Thanks Gerard, for your very useful input. Now, does SYN08 know where the 4.99KHz is coming from? Also, everyone, please note the SYMMETRY of the first two (unknown) tones measured by Gerard, and noted on Fig. 3 from our paper, of 33years ago. 4.99-3.18=? 3.18-1.37=? Close enough for me.
Now what could possibly create these (unknown) symmetrical distortion products? Hint: FM is a good possiblity, AM maybe, but no creation mechanism has been developed for this, as of yet.
Thanks again, Gerard, you may have saved me a lot of time and trouble.
Now what could possibly create these (unknown) symmetrical distortion products? Hint: FM is a good possiblity, AM maybe, but no creation mechanism has been developed for this, as of yet.
Thanks again, Gerard, you may have saved me a lot of time and trouble.
Last edited:
John,
The major problem you face here is not just blind conviction that all which was learned - often many years ago - was learned by rote, and often without fullest natural understanding - so that any challenge to a small part of the foundation of that which was subsequently built on that knowledge may just threaten the whole.
The spirit shown by those who put up argument against any novelty - or change of view-point - is rather similar to that of theological literal fundamentalists. (I was an Irish Episcopalian theological student in Dublin ....I know the signs of resistance to open-minded challenge. " It is written...." Usually with a
However I can understand that you rile people!!!! But clearly more gentle methods have failed. If the books have not given an answer does not mean that the question is invalid. Switch from an orthodox/acceptable viewpoint is the basis of quantum theory after all! The only one who is contributing as a real empiricist appears to be PMA. (Scott may have too much commercial pressure to totally disclose - hence, possibly, his reticence in talking on these matters with Dick S.)
The major problem you face here is not just blind conviction that all which was learned - often many years ago - was learned by rote, and often without fullest natural understanding - so that any challenge to a small part of the foundation of that which was subsequently built on that knowledge may just threaten the whole.
The spirit shown by those who put up argument against any novelty - or change of view-point - is rather similar to that of theological literal fundamentalists. (I was an Irish Episcopalian theological student in Dublin ....I know the signs of resistance to open-minded challenge. " It is written...." Usually with a
However I can understand that you rile people!!!! But clearly more gentle methods have failed. If the books have not given an answer does not mean that the question is invalid. Switch from an orthodox/acceptable
viewpoint is the basis of quantum theory after all! The only one who is contributing as a real empiricist appears to be PMA. (Scott may have too much commercial pressure to totally disclose - hence, possibly, his reticence in talking on these matters with Dick S.)Hi Demian,
the overtones of musical instruments rarely are integer multiples of the fundamental frequency. Strings and pipes are inherently inharmonic. Inharmonicity can also intentionally be there for a desired expression. The sound spectra of clarinets have strong odd harmonics in in the low register (characterized by a mellow warmness).
The distortion mechanism in amplifying devices and circuits is well-understood. The audibility and perception of distortions are pretty well-understood. The pronounced disability of measuring techniques to provide conforming data is not widely well-understood.
the overtones of musical instruments rarely are integer multiples of the fundamental frequency. Strings and pipes are inherently inharmonic. Inharmonicity can also intentionally be there for a desired expression. The sound spectra of clarinets have strong odd harmonics in in the low register (characterized by a mellow warmness).
The distortion mechanism in amplifying devices and circuits is well-understood. The audibility and perception of distortions are pretty well-understood. The pronounced disability of measuring techniques to provide conforming data is not widely well-understood.
Brianco, the voice of reason! I do wish that I could be more effective, but in my sorry state, I have experience, but not the energy, to 'prove' everything that I find useful.
PMA, a respected competitor, is surely on track to find the best way to do audio designs.
I had hoped that Scott could learn from another source, perhaps more effectively than from me, and pass it on. I certainly can't do it very well.
PMA, a respected competitor, is surely on track to find the best way to do audio designs.
I had hoped that Scott could learn from another source, perhaps more effectively than from me, and pass it on. I certainly can't do it very well.
I'm taking a break from my trip preparations, for one comment. 11*3180 - 2*15000 = 4980 pretty simple math. If we can't read frequencies off of a 33yr. old mechanical chart recorder to 1% accuracy, you can't either. There is no mechanism for the creation of inharmonic frequencies, I'm beginning to think we are being put on.
