8 questions
Thank you for your welcome words for me to join this forum.
Seen the reactions on my posting #704 I only have questions.
As I said in my posting #704, the focus of my work was not in high end audio and I did not follow the literature.
So I after those 37 years, what have been the key research results in audio?
1. My first question is about the state of the art of the knowledge how our ears do function. Because notwithstanding all nice circuitry, and theories, the end result should be acceptable for the ear. Is the ear a four-analysis instrument, and/or does it carefully ‘measure’ where the signal crosses the zero line, and/or does it detect phase dispersion, etc., etc.
2. If the answer on 1 is positive, in the sense that we know enough of the ear/brain mechanisms, have their been attempts to translate these to specs. of audio power amplifiers?
3. Have their been listening experiments by deliberately change the characteristics of a power amplifier on the listening results? For instance reducing with a simple network the open loop bandwidth of an amplifier that has an inherently a high open loop bandwidth?
4. Does a list of standard recordings exist (for instance CDs) with a variety of audio fragments, that are acknowledged as distortion free recordings?
5. If the answer on 4 is yes, are these standard recordings always used in all comparison listening tests?
6. Do comparison power amplifier listening tests always use the same standardized loudspeakers that are acknowledged for their good and neutral response, and done in a standardized listening room?
7. Are listening tests always done blind?
8. Have there been attempts to objectivize the outcome of the inherently subjective listening tests, or is it like winetasting, where in a few sentences the characteristics of the wine are described.
In 1972 when I designed the 25W amplifier, I had no answers on these questions; only the recommendations of Matti. Even his TIM theory I gave only the advantage of the doubt because he did not proof to me that it really existed, and if yes then may be only in mediocre designs.
If the ear would be phase sensitive, then a high open loop bandwidth just helps to reduce phase dispersion (on distortion if you like). So I thought nothing wrong to design a high open loop gain if it so easily obtained with a clever design.
The listening results in 1972 of our 25W amplifier were very good for my ears, and I could only describe it as very transparent, whatever that is. We did not use a standardized recording as a source, neither we did comparison tests with high end amplifiers, because we did not have them.
I never listened to the Electrocompaniet audio amp. Then, as a young engineer I did not have the money to buy it, and it did not happen that I found it by accident in an audio shop.
Later in my Philips time I visited the Marantz development labs in Tokyo once, in the time that Marantz belonged to Philips (not any more by now). I was there to discuss something else than their high end power amps. Before I left I could visit their demonstration room. I heard their most expensive power amp. with a batteries power supply (charged when the amp. was off), because their philosophy was not to be disturbed by 50 or 60 Hz when listening. The loudspeaker set alone did cost 50.000 Dollars they said. As I had to catch a plain I had only a couple of minutes to listen. What can I say? It sounded quite well, but I could not do a blind test to compare it with other amplifiers, and listening for a couple of minutes is much too short for a good judgment in my opinion.
Thank you for your welcome words for me to join this forum.
Seen the reactions on my posting #704 I only have questions.
As I said in my posting #704, the focus of my work was not in high end audio and I did not follow the literature.
So I after those 37 years, what have been the key research results in audio?
1. My first question is about the state of the art of the knowledge how our ears do function. Because notwithstanding all nice circuitry, and theories, the end result should be acceptable for the ear. Is the ear a four-analysis instrument, and/or does it carefully ‘measure’ where the signal crosses the zero line, and/or does it detect phase dispersion, etc., etc.
2. If the answer on 1 is positive, in the sense that we know enough of the ear/brain mechanisms, have their been attempts to translate these to specs. of audio power amplifiers?
3. Have their been listening experiments by deliberately change the characteristics of a power amplifier on the listening results? For instance reducing with a simple network the open loop bandwidth of an amplifier that has an inherently a high open loop bandwidth?
4. Does a list of standard recordings exist (for instance CDs) with a variety of audio fragments, that are acknowledged as distortion free recordings?
5. If the answer on 4 is yes, are these standard recordings always used in all comparison listening tests?
6. Do comparison power amplifier listening tests always use the same standardized loudspeakers that are acknowledged for their good and neutral response, and done in a standardized listening room?
