguing a point that has no basis in reality. THD does not cause IMD, THD is IMD, just tested with a single tone. THD and IMD are not something independent -- they ARE THE SAME THING, caused by the same non-linearity. To continue to argue that somehow they are separate and distinct is simply ignorant.

yes, saying "THD adds IMD" is technically wrong. I just wanted to emphasize "more THD does not automatically mean more IMD" (as in diff electronics = diff TFs = diff proportions of THD/IMD).
Anyway, I was wrong. Me being wrong about TFs/etc does not prove you right, though. Would be much more efficient/helpful if you brought arguments for the theory that "electronic distortion always sounds bad".

 

[NTK]

Anyway, I was wrong. Me being wrong about TFs/etc does not prove you right, though. Would be much more efficient/helpful if you brought arguments for the theory that "electronic distortion always sounds bad".

So now we are onto the "shooting blind does not always miss the target" argument?

 

[antcollinet]

Kind of agree. However, what if the original photograph/recording is not particularly high quality?
If you wanted to reproduce it to put on your wall, is it acceptable to tweak it slightly in Photoshop first?
To me, this is the equivalent of tweaking the sound profile in playback.

Yes - that is a creative process. You are modifying the photograph in a controlled/designed way to make it more visually appealing. And you are applying changes to that specific image that suit that specific image. You are not applying a global change that will be applied to every photo you ever look at.

But yes - if you want to apply tone controls or EQ to the music - not a problem, have at it. It is a controlled process using tools designed to enable you to do just that.

This is different from the distortion applied by reproduction gear which is the opposite - it is the result of weaknesses in the design of the gear (non linearity) causing non designed and unwanted changes to the sound.

In your photographic analogy it is the equivalent to viewing your carefully tweaked image in a frame where some child has smeared ice cream all over the glass.

 

[lashto]

At the production stage there is someone who makes a decision: the production is ready, the product will be deliviered as it is.
Thus the product arrives at the reproduction stage.
Now if the product is ok it should be used 1:1 without any further "sounding".

Pretty big "if" in there. And studies show that many people prefer the "further sounding".
"Should be used 1:1" is not some law or rule, just a preference ... neither better nor worse

BUT:
the decision of the producer is based on what he hears. He is using his own equipment and environment.

and his preferences! Basically, he does whatever sounds better to him.
When I add 'distiortion' at home (e.g. with tubes or whatever) I do the exact same.

At the reproduction side the equipment is different and may even sound different despite of the theoretical 1:1 transfer.

I'd say it'll surely sound different: different speakers, different room ...

So it is possible that the listener likes to have some changes, a bit here and a bit there. This could be equalizing or also adding distortion.
We may call it sounding. The problem with sounding: usually the listener has less control, the sounding is changing the complete track.

I do not hear/see "the problem with sounding". And even on engineering-heavy-ASR, people voted "signal is not holy"

An electric guitar can get heavy distortions at the production side, a voice is kept untreated.

Beg to seriously differ here.
So called tube-microphones are most popular, particularly for voices. That's anything but "untreated voices".
Also pretty sure that voice tarcks also go through exciters/reverb/effects like all else. The kind of electronics that suposedly make those tracks sound 'bad'. Many thousands of studio engineers say the exact opposite .. and billions of customers seem to agree

A very few 'purists' do stuff like record-direct-to-lp or direct-to-dsd. But even those records contain the (electronic) mic 'distortion'.

Applying the same distortion at the reproduction side will also influence the voice.

And that's probably the most liked quality of tube amps: the way they "influence the voice"

The discussion on how much of the distortion is acceptable is just a dicussion about how much is perceived or masked.

What is "acceptable" is an interesting subject too. A bit far though since we already jumped to preferable/craved

 

[pkane]

Of course you do. Unless you have a DBT, lines on a graph are just ... lines on a graph. May sound better, may sound worse ... or may not sound at all.

Sorry, but I've made no claims about audibility or preference, you did. So you do the testing and provide evidence.

Very good, we know they look different. And who says they sound bad/worse?

You did.

I did. I like your tools. Also like some of the 'distortion' they produce.

Glad you liked it. Yet, what you heard is IMD, not HD. But I think we're beating a dead horse, by now...

 

[lashto]

Well... 2nd harmonic at -60dB would be masked.

or not. And harmonics do not seem to care that much about masking

... when a noise was added that would have masked these distortions ... listeners still heard a pitch corresponding to the missing fundamental, as reported by J. C. R. Licklider in 1954.[4]

Ofcourse the IM products that always accompany harmonic distortion, may not be masked and will not contribute to pleasantness but rather smearing.
Regardless of how much you want HD to be the only 'product' it isn't.