It's pretty much irrelevant. A current source can be transformed, using the Norton theorem, in an equivalent voltage source, so a non-ideal current source maps to an equivalent non-ideal voltage source (having a non zero series impedance).
From this point, you already know the answer. The linearity is in no way better ot worse than the opamp open loop gain linearity, where the opamp is fed by a non-ideal voltage source. The source impedance may or may not matter, however I would say that for most practical cases the input capacitance nonlinearity is orders of magnitude lower than the (e.g.) Cb'c of the VAS device(s), subject to a large signal swing.
I would think that Brianco has also good chances to become a respected and appreciated high end audio designer, isn't he? He has all the required qualifications, including advanced Irish Episcopalian theological education.
You are incorrectly assuming that we have an agreement about what "inharmonic frequencies" are.
I guess, I mean 60Hz and 5kHz will never make 133Hz (at least not through nothing but a uA741).
Yes Nelson, though the quote gave it away. RB helped me escape from MIT (funny story).
So, output impedance of a voltage to current converter greatly depends on it's gain, plus input impedance of current to voltage converter depends on it's own gain.
Now, what if their output and input impedances are defined more by external resistors than by sand?
Very sarcastic - not
- we know where you stand in this matter. 
For your information I am an end user of audio equipment; I buy things. I am not a designer, have no wish to be such, but I am rightly very very concerned with the attitudes of many on this thread....You are the designers be you engineers - or not! What you are not is metaphysicians trying to get an extra few fairies dancing on a pinhead. Nor are you artists so there is no excuse for histrionics.
Feet in the clay 'scientific' blinkerdom is NOT 'scientific'; nor does it lead to advance. No advance was ever made without at least looking in the direction one may go followed by 'an act of faith' (taking a risk). But of course such vision is not required if one wants to stay rooted in the past and grow into being an ever increasing anachronism. After all, if you want to live in an ever loftier ivory tower no words from the marketplace can prevent that. Progress is about controlled advance and managment of change. Those of you who have learned the basics and may well have written doctoral theses are best positioned to lead that advance with regard to audio replay equipment. Too many in the commercial audio world have for far too long designed within what are no more than variations on a limited set of themes. There has been little breakthrough in the past many years. What has masqueraded as advancement is applying fixed ideas to new chips etc. thrown out by the makers as being advancement and few of which challenge the simple JFET properly integrated in a well conceptualized design. Those who have made changes for the better, big or small, have taken a step or two aside and assumed that all which has gone before is not of necessity holy writ. (Even J.C. - despite his initials - would not claim to be speaking from the chair of Peter!!)
All I want to see is some advancement which may give some of you exceptionally capable and highly trained people - and that includes you Syn08 - a real reason to think more freely and creatively and thereby to design BETTER audio equipment than ever before. And I , for avoidance of doubt, am talking about stuff which SOUNDS more real and is affordable, NOT stuff which is exonerated by lab results alone.
You don't have to agree with John or even listen to him, but you do have to strive toward advancement of your conceptual launch pad!. If you do not look for better designs at the 'enthusiast' budget level then the quality of music software will ever decrease and the catalogue shrink to include little more than ever more compressed standards and stuff for adolescents.
I side with John simply because he is attempting to force you to THINK CREATIVELY.
11*3180 - 2*15000 = 4980
2*15000 - 9*3180 = 1380
All the other blips will follow suit.
This is probably a waste of time, but...
I sure wish someone (ahem) would get themselves a copy of The Volterra and Wiener Theories of Nonlinear Systems(*) by Martin Schetzen. The Volterra series is an exact description of the input-output relationship for nonlinear time-invariant systems. There are a couple of chapters in the book that discuss the synthesis of the Volterra operators. For example, the first-order Volterra operator is just the convolution integral from linear system theory. The second-order Volterra operator is a double integral that determines the second-order distortion components. Likewise, the third-order Volterra operator is a triple integral that determines the third-order distortion components - and so on to infinitely-high-order operators. This analysis is able to (theoretically at least) determine the exact distortion of nonlinear systems with memory, of which PIM is just one case - provided the system is time-invariant.
The synthesis of the operators consists of finding block diagrams that implement the most general possibilities for each operator. It turns out that these block diagrams contain only three types of elements.