7. Are listening tests always done blind?
8. Have there been attempts to objectivize the outcome of the inherently subjective listening tests, or is it like winetasting, where in a few sentences the characteristics of the wine are described.
In 1972 when I designed the 25W amplifier, I had no answers on these questions; only the recommendations of Matti. Even his TIM theory I gave only the advantage of the doubt because he did not proof to me that it really existed, and if yes then may be only in mediocre designs.
If the ear would be phase sensitive, then a high open loop bandwidth just helps to reduce phase dispersion (on distortion if you like). So I thought nothing wrong to design a high open loop gain if it so easily obtained with a clever design.
The listening results in 1972 of our 25W amplifier were very good for my ears, and I could only describe it as very transparent, whatever that is. We did not use a standardized recording as a source, neither we did comparison tests with high end amplifiers, because we did not have them.
I never listened to the Electrocompaniet audio amp. Then, as a young engineer I did not have the money to buy it, and it did not happen that I found it by accident in an audio shop.
Later in my Philips time I visited the Marantz development labs in Tokyo once, in the time that Marantz belonged to Philips (not any more by now). I was there to discuss something else than their high end power amps. Before I left I could visit their demonstration room. I heard their most expensive power amp. with a batteries power supply (charged when the amp. was off), because their philosophy was not to be disturbed by 50 or 60 Hz when listening. The loudspeaker set alone did cost 50.000 Dollars they said. As I had to catch a plain I had only a couple of minutes to listen. What can I say? It sounded quite well, but I could not do a blind test to compare it with other amplifiers, and listening for a couple of minutes is much too short for a good judgment in my opinion.
Re: 8 questions
Jan, this is a wonderful set of questions.
Many of us still have few answers to these questions.
Many would disagree on what those answers are.
Some others would argue that some of the answers to the questions don't matter.
There is a great divide in the audio community between so-called subjectivists and so-called objectivists. This polarization is unfortunate. There also seems to be too much labeling of people as being one kind or the other, when indeed there are many people who live in both camps.
It is a fascinating hobby.
Cheers,
Bob
Jan Lohstroh said:
Jan, this is a wonderful set of questions.
Many of us still have few answers to these questions.
Many would disagree on what those answers are.
Some others would argue that some of the answers to the questions don't matter.
There is a great divide in the audio community between so-called subjectivists and so-called objectivists. This polarization is unfortunate. There also seems to be too much labeling of people as being one kind or the other, when indeed there are many people who live in both camps.
It is a fascinating hobby.
Cheers,
Bob
Re: Re: 8 questions
Exactly. Subconscious perception can't be measured, but it have to be addressed primarily.
Bob Cordell said:
Exactly. Subconscious perception can't be measured, but it have to be addressed primarily.
Jan L. we have gone forward since those early years. Your little amp still sounds great. Electrocompaniet seems to have gone astray, since that first design that they got from your paper. PIM and other DIM related distortions seem to be just as important as TIM. You will have to look back on this thread and my CTC thread to get newer info. AES is next to useless here, at this time, since they blocked Matti from publication back in the 1980's.
john curl said:
John,
Can you give some reference to that 'newer info'? That's what we all have been waiting for. I noticed you regularly refer to Matti's and Jan's design in these discussions as a kind of yardstick. In fact, I often hear you say that Matti and Jan had "it" right and anything that has been written about TIM, DIM, PIM or what have you after that is dead wrong. So where's the progress?
@Jan: Your remarks about "objective-ating" (sp?) listening test by controlling known-good music, loudspeakers and blind testing are very to the point. But, as Bob alluded to, these are precisely the factors that often cause a deep rift between opinions of practioners here. Well controlled tests tend (but not always) to fail to detect audible differences that were believed to be present in casual uncontrolled tests, and this is very difficult to reconcile as you undoubtedly know.
Jan Didden
What an interesting thread !
I have a similar setup as the otala (as far as trimmable control over loop gain) with a much lower current voltage amp (3.5ma).
