I think Scott is correct to insist on proper descriptions - sloppy
naming creates a lot of confusion.
As far as I can see the patent is describing phase modulation
in the same way that the classic IM test involves amplitude
modulation. Phase modulation appears to be one of the
elements in the aforementioned "grouping mechanisms", but
I suspect that your modern amplifier is going to exhibit this
at the same low levels as amplitude modulation.
OK, so maybe Bob can make a comment. Sample the output of the amplifier at 24/192 and separate the real and imaginary components mathematically and extract the AIM and PIM. I read the patent and there was nothing there that did not have a digital analog.
Wow and flutter (in their normal context) are typically non-harmonicly related to the signal frequencies, this inexactitude implies non-physical results.
Wow and flutter (in their normal context) are typically non-harmonicly related to the signal frequencies, this inexactitude implies non-physical results.
Last edited:
You have to forgive my colleague, Ron Quan. He is just a working audio-video engineer located in Silicon Valley, who can still hear differences in audio equipment, and has invested his time and effort to do a paper for the AES, even knowing in advance that many will not appreciate it. He worked with me, BEFORE L & V took over the AES, keeping out radical elements pushing such measurements. I apparently poisoned his mind toward whatever works, rather than precise definitions. I would have not put 'Wow and Flutter' in as a perfect description of what he is measuring, but you must also understand that he also worked at Ampex, when magnetic tape was popular, before digital completely took over, and wow and flutter are common measurements that we made working at such companies. A loose definition of flutter is really dynamic FM modulation. Anyone here can look it up. I hope that some do.
You're right, I forgot. L & V have not had a go at it yet.
At least everyone doesn't have to travel to Waterloo.
I heard several comments at the AES to the effect that what he measured was real enough, but that it did not disclose non-linearities that were not known already and that could be measured by more traditional means. In other words a very ingenius and valuable way of providing another perspective at known shortcomings.
I'm sure that Bob C can comment more to the point, but as far as I know he's in a red-eye en route to NJ at this moment.
John, what's your take on it?
jan didden
It has been my firm 'belief' for the last 30 years, since real (in my opinion) effort has been made to find the underlying differences in audio electronic design that make some designs sound 'better' than others.
The first effort by Matti Otala, who was prompted by Mitch Cotter (a known FM designer) to put PIM or FM distortion or differential phase, as the most likely cause, was suppressed from publication in the JAES. Then it was looked at by Cordell and discounted as not being very important.
The subject sat in abeyance for about 15 years, until Barrie Gilbert looked more fully into the problems with op amps. Then, he gave a masterful QUANTITATIVE ANALYSIS of an idealized version of an op amp commonly used in some fashion by literally thousands of designs for the last 40 years. Here was the 'smoking gun'.
As usual, it was dismissed by many.
Then, Mitch Cotter told me what he had noted in Matti's and my preprint on TIM, first given in 1976 at the NY AES. This was the famous 741 TIM(30) measurement with 5V peak-to-peak output, and 30KHz -3dB rolloff. Several UNDEFINED IM looking tones that had no easy definition in terms of IM itself, apparently were FM sidebands, instead. This was noted more than 30 years ago by Mitch Cotter. He had told Matti Otala, but he had not informed me until about 1 year ago, or so. Looking at the data, and noting this insight about the FM sidebands mixed with the traditional IM on this thread, brought the same criticisms that Ron Quan is about to experience from his looking for FM or differential phase distortion more directly, through using video techniques.
For some reason, people do not appreciate the existence of FM byproducts, and usually claim them to be 'exotic' IM byproducts, if at all possible.
I do not wish, at this time, to parse with the definitions or device selection that Ron Quan used. However, the measurements are backed up by both test procedure and by mathematical calculations. Anyone seriously interested, can duplicate his tests and expand even further, as it is clearly stated how to do it in his paper and in his patent # US 7,550,977 B2.
I am MOST interested in why op amps, in particular, sound different in my experience, and this is the most likely reason. This testing might also give a more concrete reason why open loop bandwidth seems to make a difference, at least to me. We shall see.
The first effort by Matti Otala, who was prompted by Mitch Cotter (a known FM designer) to put PIM or FM distortion or differential phase, as the most likely cause, was suppressed from publication in the JAES. Then it was looked at by Cordell and discounted as not being very important.
