Pushing the Computer Audio Boundaries :)

[jkeny]

Like the hugely expensive Wavac amp reviewed in Stereophile: subjectively sounded gorgeous, objectively measured broken! Sometimes that distortion sounds pretty damn tasty.

well that tells us that the human ear/brain is very tolerent of certain levels of distortion or that we cannot differentiate between something that has v.good or bad measureable distortion. If the opposite were true then we would not like the sound of valves or vinyl which objectively measures poorly.

Where am I going with this? I've no idea. :-) Told you it would go around in circles.

Maybe is this too simplistic but as per the review, the high distortion is mostly pure 2nd harmonic, to which the human ear is very tolerant. High order distortion is more discordant sounding. Inter-modulation distortion (IMD) is perhaps more discordant again. I would also suggest that the ear is infinitely more sensitive to timing than to frequency & if this isn't measured then we are probably leaving out the most critical factor in sound quality. Unless we put a weighting on each of these types of measurements (based on their importance to our hearing) then we are going around in circles. But it should be possible to approach this in a scientific way with psychoacoustics & current models of the ear.

[Diapason]

Makes sense to me, John. Told you you'd get that Nobel prize!

[jkeny]

Makes sense to me, John. Told you you'd get that Nobel prize!

Could you pick it up for me - I don't have the time?

[DaveF]

Unless we put a weighting on each of these types of measurements (based on their importance to our hearing) then we are going around in circles. But it should be possible to approach this in a scientific way with psychoacoustics & current models of the ear.

I wonder though if it is indeed possible to come up with such a model for the ear. Would the attempt be just too complex? Are there too many variations in hearing from one person to the next or would a general model be enough?

What contributes most in our ability to hear and hear something consistantly? Is it the physical make up of the ear and its health and how much of ones brain is involved in this?

Let the sh*te talk begin. :-)

[Diapason]

Here's another question, John: if second-order distortion sounds good, can you introduce some to your DAC? Just so we're clear, I'm not joking!

(If you could introduce it in a controllable way, the listening tests could be very interesting.)

[Ciaran]

Here's another question, John: if second-order distortion sounds good, can you introduce some to your DAC? Just so we're clear, I'm not joking!

(If you could introduce it in a controllable way, the listening tests could be very interesting.)

I remember reading about an experiment of that kind in Stereophile some years ago. The conclusion was that while we might be more tolerant of some sorts of distortion than orders, nobody seemed to prefer any of the artificially distorted tracks to the original unaltered one, even if the distortion was the cuddliest kind you could think of!

I'll see if I can find a reference.

[Diapason]

Curiouser and curiouser.

[Ciaran]

Here's the reference: it's "Euphonic Distortion: Naughty but Nice?" by Keith Howard from 30 Apr, 2006.

For the TL/DR crowd, the main conclusion (Page 2) is

The most important finding was that none of the different patterns of nonlinearity sounded in any way preferable to the undistorted reference. They all sounded worse, albeit in different ways. Pattern 1 added a distinct "dirtiness" to the sound that was not unpleasant but did change the instrumental timbre and diminish the sound's sense of fidelity—there was something clouding the sound. Pattern 2 was much better, with the closest sound to the undistorted reference. But I thought I could also detect it just beginning to muddy the presentation. Pattern 3 was unpleasant, adding an edge to the sound that would surely become fatiguing over extended listening. Pattern 4 wasn't as bad, but there was still something unnatural about it. Although Patterns 3 and 4 both introduce less distortion than Pattern 1, it was Pattern 1 that proved less subjectively annoying. While I could hear its effect, it did not threaten to send me screaming from the room after 10 minutes' listening.

The article includes a link to Howard's website, where you can download a program to add distortion of your choosing to any soundfile. I don't think I've seen any references to the article since then, but it does seem to undercut a piece of popular audio "wisdom", viz. that distortion added to sound can make the sound pleasanter if it's the right kind of distortion. It does confirm the idea that the type of distortion is important, in that smaller amounts of the wrong kind will sound more unpleasant than larger amounts of the right kind.

It's also interesting that

the first surprise awaiting you when you experiment with distortion synthesis is just how large the distortion must be to become audible.

[Diapason]

Right then, another piece of my audio "knowledge" has been seriously upended. It's pretty stark too:
"for me the issue is now settled. Unless and until somebody comes up with a "magic" pattern of nonlinearity that truly enhances sound quality, I will believe euphonic distortion to be a fantasy."

[jkeny]

Unless we put a weighting on each of these types of measurements (based on their importance to our hearing) then we are going around in circles. But it should be possible to approach this in a scientific way with psychoacoustics & current models of the ear.

I wonder though if it is indeed possible to come up with such a model for the ear. Would the attempt be just too complex? Are there too many variations in hearing from one person to the next or would a general model be enough?

What contributes most in our ability to hear and hear something consistantly? Is it the physical make up of the ear and its health and how much of ones brain is involved in this?

Let the sh*te talk begin. :-)

Let me start the sh*te talk !!

Let's separate the ear from the brain (for the moment). The ear appears to work, not like a microphone but more like a variably tuned receiver which can & does alter it's sensitivity depending on the circumstances. Remember, the ear was never designed for listening to music, it's role in survival has to be understood. Primarily then, it is an early warning device of any anomalous sound in our environment to sense danger i.e the approach of an attacking animal, for instance. The direction from where this sound is coming has to be quickly determined & that's why we are probably exquisitively sensitive to timing information because it's the minute difference in timing of the sound wave hitting our ear that gives us the location of a sound - and this can be shown to be very accurate. As an aside there is an interesting Youtube video of a blind cyclist who echo locates around by making clicking sounds http://www.youtube.com/watch?v=Z_E3zxx2l9g
So anyway, this is the threat correlation function of the ear. It may also explain some issues that people have with some digital - that uncomfortable, uneasy feeling where long term listening to some digital is impossible. Digital audio, in my opinion, made a fundamental mistake in focusing, almost exclusively, on frequency accuracy & ignoring perhaps the more important timing element of signals. It's only now beginning to rectify this mistake by paying more attention to it.

Dave, I'm not sure why you appropriate intractable complexity to the functioning of the ear? As in all models of operation, the model describes the ideal & deviations from that are seen as individual exceptions or diseases. I don't think that hearing is so fundamentally different from person to person that a good working model can't be developed.

Once the ear model has been developed the era/brain interaction can be added to this as a meta-model