What I have found objectively is no change in overall slew or phase margin at UG. What does change is the phase in the audio band and the OLG (naturally), as well as the sonic character of the amp. (attached pix). I do not have to A/B the changes in loop gain (can adjust OLG AS I listen in real time). Going from 40db to 70+db can be heard easily. I always wondered what the 68k/220pf VAS shunts were in the symasym and how they affect the overall sonics of the amp. I haven't as yet experimented with the caps except in simulation but just the basic "crippling" is an eye opener.
"A" in pix is where I am subjectively evaluating now ,"B" is the maximum "cripple" I have attempted.
OS
I have a similar setup as the otala (as far as trimmable control over loop gain) with a much lower current voltage amp (3.5ma).
What I have found objectively is no change in overall slew or phase margin at UG. What does change is the phase in the audio band and the OLG (naturally), as well as the sonic character of the amp. (attached pix). I do not have to A/B the changes in loop gain (can adjust OLG AS I listen in real time). Going from 40db to 70+db can be heard easily. I always wondered what the 68k/220pf VAS shunts were in the symasym and how they affect the overall sonics of the amp. I haven't as yet experimented with the caps except in simulation but just the basic "crippling" is an eye opener.
"A" in pix is where I am subjectively evaluating now ,"B" is the maximum "cripple" I have attempted.
OS
Attachments
john curl said:
Hi Jan,
I agree with John that we have moved forward a great deal since those days. Matti was instrumental in raising many people's awareness of high-frequency distortion, even though not everyone agreed with his conclusions. One of the ways that we have moved forward is due to the fact that Matti proposed and used tests that could actually measure the distortions he researched, including TIM, PIM and IIM.
Armed with those tests, people were able to objectively assess the levels of those distortions of amplifiers with many different toplogies, open loop bandwidths, and amounts of NFB. A related way in which things have moved forward is that through those tests most designers realized and could show that those distortions could be minimized without resort to wide open-loop bandwidth and low feedback. In other words, one might conclude that Matti's approach to minimizing these distortions was sufficient but not necessary.
Designers also learned how to better frequency compensate NFB loops so that higher slew rates could be achieved.
Properly designed amplifiers of today of many different toplogies and amounts of feedback enjoy very very low values of these distortions.
Another great area of improvement in well-designed amplifiers is lower crossover distortion (Matti was very right to warn you about crossover distortion). Although not from the same mechanism as TIM, crossover distortion also shows up on the DIM test. The better availability of very fast bipolar power transistors (like RETs) and MOSFETs has enabled significant reductions in dynamic crossover distortion.
I think it is also fair to say that designers have a better understanding of how best to bias a class AB output stage (thanks largely to real-world application and adaptation of Barney Oliver's work). This results in much reduced static crossover distortion.
There is also much better awareness and attention to BJT power transistor beta droop and ft droop at high currents.
The more widespread understanding and use of DC servos has made it possible to get rid of the nasty electrolytic capacitor in the feedback network, largely eliminating a serious source of distortion.
More recently, ThermalTrak output transistors were introduced by OnSemi. By incorporating a sensing diode right on the header for the output transistor die, far better thermal tracking is available for use in setting up the class AB bias in the output stage. In my view, the availability of these transistors is nearly a game-changer.
Cheers,
Bob
Bob Cordell said:
Hi Bob
In my project amp (What is simple), I have a TDH in open-loop of ~0,6%-20KHz and ~0,1%-1KHZ with 50Wrms, and Slew Rate de 90 V/us.
-obviously with the negative feedback reduces the distortion....
I would like to know if with this distortion in open loop, if I would low distortion TIM, AIM and PIM?
In your project "Mosfet Error Correction" by my calculation, you have 0,1%-20Khz in open-loop which is very good...
I would like to know your opinion
Thanks
Hi Dimitri
Is an article, the bibliographic reference is cited in article...
Here is the full article:
http://ifile.it/dk3ytpv
Is an article, the bibliographic reference is cited in article...
Here is the full article:
http://ifile.it/dk3ytpv
Wavebourn said:
The circuit that I was referring to the previous post, would be to use only one common-emitter, that saturates easily, it starts sending strong second-harmonic distortion, is used in valve amplifiers that do not use negative feedback.
