john curl said:
I agree, but this is not the issue I am trying to address, which is:
As usual, you did not even bother to read my comment. This won't make your blurbs about "Linear distortion" you anywhere right.
Of course objectively, i.e. DA and distortion. But there's one more point: if used for compensation, in particular above say 1MHz, polystyrene caps have too much self induction.
edit: in case of the same voltage rating polystyrene is of course better (less DA etc.)
edit: in case of the same voltage rating polystyrene is of course better (less DA etc.)
OT
SY, what do you think about Robert T. Hodgson´s result in comparing wine testers judgements?
SY, what do you think about Robert T. Hodgson´s result in comparing wine testers judgements?
john curl said:
Hi John,
DA was modeled by Pease and others as a group of R-C's in parallel with the main capacitor. This small-signal model would cause small frequency response and phase response errors, what I refer to as linear distortions. In the time domain, they would cause analogous errors in integrators, and that is why there was so much study of DA back when analog computers were being built and used.
Note that some of the time constants in those DA parasitics are very long.
The process described thusfar, by itself, will not cause new frequency spectra to be created by a nonlinear process.
Whether or not DA is itself non-linear is a matter of whether the elements in the small signal model of DA change appreciably with voltage and current conditions on the capacitor. This I am not sure of. I gather that Syn08 thinks the answer to this is yes, if I have understood him correctly.
In any case, if DA is nonlinear, then it should make an appearance in THD measurements under the right conditions or under SMPTE IM measurements under the right conditions.
Cheers,
Bob
Bob Cordell said:
Exactly. Bob Pease linear approach and model holds only as a small signal approximation. This does not mean that DA is anywhere linear, and does not mean that a sensitive differential method as DiffMaker could not reveal the inherent nonlinearities.
I maintain that DiffMaker cannot reveal any linear distortions in the sense of e.g. phase response errors.
Hi Jan,
I just finished your interview with Erno Borbely published in AudioXpress. I found his reason for refusing to participate in these discussions to be particularly interesting.
I just finished your interview with Erno Borbely published in AudioXpress. I found his reason for refusing to participate in these discussions to be particularly interesting.
Oh come on, Alan ... Don't be so hard on John, I'm sure he's right.
I just realized a lot of people in this thread probably didn't see this post by John C in my hearing thread, and I thought this was just a great post ... so I hope you don't mind John.
I just realized a lot of people in this thread probably didn't see this post by John C in my hearing thread, and I thought this was just a great post ... so I hope you don't mind John.
john curl said:
syn08 said:
I have used those caps for over 20 years. The only problem I have with them is that are readily available and inexpensive.
Bob, the Pease model was developed in the '50's by others. It is a MODEL, and not completely correct. There is SOME non-linearity in DA, but it is usually a LINEAR DEVIATION (like phase shift) from the IDEAL signal. That is why the DA difference essentially sounds like the original music, with EQ and sometimes added non-linear distortion.
Many here are not even up to the engineering level that I reached 25 years ago. Please get to that level, before judging me or my efforts.
Many here are not even up to the engineering level that I reached 25 years ago. Please get to that level, before judging me or my efforts.
Johnloudb said:
For me the Maquee in Wardour street was a regular tinitus invoker back in the late 1970's. The Clash at the roundhouse in Camden, the Buszcocks, and Eddie and the Hot rods were the main culprits. Outdoor events such as Reading were never a problem.
anatech said:
Chris,
You post is well-intentioned, but it misses on a number of points. In an effort to be brief (the boys are napping, but will be waking at any moment) I'll try to answer your post, knowing that it will be incomplete at best.
--I, for one, am not afraid of measurements. I own--and use--an HP339A. I'd like to get an Audio Precision, but can't justify the money. Note that John has mentioned measurements many, many times, yet he's clearly in the listening camp.
--'Most audio engineers' don't post here. Just John, Nelson, and Charles (rarely, and I can't say that I blame him). Instead, we get people who might be called advanced hobby designers and folks who are just starting and people who are at various intermediate stages. But don't confuse the offensive and arrogant posts by some members here with what might be said by 'most audio engineers.'
--And how, pray tell, do you think reviewers come up with 'audiophile approved' ideas since they don't read catalogs? This deserves a much longer, dedicated post, but I don't have the time. The short answer is obviously that they listen to a component, find that it sounds good, then seek a quick, layman's answer as to why. After you've listened to ten components that had better detail and high end, eight of which used polystyrene caps, unless you're brain dead, you begin to see a trend. Are polystyrene caps a guarantee of good sound quality by themselves? No. Is a single ceramic cap (for example) the end of the world as we know it? No. To absolutists who insist that measurements are everything, a single opamp/ceramic cap/tantalum cap/etc. in the signal path should degrade the sound to unlistenability. Since that clearly doesn't happen, in their view the whole question of, say, opamps in the signal path becomes bogus; a lie promulgated by gurus preaching to gullible followers who then mindlessly repeat the words they were taught by their masters. Nowhere in their worldview is there room for the idea that the sound is degraded by small degrees, not by zero-or-one binary choices--one opamp in the mixing board at the studio and the sound is utterly destroyed. That all-or-nothing mentality is part of the problem.