Did not say that I "want HD to be the only product". My ears are also "IMD factories" and I like them

So what we can conclude is that you say it should not be too much distortion but it is not known when it is too much and it also is not clear what the minimal amount of distortion is.

Do you know of any research that shows what the minimum is ?
I mean ... no distortion is said to sound 'lifeless' so there must be some minimum (subjectively determined) where sound will 'improve' and as above a certain point it must become less 'nice' again.
There thus must be a sweet spot and that should be quite provable with listening tests.

some things are clear about HD (e.g. we crave it). Most are not and one can easily reach maybeland.

Talking about audible HD, studies put it at -60db. But that is usually the "audible as distortion" limit. Linkwitz says that "audible as improvement" starts much earlier. Don't know where, did not find a published/academic study.

Here's what Katz says

The added distortion should subjectively improve sound quality without degrading clarity, depth, or definition. If not, then the device must be adding too much distortion. Based on these listening tests, I recommend a maximum level of second harmonic no higher than 60 dB below the fundamental at nominal level (mezzo forte), typically -20 dBFS. The sound of -60 dB second harmonic is pleasant to the ear and enjoyable to many listeners, but perceived as a bit thick sounding with some loss of detail. Probably -66 dB second harmonic is the ideal setting, as in our listening it produces an attractive, warm, three dimensional quality with no apparent loss of detail or other side effects on any amplifier we tried. With some musical sources and some listeners, we may have to further lower second harmonic. I also suggest that the processor and amplifier be otherwise very linear..

As about the upper limit, the SINAD level where things start to sound 'lifeless', I also have no study.
My oppinion: I think lifeless/sterile sound does not depend on the HD level but (mostly) on the HD-spectrum. At least when played through 40-SINAD speakers. If we had 120-SINAD speakers, that opinion may be very different.

 

[UliBru]

Pretty big "if" in there. And studies show that many people prefer the "further sounding".
"Should be used 1:1" is not some law or rule, just a preference ... neither better nor worse

and his preferences! Basically, he does whatever sounds better to him.
When I add 'distiortion' at home (e.g. with tubes or whatever) I do the exact same.

I'd say it'll surely sound different: different speakers, different room ...

I do not hear/see "the problem with sounding". And even on engineering-heavy-ASR, people voted "signal is not holy"

Beg to seriously differ here.
So called tube-microphones are most popular, particularly for voices. That's anything but "untreated voices".
Also pretty sure that voice tarcks also go through exciters/reverb/effects like all else. The kind of electronics that suposedly make those tracks sound 'bad'. Many thousands of studio engineers say the exact opposite .. and billions of customers seem to agree

A very few 'purists' do stuff like record-direct-to-lp or direct-to-dsd. But even those records contain the (electronic) mic 'distortion'.

And that's probably the most liked quality of tube amps: the way they "influence the voice"

What is "acceptable" is an interesting subject too. A bit far though since we already jumped to preferable/craved

I'm not an artist or producer but a consumer. So I like to hear what others create and try to tell by their art. I prefer to hear it as original as possible.
Of course I also appreciate a voice recorded by a proper microphone and treated with special effects. And I can even enjoy distortions.
So a recording does not need to be free of distortions.

But I do not want to create my own post-production and to put the same spices over all instruments and voices.
So e.g. adding some reverberation to a voice or instrument at the production stage can sound pleasing. Adding the same reverberation over the complete track, the complete album or music library will not create goose bumps by highest chance. Instead of reverberation you can also read vibrato, phasing, echo and other effects AND distortion.

You are free of course to do what you like and enjoy it.

 

[Keith_W]

Ummmmm. I would not be arguing with Uli or pkane if I were you @lashto . I am in awe of their presence on this thread.

 

[lashto]

I'm not an artist or producer but a consumer. So I like to hear what others create and try to tell by their art. I prefer to hear it as original as possible.
Of course I also appreciate a voice recorded by a proper microphone and treated with special effects. And I can even enjoy distortions.
So a recording does not need to be free of distortions.

But I do not want to create my own post-production and to put the same spices over all instruments and voices.
So e.g. adding some reverberation to a voice or instrument at the production stage can sound pleasing. Adding the same reverberation over the complete track, the complete album or music library will not create goose bumps by highest chance. Instead of reverberation you can also read vibrato, phasing, echo and other effects AND distortion.

