Hi,
This we agree on. There is as such no point to make an amplifier behave like the ear, HOWEVER the point is that KNOWING the behaviour of the human hearing may inform our design priorities and give a different weight to certain aspects of reproduction than we might otherwise.
It basically exposes and busts the various audio myth continuously perpetuated.
HD and IMD are normally closely linked, if I see the two much differing when measured I already the Amplifier has non-linearities stemming from sources other than simple non-linearity of the gain devices.
I would not venture as to why SOME engineers LIKE ultra low distortion any more than I would venture as to why some people like Marmite..
Though the explanation why the urban engineering myth of the "low distortion" persists may more direct than yours, insofar that there is a belief that closer a given device comes to what is considered theoretical perfection, the better it is. Perfection in this case is an amplifier who's output is a complete and fathful replica of the input signal but amplified, into any load, for any signal.
The problem in this is of course that the commonly used measurements are quite myopic and do not account for many factors that anyone who does keep up with what is going on should know.
In reality ideal amplifiers do not exist. Real amplifiers are subject to many limitations and imperfections. A real engineer also understands that all design is a compromise and that if we make some aspects better there is invariably a price to pay, be it in actual increased cost or in degradation of other parameters.
So, for example relaxing the requirements for ultra-low THD & IMD to levels that are sensible given what we know about the human hearing can allow us to focus on improving transient behaviour and produce better immunity to RFI et al and can give a simple means by which thermal memory distortion may be reduced.
I would indeed note that focus on Low THD/IMD etc. seems to have more to do with the fact that for decades and indeed to this day and age "My Amp's distortion is lower than anyone else's" is a routine staple of advertising audio and because it is trivially easy to measure nowadays.
It is not being advanced as cause celebre because ultra-low distortion provides an improvement over "low enough" distortion that has been proven to be audible.
This may be so, however the results of such tests are very rarely if ever entered into debates here and even in scholary journals.
Reading thishere forum and some members posts one doubts rather doubts this. I do agree that a great deal is known, though rarely read and applied (one look at most modern amplifier designs suffices - most clearly have never even read Self, Groner and Cordell).
Yet more remains either unknown or worse actively disputed by the technocrati who often refuse to acknowledge real issues long after they have been demonstrated to be real by measurements and when test methods are in fact available to deal with them.
So the old "distortion" thing is thub-tumbed and welking-rung again and again and anything else is routinely rubbished as "illusion" and "myth". Hence the myopic response.
So if you disagree with the state of affairs, why not stop banging on about ultra low distortion and instead cover the different design compromises, show some of the results of noise loading and so on? It would make for a by far more interesting and useful debate.
Ciao T
This we agree on. There is as such no point to make an amplifier behave like the ear, HOWEVER the point is that KNOWING the behaviour of the human hearing may inform our design priorities and give a different weight to certain aspects of reproduction than we might otherwise.
It basically exposes and busts the various audio myth continuously perpetuated.
HD and IMD are normally closely linked, if I see the two much differing when measured I already the Amplifier has non-linearities stemming from sources other than simple non-linearity of the gain devices.
I would not venture as to why SOME engineers LIKE ultra low distortion any more than I would venture as to why some people like Marmite..
Though the explanation why the urban engineering myth of the "low distortion" persists may more direct than yours, insofar that there is a belief that closer a given device comes to what is considered theoretical perfection, the better it is. Perfection in this case is an amplifier who's output is a complete and fathful replica of the input signal but amplified, into any load, for any signal.
The problem in this is of course that the commonly used measurements are quite myopic and do not account for many factors that anyone who does keep up with what is going on should know.
In reality ideal amplifiers do not exist. Real amplifiers are subject to many limitations and imperfections. A real engineer also understands that all design is a compromise and that if we make some aspects better there is invariably a price to pay, be it in actual increased cost or in degradation of other parameters.
So, for example relaxing the requirements for ultra-low THD & IMD to levels that are sensible given what we know about the human hearing can allow us to focus on improving transient behaviour and produce better immunity to RFI et al and can give a simple means by which thermal memory distortion may be reduced.
I would indeed note that focus on Low THD/IMD etc. seems to have more to do with the fact that for decades and indeed to this day and age "My Amp's distortion is lower than anyone else's" is a routine staple of advertising audio and because it is trivially easy to measure nowadays.
It is not being advanced as cause celebre because ultra-low distortion provides an improvement over "low enough" distortion that has been proven to be audible.
