Hi, this question is for people who like to ramble a little
My doubt is that below a certain level (that i do not know) distortion is not a problem for the listeners.
Taken from Stereophile magazine this amp, hardly low in distortion, is producing spectacular sound (the reviewer says)
So my question is ... when distortion is low enough for you ?
When you design something do you put very strict requirements for distortion or you focus instead on other parameters and which ones ?
As an aside i think instead that power supply noise is always detrimental for sound.
Thanks a lot, gino
My doubt is that below a certain level (that i do not know) distortion is not a problem for the listeners.
Taken from Stereophile magazine this amp, hardly low in distortion, is producing spectacular sound (the reviewer says)
So my question is ... when distortion is low enough for you ?
When you design something do you put very strict requirements for distortion or you focus instead on other parameters and which ones ?
As an aside i think instead that power supply noise is always detrimental for sound.
Thanks a lot, gino
Last edited:
Well that's a nice can of worms you're opening there 
But I find your statement on psu noise too general. I could also say the psu noise has NO impact on sound because it is on the supply not on the amp output.
But we know that some of it will sometimes end up at the amp output. How much, depends on the psu, the amp and probably a few more things like wiring and construction.
So is psu noise detrimental to the sound? It depends, on many things, so this can't be answered with a general statement.
Jan
But I find your statement on psu noise too general. I could also say the psu noise has NO impact on sound because it is on the supply not on the amp output.
But we know that some of it will sometimes end up at the amp output. How much, depends on the psu, the amp and probably a few more things like wiring and construction.
So is psu noise detrimental to the sound? It depends, on many things, so this can't be answered with a general statement.
Jan
Well that's a nice can of worms you're opening there
Hi and thanks for the kind warning
I would appreciate very much your opinion. I am quite puzzled at the moment.
The distortion showed is not very minimal ... and still you know ... the reviewer bought the testing samples and so on.
This is love at first listening ...
I think i should have kept the mouth shut
I agree with you ... there are things like the PSRR ... and so on.
I was trivial as always ... and please forget this silly digression of mine.
Going back to distortion ... what is your point of view ?
I am seriously interested.
Thanks a lot, gino
P.S. by the way to catch the fish you need the worm
There have been a number of well documented studies on the subject over the past 100 years or so. Google may turn up a few. Earl Geddes - a member here (Gedlee) has done some himself.
The audibility of distortion may be different in electronics and in loudspeakers. I seem to recall that some studies pointed to that, but you'd need to verify it.
The audibility of distortion may be different in electronics and in loudspeakers. I seem to recall that some studies pointed to that, but you'd need to verify it.
Hi and i hate electric guitars
No seriously there are not only electric guitars ... there are voices, unamplified instruments ... one thing can be good in one case and bad in another
I am looking for something of more universal
I understand that you have nothing against distortion in general
Thanks a lot, gino
Hi and thanks for the reply.
The fact that
to me means that the amount is much higher in speakers (i have seen impressive figures ... like 10% of distortion or more)
On this basis a 1% of distortion in the amp can be quite tolerable
That could very well be ... this is interesting for me because i am designing my first line driver preamp unity gain ... the fft is attached (thanks a lot again Mr Mooly
)Not bad for a single npn i guess
Maybe i should be less obsessed by distortion
Thanks and regards, gino
Attachments
Last edited:
I think it's a combination of two factors: is the distortion audible, and is it preferred to a less-distorting reproduction.
Remember that what makes musical instruments sound different is the difference in harmonics structure. So harmonics in and of itself are not always disagreeable (although harmonics may be dissonant).
The audibility depends very much on the experience of the listener, the type of music, the type of distortion, all kinds of psycho-acoustic factors etc etc. So this is a question that will never be answered for all cases.
Jan
Remember that what makes musical instruments sound different is the difference in harmonics structure. So harmonics in and of itself are not always disagreeable (although harmonics may be dissonant).
The audibility depends very much on the experience of the listener, the type of music, the type of distortion, all kinds of psycho-acoustic factors etc etc. So this is a question that will never be answered for all cases.
Jan
Notwithstanding my previous post, my own rule of thumb is that when there's at least one zero after the decimal in the THD, and the structure shows diminishing levels with increasing order, I'm not worrying
Jan
Hi and thanks again for the very interesting advice
If the original sound has harmonics i agree the replaying system should give back them
But it is the adding action ... the generation of harmonics not present in the original signal that it is hard for me to accept.
I would like a situation with nothing added and nothing taken out, just amplified
The audibility depends very much on the experience of the listener, the type of music, the type of distortion, all kinds of psycho-acoustic factors etc etc. So this is a question that will never be answered for all cases.