You are free of course to do what you like and enjoy it.

Each on of us has preferences, that none of my (or anyone's) business.

I do get a bit of a headache though when the preference for fidelity is presented as the must-have/better/etc preference. And people who like some distortion spice are called "sucker likes distortion".
Not nice. Not fair. Not true. Not helpful. Not .. anything.

Also a few here presented interesting statements like "distortion always sounds bad" or "electronic distortion always sounds bad" (mostly as in "the D you add at home is bad/worse/ugly/etc").
Still waiting for a single trace of proof/study/etc ... and still failing to compute how my tube amp is 'bad' but the music studio's tube mic/effect/amp is 'good'

 

[lashto]

Ummmmm. I would not be arguing with Uli or pkane if I were you @lashto . I am in awe of their presence on this thread.

I like & respect Paul's work a lot.
I like & respect Uli's work a lot.
Both have a lot more EE experience than me. As in, I have ~none

That doesn't mean I cannot argue. Logic is universal, math is too and I studied both. Also, preferences are not an EE-question and anyone can read & quote studies.
And of course, anyone can be wrong. If I am that one, I'd like to see the study/math that says so. Still waiting...

 

[ahofer]

But I do not want to create my own post-production and to put the same spices over all instruments and voices.

..and recordings. I really don't get the idea behind a fixed/unswitchable distortion, EQ, or other effect imparted by colored equipment.

 

[kevinh]

I'm not an artist or producer but a consumer. So I like to hear what others create and try to tell by their art. I prefer to hear it as original as possible.
Of course I also appreciate a voice recorded by a proper microphone and treated with special effects. And I can even enjoy distortions.
So a recording does not need to be free of distortions.

But I do not want to create my own post-production and to put the same spices over all instruments and voices.
So e.g. adding some reverberation to a voice or instrument at the production stage can sound pleasing. Adding the same reverberation over the complete track, the complete album or music library will not create goose bumps by highest chance. Instead of reverberation you can also read vibrato, phasing, echo and other effects AND distortion.

You are free of course to do what you like and enjoy it.

Of course the issue ofr me is we don't really know what the original sounded like (I am thinking of older recording which I listen a lot to). In that case we need a tone control like the old Cello Audio Pallette rendered in software at a good prince point (ie <$500).

 

[solderdude]

harmonics do not seem to care that much about masking

You do not seem to understand masking ..

At least when played through 40-SINAD speakers

What is a 40 SINAD speaker ?

Did you read what Katz stated:

I also suggest that the processor and amplifier be otherwise very linear..

So no high output Z, no roll-off or phase shifts and distortion below 66dB. (0.05%)

Do yourself a favor. Take your favorite music. Play it loud. Now attenuate -66dB. What do you hear ?
All harmonics and IM products should be below that level.

Do yourself another favor.
Take 2 different recordings.
Attenuate 1 of them by -66dB.
Mix one of the recordings with the other attenuated one.
Do an AB test with the original recording and the one with the -66dB recording mixed in there.
No cheating with quite passages in the loud recording.

 

[DonH56]

"Do we crave distortion?"

Only hearing in high-end audio, naturally. In vision, smell, touch, taste, bridge construction, most anything else it would be a bad thing.

 

[audiofooled]

Raising or lowering the playback level is possibly the only fully transparent change (thinking about loudness curves a different level will also change the timbre)..

Fully agree.

BUT:
the decision of the producer is based on what he hears. He is using his own equipment and environment.

And seldom evaluates the mix at low to moderate listening levels, not to disturb the neighbors?

At the reproduction side the equipment is different and may even sound different despite of the theoretical 1:1 transfer.

Yes, I think there is something to it that systems which are neutral and transparent, tend to sound less exciting when they are basically idling or not being used to the slightest of their potential, or at least approaching some threshold of listening level with regards to what is intended during production stage.

For example, my system is far better than my hearing when it comes to distortion and noise and I see no point in adding any coloration to it. But when it comes to recordings, there are simply ones that are made for low listening levels, and some that are very dynamic and beg for a gradual level increase until they sound just about right, and there are extremely dynamic ones that sound better the louder they are reproduced. I think this is, amongst other things, due to a range of psychoacoustic phenomena, one being that low distortion system never seams to be loud enough.

As an extreme example there are two performances by the same artist that I equally like, but are performed very differently and in completely different environments:

This is obviously recorded the way it was and there is nothing more to expect. Everything seems to be "wonderfully smeared" by all the reflections and I like to listen to this at low to moderate levels, casually, with no attention to detail, just enjoy the performance. Try to raise the level a bit loud and the SQ degrades.