This may be so, however the results of such tests are very rarely if ever entered into debates here and even in scholary journals.
Reading thishere forum and some members posts one doubts rather doubts this. I do agree that a great deal is known, though rarely read and applied (one look at most modern amplifier designs suffices - most clearly have never even read Self, Groner and Cordell).
Yet more remains either unknown or worse actively disputed by the technocrati who often refuse to acknowledge real issues long after they have been demonstrated to be real by measurements and when test methods are in fact available to deal with them.
So the old "distortion" thing is thub-tumbed and welking-rung again and again and anything else is routinely rubbished as "illusion" and "myth". Hence the myopic response.
So if you disagree with the state of affairs, why not stop banging on about ultra low distortion and instead cover the different design compromises, show some of the results of noise loading and so on? It would make for a by far more interesting and useful debate.
Ciao T
Hi,
I know you intended what you wrote to be cynical, but what makes it really funny (I am still holding my side) is that it actually describes reality rather well...
The whole concept of a linear frequency response is a complete illusion, created in "my day" by judicious pen-damping and nowadays by the use of averaging.
If you ever saw the raw, unaveraged frequency response of speakers that are presented as having a flat frequency response, you would be appalled at the fact that they are not even +/-10dB flat.
Another facette is one particular speaker myth that I keep noticing, namely that of the "Baffle Step", which actually produces a 6dB LF Boost in the power response.
Yet most here and many commercial speaker designers use it anyway and insist it must be used, creating what is in effect (and this indeed is the reason for preferring to apply it) a fixed, undefeatable "Loudness" contour to compensate for listening levels that are lower than "realistic".
Ciao T
I know you intended what you wrote to be cynical, but what makes it really funny (I am still holding my side) is that it actually describes reality rather well...
The whole concept of a linear frequency response is a complete illusion, created in "my day" by judicious pen-damping and nowadays by the use of averaging.
If you ever saw the raw, unaveraged frequency response of speakers that are presented as having a flat frequency response, you would be appalled at the fact that they are not even +/-10dB flat.
Another facette is one particular speaker myth that I keep noticing, namely that of the "Baffle Step", which actually produces a 6dB LF Boost in the power response.
Yet most here and many commercial speaker designers use it anyway and insist it must be used, creating what is in effect (and this indeed is the reason for preferring to apply it) a fixed, undefeatable "Loudness" contour to compensate for listening levels that are lower than "realistic".
Ciao T
This situation is quite easily understandable. We must admit, that there is no mechanism for "truth" verification in the field of audio.
What I mean, just compare audio and space exploration. Space engineer can not hide behind dogms or believes, since his erratic vision can lead to a crash. In the field of audio, audio engineer is responsible mainly for an incomes from mass produced mid-fi equipment, that is orders of magnitude higher than at the high-end market. So, what could be a stimulus for changing this convenient status-quo?
Not forget also about strong inertia, both among university professors and among reviewers of scientific magazines. Are professors interested to agree, that they must learn everyday and hard work over updating their lecture courses?
Are reviewers interested to recognize, that all they achieved due the old fashioned published researches, does no longer meet requirements of society?
Better if forum discussions would be free from the constraints mentioned, and people would be in "free" flight, giving freedom for their creativity. Thanks, Thorsten, for your efforts at this forum. The fight is nothing of being easy.
Last edited:
This is utterly true, "my amp's distortion is lower than your's" is unfortunately a paradigm, that is convenient to almost everybody, at first, for unexperienced and uninformed buyers of mid-fi equipment. Manufacturers and engineers are oblidged to "to keep face" at this situation, prooving by all measures, that opinion of such buyers is not wrong.
Reproduction, in your understanding, has been already done for decades. Other camp, is trying to enlight, that everything is relative, temporal, and does not fit to dogmatic vision, that all the measures you use to control "reproduction", suites well designing of sensors and technical instrumentation, but can not cover situation with human hearing. Do we need prooves for such a position? One of the prooves, that such camp, which doubts that "reproduction" has already been solved, becomes wider and wider.
None of the amplifier could achieve absolute copy of reality, good ones must be close to reality according to correct metrics, not THD. A given amp device is a collection of various compromises, and its impossible to made one, being good both in correct metrics AND in THD. On the contrary, pursuing THD, we INEVITABLY trash out correct metrics, dynamics precision ets. Therefore, we observe more and more opposition to using very deep NFB.