1) Single-input, single-output linear time-invariant (LTI) blocks
2) Ideal two-input analog multipliers
3) Summers
As you might imagine, using the Volterra series for the computation of the output for a given input gets very complicated very fast. But if all you want to do is determine what the frequency components of the output are for the case of, say, an input which is the sum of two sinusoids at frequencies f1 and f2, this is much easier than explicitly computing the exact output. The LTI blocks do not add any frequency components not present at their input, so they can conceptually be replaced by "gain of one" blocks. Once this is done, what you end up with is nothing more than the block diagram of a simple Taylor series input-output relationship that one would use to analyze distortion of a static, memoryless (AM-to-AM) nonlinearity. Hopefully, we all agree that for static nonlinearities, the output frequency components are |n*f1 +/- m*f2|, where n and m are integers. This reduction of the block diagrams for the Volterra operators shows that even for dynamic nonlinearities such as PIM, these frequency components are the exact same as the static-distortion-only case.
Therefore, if distortion frequency components other than |n*f1 +/- m*f2| are present, the system must be non-time-invariant.
Just for grins, I've included a capture of a block diagram of the most general case of second-order Volterra operator. This is for second-order distortion components only, and does not include the convolution integral for computing the "linear part" of the output. Each of the M tandem branches leading into the summer has the same frequency components at its output as its adjacent branch, so for determining the frequency components of the output, there is no loss of generality in setting M=1 here.
(*)Heh heh, he said "Wiener"
I sure wish someone (ahem) would get themselves a copy of The Volterra and Wiener Theories of Nonlinear Systems(*) by Martin Schetzen. The Volterra series is an exact description of the input-output relationship for nonlinear time-invariant systems. There are a couple of chapters in the book that discuss the synthesis of the Volterra operators. For example, the first-order Volterra operator is just the convolution integral from linear system theory. The second-order Volterra operator is a double integral that determines the second-order distortion components. Likewise, the third-order Volterra operator is a triple integral that determines the third-order distortion components - and so on to infinitely-high-order operators. This analysis is able to (theoretically at least) determine the exact distortion of nonlinear systems with memory, of which PIM is just one case - provided the system is time-invariant.
The synthesis of the operators consists of finding block diagrams that implement the most general possibilities for each operator. It turns out that these block diagrams contain only three types of elements.
1) Single-input, single-output linear time-invariant (LTI) blocks
2) Ideal two-input analog multipliers
3) Summers
As you might imagine, using the Volterra series for the computation of the output for a given input gets very complicated very fast. But if all you want to do is determine what the frequency components of the output are for the case of, say, an input which is the sum of two sinusoids at frequencies f1 and f2, this is much easier than explicitly computing the exact output. The LTI blocks do not add any frequency components not present at their input, so they can conceptually be replaced by "gain of one" blocks. Once this is done, what you end up with is nothing more than the block diagram of a simple Taylor series input-output relationship that one would use to analyze distortion of a static, memoryless (AM-to-AM) nonlinearity. Hopefully, we all agree that for static nonlinearities, the output frequency components are |n*f1 +/- m*f2|, where n and m are integers. This reduction of the block diagrams for the Volterra operators shows that even for dynamic nonlinearities such as PIM, these frequency components are the exact same as the static-distortion-only case.
Therefore, if distortion frequency components other than |n*f1 +/- m*f2| are present, the system must be non-time-invariant.
Just for grins, I've included a capture of a block diagram of the most general case of second-order Volterra operator. This is for second-order distortion components only, and does not include the convolution integral for computing the "linear part" of the output. Each of the M tandem branches leading into the summer has the same frequency components at its output as its adjacent branch, so for determining the frequency components of the output, there is no loss of generality in setting M=1 here.
(*)Heh heh, he said "Wiener"
Attachments
Last edited:
Really?
"You are dork! We did all that of this and that 40 years ago!"
Can you show me please something revolutionary that John did, or demonstrated, during the life of this particular thread, except empty statements about how progressive he and his friends are, and how badly we oppose him?
What John tries to suggest, only that on the picture made 30 years ago he sees presence of PIM, and this presence is signified by non-harmonic sidebands
If it is a New Word in The Progressive Science, I am The Pope.
Last edited:
Math again? We need no stikin' math. You better look at Fig. 3 from our paper, of 33 years ago, that's the Truth.
- Status
- Not open for further replies.