ra₇;4565473 said:
THD is "H" because the driving signal is also "H", either nominally sinusoidal either in simulation, or in measured situation, or rarely another periodic signal function. Square waves being second most common. Triangle-waves are popular, as are the good old crisp sawtooth. Each has very well known harmonic structures, so as the (invalid) theory goes, one computes the frequency transform (Fourier, stepwise bandpass, z-gate, etc) of both the driving signal and the output signal, scaled to normalize out amplitude differences maximally, and there you go. “Harmonic distortion”.
MUCH more interesting is what happens when you feed the same amplifier either pink or white noise. Real noise or synthetic, makes no difference. Well behaved noise generators smoothly bind noise within a maximum amplitude envelope, so as to keep things in bounds. The most notable thing when doing the analysis this way is that “harmonic distortion” essentially evaporates. The decoherence of linear transfer distortion thru use of white/pink noise erases the convenient - but wrong - idea that the nonlinearities of the transfer curve are actually “harmonic”. There are by definition, no harmonics to a noise spectrum. There are though, for a given window, specific frequencies and phases which mathematically characterize the sampled waveform. Since the results are not “easy to measure”, and even less easy to “characterize” (no longer being harmonically related to the constant, unchanging driving periodic waveform) … well, they're not done.
By anyone.
Except us sneaky mil-spec RADAR theorists.
Who can definitely deconvolve noisy reflections and derive signal from them.
'Cuz we like math.
LOL
GoatGuy
So you're betting on the "possibility" that all the scratches on your car will be in select, non observable places?
The very definition of optimism.
Then call it what it is. Clipping. Everything goes for a dump when you start clipping.
dave
Exactly. Once the information in the HD is collapsed, you don't have much left.
HD should be displayed as a 4 dimensional chart showing frequency of input, level of input, quantity and frequency of output. Since we are usually viewing a 2D display, a rotatable projection of a 3D graph of 3 of them, with a slider for the 4th. I'd guess input level would be the most useful, but each permutation likely shows something useful.
That would require a higher level of understanding to interpret.
dave
Wow this thread exploded. I'm glad. Most of the comments I would have made were made already. Mostly by Arny. I just wish there was an accepted HiFi 101. Like any other topic, all the first year biology students learn what a cell is, they all speak the same language. I think in terms of home audio Geddes work is a great place to start. I just wish more people agreed. He has done so much of the work to cut the bull and get to the point.
Yes. This obvious point seems to confuse some people. They ask about distortion of components or circuits, not realising that it is only the triple (components, circuit, signal) which has distortion. I'm sure some people think that a circuit (or a component) simply sprinkles in distortion like it sprinkles in noise.arnyk said:
The fact that 50% THD seems to alarm people shows that THD tells us something about an amplifier - perhaps that it is broken.
Wow (what's wrong with repeating what Arny had said?). I automatically glanced at your "join date". August 2015? Many new comers had come to this site with typical unique perspective and assumptions, thinking that they might have brought new things to the table. Look at Dave's join date. 2001. He must have read the same comments more than a hundred time 😀 Many of the old members are even bored to repeat or discuss the same topic again and again.
But I have also observed some new comers coming with a typical perspective, talking about science, golden rules, bla bla bla then disappeared. One observation I made is that we who stay here longer is not because we like to have discussion. We are here because we build things, even chasing the holly grail of sound... 🙂
How typical dyiaudio.... insulting and chiding other members, without offering any counterpoint/ argument.... because, see, if all you spew is insults, how could you be proved wrong?
That, my good chap, was … useless. How about starting with any critique of the paragraphs I posted? Factual, not surmise! Tha'd be nice.
Happy New Year too!!!
GoatGuy
I don't see the basis for the excitement over multitone or noise test methods in audio
for audio we really have circuits that can be very successfully described as "weakly nonlinear" if operated within design limits
and the tests are hardly unknown in audio with Klippel, AP analyzer menu selections, and whitepapers, journal articles...
for audio we really have circuits that can be very successfully described as "weakly nonlinear" if operated within design limits
and the tests are hardly unknown in audio with Klippel, AP analyzer menu selections, and whitepapers, journal articles...
Oh for sure, I've seen the same thing in a particular group that I had been a part of from the beginning. New people would come in and point things out that we all knew. The problem was though that they were able to point it out, it means us 'old stock' hadn't dealt with it.
The best way to chase the holy grail of anything that adheres to the laws of nature. (Sound waves and our brains interpretation of them being such) is through science.
Last edited:
Isn't it about reproduction of sound fields? Ideally a 1:1 correspondence. But of what exactly? The impression of sound and musical sound is first of all a computation of the human auditive "system" . Visual and auditive perception is "logarithmic" this property is quite important for survival in wilderness. This is more or less "mechanical". But the other features of impression of some perception are beyond the realm of scientific methods.