Last edited:
Hi,
Depends on SPL and frequency.
Yes. It would etc etc
I stand by my assertion that, all the engineering caveats considered and fulfilled (and please note that caveat), lower distortion is better. And no amount of pseudo techno babble will convince me otherwise.
Better for the ear brain system to introduce the non linearities that the man made part of the signal chain. Period.
Thankfully, I will never share the same space as some people here when it comes to thinking about audio systems and design. And that includes the gold fuses.
Hi,
Lower distortion is a nebulous term.
I believe it needs clearer definition as to what do you advocate?
Do you advocate lower transient distortion? Lower thermal memory distortion? Lower harmonic distortion? Lower Intermodulation Distortion? Lower "fuzzy distortion"?
If you advocate that ALL form distortion should be lowered, are they all equally important to be lowered, or should we actively attempt to minimise certain types more than others?
And if so, why?
Sadly the man-made part of the signal chain invariably introduces non-linearities. So it would seem we are back at needing to define what forms of distortion we wish to minimise, how to quantify them and what measures to take to achieve this end.
I am unsure as to what gold fuses have to do with the debate if and what kinds of distortion in amplifiers are a valid indicator of quality or not.
I do observe that fuses as well as fuse holders can measurably and appreciably contribute to distortion in amplifiers, however I somehow get the feeling that you are not in support of this "fuses make observable differences" observation.
I think you may need to clarify your point of how gold fuses fit into this debate.
Ciao T
Lower distortion is a nebulous term.
I believe it needs clearer definition as to what do you advocate?
Do you advocate lower transient distortion? Lower thermal memory distortion? Lower harmonic distortion? Lower Intermodulation Distortion? Lower "fuzzy distortion"?
If you advocate that ALL form distortion should be lowered, are they all equally important to be lowered, or should we actively attempt to minimise certain types more than others?
And if so, why?
Sadly the man-made part of the signal chain invariably introduces non-linearities. So it would seem we are back at needing to define what forms of distortion we wish to minimise, how to quantify them and what measures to take to achieve this end.
I am unsure as to what gold fuses have to do with the debate if and what kinds of distortion in amplifiers are a valid indicator of quality or not.
I do observe that fuses as well as fuse holders can measurably and appreciably contribute to distortion in amplifiers, however I somehow get the feeling that you are not in support of this "fuses make observable differences" observation.
I think you may need to clarify your point of how gold fuses fit into this debate.
Ciao T
jkeny;2670573[snip said:
It is IRRELEVANT what the ear does, as long as the amp reproduces the original source faithfully. To do that, the amp needs low distortion. Low enough so that it is ideally below the threshold of the ear to hear a difference. If the ear can discern a difference between 1% thd and 2% thd than obviously for faithfull reproduction, the amp should distort less than that. It is IRRELEVANT that the ear internally distorts 10, 20, 30 or 40% to finally conclude that two incoming sounds are different because one has 1% 2nd harmonics and the other has 2% 2nd harmonics + 1% 3rd, to mention just an example. And I believe we agree that the ear can do that, although the absolute lower threshold is up for discussion.
jan didden
Last edited:
It is easy to buy an amp that is so transparent that nobody can hear the difference between the input signal and the output signal. Correct reproduction, for all practical reasons, has indeed been reached long time ago.
All the doubts that you mention, and the 'opposition to deep feedback' always comes with either no arguments, or with arguments that indicate a lack of understanding or confusion. Like one post I read that said: 'NFB destroys transients', a nice example of something that simply isn't true but lots of people that do not have a good understanding, read it and say: see! I told you NFB is no good! Great entertainment!
jan didden
As low as possible difference between input and output (we are talking about amps..).in all aspects (linearity, speed, "robustness"), at all power levels, with all "realistic" audio signals and loads.
Do You expect by musicians at live concert they should "improve" their production according Your taste? Or should they add in Beethoven music something by Mozart or something remove?
But the same You are expecting by reproduction chain... Reproduction is not "relative", only perception of different individuals..
Last edited:
Great. Not much reason then to continue spending time reading and contributing in this thread.
NOBODY can hear a difference, perfect.
We shall better enjoy the music then, instead of spending hours teaching successful, but stubborn designers which do not share this truth and protecting "people that do not have a good understanding" for spending too much money.
We my even shut down half of this forum, in that case.
- Status
- Not open for further replies.