Notwithstanding my previous post, my own rule of thumb is that when there's at least one zero after the decimal in the THD, and the structure shows diminishing levels with increasing order, I'm not worrying
Jan
I agree there is no need to try to reach the 0.00001%
The problem is that i would expect to see some kind of convergence of designers on few solutions/approach
Maybe more than the actual parts/devices what matters is how they are used
In the sense that is possible to make great amps with tubes and solid state chips
Thanks again and kind regards, gino
Yes that's also what I want. Not easy - in fact, in theory impossible!
But have you ever looked at clipping? It is very well possible that audible differences between amps come from differences in clipping. Tube amps clip softly so can clip without very hard distortion.
SS clips hard; so even if it's a 100W amp. IF it clips it sounds horrible. That's why a 30W tube amp can sound as good or better than a 100W SS amp, even if the SS has lower THD!
The universe is not simple, nor is audio electronics
Jan
I've seen that written often, but haven't seen that on the bench. Single-ended stages will give a mix of even and odd order; push-pull stages will give almost exclusively odd order. But transient clipping doesn't tend to be that audible, and continuous clipping is bad no matter what.
The real issue in tube stages is blocking, which can turn a transient event into a very audible and obnoxious "choking." A lot, maybe the majority, of tube designs are prone to this issue. it's something that a few of us consider very significant in our designs and there's different ways of preventing it- Morgan Jones's direct coupled power amps are a good example. With all due modesty
, the Red Light District is another approach to minimizing or eliminating blocking.
And it's as simple as that. All audio systems add distortion to the original recording, extra to that in the encoded performance, and one always hears it; the key thing is then how the brain registers it, deals with it.
There's obvious distortion: guitar amps, transistor radios overloading; and then there's more subtle distortion, which is usually called the sound of a system, acoustic signature, etc, etc. No matter how expensive or refined a system is it's always there - so, what to do with it? The typical solution is to massage or flavour the distortion to suit, hence the constant cycling through of expensive gear by some, and endless fiddling by others, to get the 'nicest' distortion from their setup.
Which is where 'badness' comes in - this is when distortion personally offends you, disturbs the enjoyment of listening.
For myself, the solution is to minimise the subjectively perceived distortion. Nearly all conventional systems have too much of it very clearly audible, and I find it disturbing because it severely degrades more complex, so-called 'poor' recordings, at the subjective level. Luckily, if one goes about optimising systems in the right manner the distortion added by the playback is reduced to a level where it subjectively disappears, only the content of the recording, and its characteristics, registers strongly.
In direct response to your question, gino, distortion is the most critical problem in reproduction - only by substantially taming it can convincing, highly satisfying reproduction be achieved, IME.
Last edited:
Achieving consensus on allowable distortion is unlikely as it appears to be a matter of taste, although this is frequently denied. If you could be sure that smaller distortion is always better, for all listeners, then you could find out how the minimum detectable distortion varies from one person to another. If it turns out, as I believe, that some people (many people?) prefer 'some distortion' over 'no distortion' then it gets more complicated. You would first have to decide whether a person fits into 'less is better' or 'some is better' camps, then test them to find the boundaries of their perception.
My personal view is that distortion should be somewhat less than 1%, and dominated by low-order products. Don't press me on what "somewhat less" means!
My personal view is that distortion should be somewhat less than 1%, and dominated by low-order products. Don't press me on what "somewhat less" means!
I tend to focus on noise rather than distortion. That's because distortion as shown in your first plot is an artifact of the measurement stimulus. Single sinewaves are highly improbable in music so in my view are an idiotic way to test an amplifier. When an amp is stimulated with music (by which I mean the kind I mainly listen to, classical instrumental, orchestral and choral) the non-linearities in the amp produce modulation noise (noise correlated with the signal), not discrete harmonics.
Sine wave measurements (single, dual tone-CCIF, multitone) are the straight way to detect exactly the same nonlinearities (and than find way to supress them) , those caused "modulation noise" (intermodulation) at music reproduction..Here is nothing else, no magical "X" distortion. Amp with distortion as in first plot will create with complex signal very high intermodulation backround .
Last edited:
Hi and thanks for the valuable advice
For me the aim is, given a very well done recording, a realistic sound
For me distortion is the same as deformation.
Like a deforming mirror ... cannot give a realistic image.
So i suppose that very low distortion is important.
Moreover to perceive more minute sounds is also telling of the ability of the amp to resolve the small signals.
Yes i think very low distortion for me is important.
If the sound is clinical ... well let's go to the hospital then.
Thanks and regards, gino
Try to look at envelope of 30tone multitone signal, look like bandlimited noise., with amplitude varying from zero to maximum. Music is nothing special, for amp it is only signal, more or litle complex, but only signal. Same "work" for amp with e.g CCIF test or with symphonic orchestra. Do not look at amps "antropomorphically".
some speakers such as those using distributive hole ports or slits can have distortion >100% relative to a fundamental when driven by sine but with music - or as music used to be with acoustic or regular electric guitar and bass instruments, sound fine. I've heard speakers which can play sine at fb struggle with music.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.