--There are two main problems with audio reviews. First is the fact that print readership is declining across the board. (When I say across the board, I mean everybody, in all fields. It's true in science fiction, general interest, cooking, etc. All print media are hurting. Partly due to a shift to the web, but mostly due to the fact that people simply don't read as much any more.) With the drop in readership comes a drop in revenue. The drop in revenue means that you have less money to hire good writers. The not-so-good writers are more prone to write formulaic reviews. Now, if you have a solution for this problem, I suggest you speak up quickly. A lot of editors I know would love to hear from you.
The second problem is that reviewing itself is a difficult process. It's easy to criticize a given review and smugly think you could do a better job. It's another entirely to write your fiftieth review without falling back on something that you or someone else have said before...the dreaded cliche. Try it before you say anything further along these lines.
(My personal tooth-grinder is the "if you're thinking about buying an amp in this price range, you owe it to yourself to listen to the Megawhooziz 2000!")
--My checklist on Andy's post (and syn08's) is part of a larger project I've been working on. You may have run across my phrase 'science as religion' once or twice. There's more to this than meets the eye. I'm in the process of studying the interaction between the "Measurements-are-everything" folks and the folks who listen. My use of the word religion is not an accident. At first blush, you might find yourself swayed by the idea that John (as an example) is a cult leader and he has somehow gained control of the minds of impressionable people who need to be deprogrammed and returned to rational mainstream society. This is the view explicitly promoted by syn08, amongst others. However, if you think about it, the reverse is actually a much better fit. The Measurements people are the religious Fundamentalists (if it's not in The [Text] Book, it can't be true) and the listeners are more like those excitedly building on something like Darwin's theory of evolution, which the Fundamentalists find find inherently threatening. As in any healthy scientific endeavor, there are a lot of ideas being tossed about, from magic dots to cable direction to the sound of various metals. In a purely Darwinian sense, some of these ideas survive the winnowing because they work. Others fall by the wayside because they didn't (Bear mentioned the Tice clock, my favorite was the VPI Brick [which made rational sense...but no sonic difference]). The Fundamentalists see this outpouring of ideas--some fairly easy to grasp, like caps, others more difficult, like magic dots--as chaos, and they instinctively recoil from the inherent messiness of it all.
The analogy goes both wide and deep--I've been working on this for well over a year--and the parallels are astonishingly accurate. The Fundamentalists even use some of the same phrases in their arguments that you find in Creationist arguments against evolution. It would take waaaay too many words to completely lay out my case. I have yet to decide where to take this, but it's not impossible that I may work it into a story. We'll have to see.
Grey
A perhaps relevant point of reference is a paper presented at the 1998 IEEE Frequency Control Symposium, "PM Noise Generated by Noisy Components", where the authors compared reactive fluctuations in capacitors and inductors using phase noise measurements. Mylar and NPO demonstrated low intrinsic reactive fluctuations as compared to mica and ceramic disc. Their work was referenced at 5MHz rather than audio, however. Also of note is their verification that LCR circuits generate more noise (reactance fluctuations) at resonance by a factor of Q squared.
I've been working on a preamp for awhile now. And, you really do have to listen. Measurements don't tell all, by any means. Meters don't have ears ... Hello!!
john curl said:
One of my first jobs (1968) after graduating was working on (and improving) an analog computer used for simulation in a UK nuclear power plant. I never got inside the exciting reactor bits, bit did learn a LOT about what was necessary in a cap to produce linear integrators. Nor was this a big secret, caps were well understood in terms of linearity, DA, etc
Of course the PDP-8 and Dartmouth Basic killed this stone dead ....
Some 25 years later I was using the same knowledge in designing highly linear 2KV ramps for SCR dv/dt testing.
Audio really is a bit of a backwater .... 😉 😉
syn08 said:
I believe Diffmaker does let linear distortions through, if we agree that for instance freq response changes are a form of linear distortion. Some of that will come through after the subtraction.
I am not an expert on DA, but what puzzles me is that I hear a residu in the Z5U difference track that sounds pretty much like the original track. I have written to Bill Wazlo to ask what he thinks of it, because in the polyprop difference test there is no signal audible whatsoever, as if this was a perfect cap. I realize that points to a DA issue but why would it sound like the original?
Jan Didden
john curl said:
My post was not a judgment of you or your talents. I did not say that Pease was first, and I deliberately used the word model. Your insinuation that I am not up to the level you reached 25 years ago is a childish personal attack, but typical of your approach to others whose philosophy you don't share. It is truly a shame that a talanted person like yourself is so threatened by other smart people chiming in on an issue.
Bob
- Status
- Not open for further replies.