Here, the artist is in a recording studio, he puts on a different hat, as if knowing his performance here would be unrestricted, also there is an engineer at the console who is not hiding how much he loves his job. This is quite something. Crank this up if your system is capable of these dynamic peaks, and enjoy all the transients, harmonics, also distortion if you will... Even the first few seconds will leave very little to be desired. But, at low listening levels, you don't even scratch the surface and the performance may seem to offer nothing more than the previous one.

To me, this begs the question, is there or is there not, a misunderstanding that coloration or added distortion on the reproduction side, is a form of poor compensation for lack of dynamic performance?

 

[lashto]

the talk is quite "circular" again so I'll pick up where I left some open points.

None of the people-prefer-distortion studies go as far as "distortion is always good for everyone, everywhere". Caveats galore as in "sometimes only", "some people only", "some distortions only" (and unclear which)...

But there's one kind of distortion that everyone likes. Always!
And it is not just good-ish but the absolute maximum level of "the best sound I ever heard".
It's so magically good, that the most hardcore engineer will hear it and say "I love it" .. guaranteed

And that is the 'distortion' of your own years. To note: it is also highly non linear distortion (i.e. like your amp's D) and the math/graphs/conclusions about HD/IMD from pkane's posts above apply here too.

The graph matches the (more recent) findings that D is a lot less audible at lower frequencies. The experimantally found D-audibility numbers are an ~perfect fit (see the studies quoted under "Perception" for "Just Detectable Distortion")

And a few more for extra fun

And a truckload of IMD from the same ears

That should be more than enough 'distortion' for today ... comments/conclusions highly welcome

 

[Sined]

if we reduce the question to the most basic/simple form, the answer is a giant and resounding
YES, WE DO CRAVE DISTORTION.

Very simple proof: just listen to the musical note C as pure sinewave compared to the piano note C.
"Surprise": ~everyone prefers the piano note and hates the sinewave.
One of the major extra-ingredients that make the piano note sound more enjoyable is, suprise again, Harmonic Distortion: the piano note contains 10+ HDs which contribute to its pleasant/musical timbre (and make it sound different from the same note/frequency played on guitar, trumpet etc...). A nice video explanation.

Harmonics generated by an amplifier are considered as distorsion in world of audio since something is added that was not on the source. But in the world of music, harmonics added to the original makes it sound richer, especially the 2nd harmonics: it is like adding other performers that plays the same instruments and the same notes but one octave higher: our ears like it !

 

[lashto]

"Do we crave distortion?"

Only hearing in high-end audio, naturally. In vision, smell, touch, taste, bridge construction, most anything else it would be a bad thing.

D in bridge construction, yep quite bad. And in anything that needs precision engineering.

But with everything else you mention, I can only see the opposite:

• Vision: Van-gogh's heavily 'distorted' sunflowers are about $20 millions more prefered than any 'pure' ones.
• Smell: many people prefer to 'distort' their body-smell with perfume.
• Taste: people prefer the taste of eggs 'distorted' with lots of heat and salt. Or maybe you like to taste them "how the chicken intended" !? .
• Touch: sorry, cannot think of an example on spot...

And as a welcome-to-the-thread present: another one of those more-thd-was-prefered studies

 

[DonH56]

D in bridge construction, yep quite bad. And in anything that needs precision engineering.

But with everything else you mention, I can only see the opposite:

• Vision: Van-gogh's heavily 'distorted' sunflowers are about $20 millions more prefered than any 'pure' ones.
• Smell: many people prefer to 'distort' their body-smell with perfume.
• Taste: people prefer the taste of eggs 'distorted' with lots of heat and salt. Or maybe you like to taste them "how the chicken intended" !? .
• Touch: sorry, cannot think of an example on spot...

And as a welcome-to-the-thread present: another one of those more-thd-was-prefered studies
View attachment 335552

Distortion implies no longer true to the source material. Your examples are of modifying the source, not necessarily adding distortion from our own senses. And I'd like to see the reference for the study that you linked; most (not all) of the ones I am familiar with have been debunked over the years. In any event it does not violate my premise that distortion in audio is preferred by some listeners, albeit work by Olive and Toole show the opposite. It's usually pretty easy to find studies on both sides of any given issue, though I have no desire to go digging for them. Or to read this thread more than a post or two every other week or so.