You are correct, with a cascode increases the region of amplification, but remember! using a differential pair without negative feedback, You will have a cascode configuration (Attached).
Attachments
What is ironic, Jan L. , is that virtually all the guidelines for proper amp design were published by Daugherty and Greinier in 1966, in the IEEE Trans. as footnoted in your later paper.
Your amp, using these principles, PROVED it sonically, and a later paper on measurement of TIM where I contributed, given in 1976 MEASURED both TIM and PIM in one test, even though Matti and I didn't not see it, at the time.
Yet, decades go by, and nobody, except some of the most successful designers of audio equipment, give it any notice, yet engineers still talk in circles about "what could be important?"
Your amp, using these principles, PROVED it sonically, and a later paper on measurement of TIM where I contributed, given in 1976 MEASURED both TIM and PIM in one test, even though Matti and I didn't not see it, at the time.
Yet, decades go by, and nobody, except some of the most successful designers of audio equipment, give it any notice, yet engineers still talk in circles about "what could be important?"
Tribute to Matti
To Bob Holmdel and John Curl,
Thanks for your reactions on my post #730 and for reporting that a lot has happened the last 40 years and that a lot of distortion phenomena are better understood and can be measured. Apparently some classes of circuitry are effective in producing almost distortion-free drivers.
I guess the first publication of Matti was: Otala, M.: Transient distortion in transistorized
audio power amplifiers. IEEE Trans. AU-18 no. 3, 1970, pp 234-239.
As this thread is a tribute to Matti Otala, may I suggest that the two of you write together an overview article with a title like: “40 years of understanding distortion in audio power amplifiers, a tribute to Matti Otala”, in which you can mention what, in your opinion, was correct what Matti and others said, and what not. You could then also refer to my 8 questions from post #730 and indicate what is needed to bring this profession further to a higher level of objective understanding.
I can offer to proof-read your article.
Cheers,
Jan Lohstroh
P.S. It is new to me that Matti was blocked by AES in the 80's. Why was that? I guess by now AES would enjoy the article I propose you to write.
To Bob Holmdel and John Curl,
Thanks for your reactions on my post #730 and for reporting that a lot has happened the last 40 years and that a lot of distortion phenomena are better understood and can be measured. Apparently some classes of circuitry are effective in producing almost distortion-free drivers.
I guess the first publication of Matti was: Otala, M.: Transient distortion in transistorized
audio power amplifiers. IEEE Trans. AU-18 no. 3, 1970, pp 234-239.
As this thread is a tribute to Matti Otala, may I suggest that the two of you write together an overview article with a title like: “40 years of understanding distortion in audio power amplifiers, a tribute to Matti Otala”, in which you can mention what, in your opinion, was correct what Matti and others said, and what not. You could then also refer to my 8 questions from post #730 and indicate what is needed to bring this profession further to a higher level of objective understanding.
I can offer to proof-read your article.
Cheers,
Jan Lohstroh
P.S. It is new to me that Matti was blocked by AES in the 80's. Why was that? I guess by now AES would enjoy the article I propose you to write.
Daugherty and Greiner
JohnC,
Do you have the article "Daugherty, D.G. and Greiner, R.A.: Some design objectives for audio power amplifiers. IEEE Trans. AU-14
no. 3, 1966, pp 43-48." available. I lost it.
Would be interesting to re-read it after all those years. Certainly when you say that they had it already right in 1966.
JanL
JohnC,
Do you have the article "Daugherty, D.G. and Greiner, R.A.: Some design objectives for audio power amplifiers. IEEE Trans. AU-14
no. 3, 1966, pp 43-48." available. I lost it.
Would be interesting to re-read it after all those years. Certainly when you say that they had it already right in 1966.
JanL
Let us ask Dimitri, first. He is our resident librarian. I do have the article, but I don't know where the LAST page (the important one with the equations) is, at the moment I used to carry it with my other info to tech meetings. We can get it in some easy way. Have you ever seen the whole PhD dissertation? Matti let me read it once.