The subject sat in abeyance for about 15 years, until Barrie Gilbert looked more fully into the problems with op amps. Then, he gave a masterful QUANTITATIVE ANALYSIS of an idealized version of an op amp commonly used in some fashion by literally thousands of designs for the last 40 years. Here was the 'smoking gun'.
As usual, it was dismissed by many.
Then, Mitch Cotter told me what he had noted in Matti's and my preprint on TIM, first given in 1976 at the NY AES. This was the famous 741 TIM(30) measurement with 5V peak-to-peak output, and 30KHz -3dB rolloff. Several UNDEFINED IM looking tones that had no easy definition in terms of IM itself, apparently were FM sidebands, instead. This was noted more than 30 years ago by Mitch Cotter. He had told Matti Otala, but he had not informed me until about 1 year ago, or so. Looking at the data, and noting this insight about the FM sidebands mixed with the traditional IM on this thread, brought the same criticisms that Ron Quan is about to experience from his looking for FM or differential phase distortion more directly, through using video techniques.
For some reason, people do not appreciate the existence of FM byproducts, and usually claim them to be 'exotic' IM byproducts, if at all possible.
I do not wish, at this time, to parse with the definitions or device selection that Ron Quan used. However, the measurements are backed up by both test procedure and by mathematical calculations. Anyone seriously interested, can duplicate his tests and expand even further, as it is clearly stated how to do it in his paper and in his patent # US 7,550,977 B2.
I am MOST interested in why op amps, in particular, sound different in my experience, and this is the most likely reason. This testing might also give a more concrete reason why open loop bandwidth seems to make a difference, at least to me. We shall see.
Last edited:
I think what would be of interest whether this new measuring method found artifacts that we were not aware of. That would really be something!
Do you believe that is the case?
As you said, the existence of FM modulation and other stuff mentioned by you was established some 30 years ago. Is Quan looking at the same phenomenon with a different pair of glasses, or did he see something new?
jan didden
Last edited:
Ahh yes I see. We're better at measuring now indeed, but often I am surprised that it isn't as much as I thought. For instance, at the AES I was again confronted with the AP S2 as 'the worlds highest resolution THD analyser', even when that equipment is at least 15 years old. The newer AP's have other features but are less sensitive. Why? There's apparently no (commercial) need for ultra-sensitive THD measurements.
Anyway, re: Quan's paper, I don't think the issue is wheter he brings more measurement sensitivity to the table, but I may be wrong. I hope John would clarify what he thinks about the method, whether it uncovers hithero unknown phenomena or whether it provides a new view into known shortcomings (which can be very valuable to be sure).
jan didden
Jan,
That is what I use and most of my research effort requires finding ways to extend it's range!
ES
Hi John,
It was nice to see you, even if only briefly, at AES. I had a chance to talk with Ron Quan after his paper and he seems to be a nice guy who is level-headed and credible. I have not yet had a chance to read his patent and preprint, but I certainly will in the next couple of days now that I am back.
One thing I recommend is that we not take him to task for his choice of semantics and terminolgy. I think we all know what he is talking about.
Rather, the key is to learn whether his work has uncovered any new insight into PIM or behaviors that were not previously discovered and/or predicted by the theory.
For example, one of the things I will look for is whether he has uncovered significant amounts of PIM in the absence of AIM. As I know it, PIM caused by negative feedback results from nonlinearity in the input stage (AIM) that is converted to phase modulation by the resulting variation of open loop gain. This mechanism of PIM creation thus fundamentally depends on the presence of AIM to make PIM. However, that conversion is well less than 100% efficient, so an amplifier creating PIM will also have substantial AIM.
The other thing that I will look for is an apples-apples comparison that shows whether PIM is exacerbated by larger amounts of NFB and smaller open loop bandwidth when closed loop bandwidth is held constant.
Cheers,
Bob
We were well beyond it 25yr ago with a Sound Tech 1700 and passive filtering.
Ed, I understand that, for instance from your interesting work on resistor distortion.
My point was that for producing better-sounding amps, there's no need for even lower THD sensitivity in our instruments than what we had the last 20 years or so.
jan didden
>>Several UNDEFINED IM looking tones that had no easy definition in terms of IM itself<<
So as not to be censored I will simply state again the relatively poor quality of the graphic misled people for a while. Every line on that graph was at an IM component, period. Let go of this fantacy and move on.
So as not to be censored I will simply state again the relatively poor quality of the graphic misled people for a while. Every line on that graph was at an IM component, period. Let go of this fantacy and move on.
- Status
- Not open for further replies.