No contemporary science can explain, that is derive a model, of the auditive system of a musician who simply does not hear the crackling and resonances of a original Caruso shellack. He hears the music only.
Thus one has to apply scientific methods on the original sound field. What is interpreted as musical sound after said unknown information computing looks objectively in time domain as a series of bursts, envelopes formed by sinusoidal signals. Depending also on the acoustical properties of the environment for ex. concert hall, crest factor can be up to 30 dB. Not a problem for the logarithmic perception, but can be a problem for a linear recording and reproduction system. psychoacoustic experiments have shown that the , say "mood" that music is intended to transport depends much on the said envelopes and
crest factors. Human musical sound recognition appears to be very sensitive to errors in rise rate of slope of the envelope, that is errors in time. If crest factor is limited or compressed, the impression becomes "unexciting".
Insofar one could establish requirements of reproduction of sound fields which would however be based on systems theory, because human auditive system must be included.
And of course loudspeakers and acoustical properties of living room.
Btw. the crest factor of ideal mathematical Gaussian noise is "infinite". The sound field of strong wind with heavy gusts comes a bit close to that ideal, the crest factor is much higher than present recording and reproducing can handle. Thus even with ultra high end you will not hear a storm with gust in your living room.
No contemporary science can explain, that is derive a model, of the auditive system of a musician who simply does not hear the crackling and resonances of a original Caruso shellack. He hears the music only.
Thus one has to apply scientific methods on the original sound field. What is interpreted as musical sound after said unknown information computing looks objectively in time domain as a series of bursts, envelopes formed by sinusoidal signals. Depending also on the acoustical properties of the environment for ex. concert hall, crest factor can be up to 30 dB. Not a problem for the logarithmic perception, but can be a problem for a linear recording and reproduction system. psychoacoustic experiments have shown that the , say "mood" that music is intended to transport depends much on the said envelopes and
crest factors. Human musical sound recognition appears to be very sensitive to errors in rise rate of slope of the envelope, that is errors in time. If crest factor is limited or compressed, the impression becomes "unexciting".
Insofar one could establish requirements of reproduction of sound fields which would however be based on systems theory, because human auditive system must be included.
And of course loudspeakers and acoustical properties of living room.
Btw. the crest factor of ideal mathematical Gaussian noise is "infinite". The sound field of strong wind with heavy gusts comes a bit close to that ideal, the crest factor is much higher than present recording and reproducing can handle. Thus even with ultra high end you will not hear a storm with gust in your living room.
I also look at join dates, but in this case the short time is GOOD 😱
Mindsource just arrived at the Palace party and not being a longstanding Court member .... sorry .... a longstanding Forumite , he does not know that "some things are best not said, even if glaringly obvious" and so said in a loud and clear voice: "but .... but .... the Emperor is naked !!!!!!!! " .
Boy!!!!! , what a brawl he triggered !!!!!!!!
Personally, I'm happy every now and then somebody says something true and meaningful, even if Politically Incorrect ..... at least to the eyes of some 🙄
No insulting there. It was mentioned because I thought it was not known. It might help people to change the way they run the debate or discussion.
There are different kind of debates. For me, it is the way to achieve agreement, including agreeing to disagree. Each party must try to understand what others are trying to say. A party may want to seek something (truth, knowledge, etc), other may want to share knowledge (sincerely) to others who appear to need it.
Others simply want to appear important, more knowledgeable than others, want to prove that others are wrong or simply want to win debates...Or practicing those debating tricks 😀
What I was saying is, a new member might come here from different neighborhood which most probably less knowledgeable than DIYaudio (such as other websites on the net). Thinking that they brought something new and never heard before. Assuming that others don't know anything. This is what prevents parties to have fruitful debate/discussion.
Last edited:
Somebody says something true and meaningful AND (assuming that) somebody else says something untrue and meaningless?? Problem is, I often found that many debates are actually like the story of an elephant and seven blind men. Each blind man can be considered correct, but can they see the big picture?
What is often forgotten is that we are talking about the human experience.
The scientific types here on this forum assume if it cannot be heard by the majority then it is inaudible.
And they show tests which prove the majority cannot hear "x" therefore no one can.
They do not seem to remember in their statistics class that the tests have an error of +/- % of the sample set and it does not apply to everyone.
Still others use themselves as a measure.
They cannot hear it therefore no one can.
The scientific types here on this forum assume if it cannot be heard by the majority then it is inaudible.
And they show tests which prove the majority cannot hear "x" therefore no one can.
They do not seem to remember in their statistics class that the tests have an error of +/- % of the sample set and it does not apply to everyone.
Still others use themselves as a measure.
They cannot hear it therefore no one can.
- Status
- Not open for further replies.