John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Sy,

No one else ever has, either, assuming levels are set correctly.

This is obviously untrue. Even if I had no opposing episodes (which if I find the reference become "observed fact"), to claim "no-one ever" is unscientific.

If you claimed "no-one has published any results in peer reviewed journals that I am aware of" this would be defensible, however the way you formulated your claim falsifies it instantly.

I do, as it so happens, remember for example an article from either the IRT (IRT :: Home) or the VDT (http://www.tonmeister.de) Journal quite a few years back where they evaluated AD/DA Studio gear for use in the Federal Radio / TV Stations in Germany (this was around Y2K).

They found significant audible differences in their testing and settled on gear by an obscure German Company which used my "I love to hate them" Cirrus Logic converters (that is why I remember it). I will try to dig out the reference for this, but it will take time, the actual stuff may be in storage.

On the other hand it is interesting to note that the old non-oversampling Sony PCM-F1 Adaptor was tested repeatedly and found "transparent".

So as we often find in Audio, we have a range of differing and conflicting testimonies and not all of the ones that conflict with the orthodox view come from "golden-ears" or uncontrolled test.

Ciao T

PS, for those who manage German, the VDT and IRT sites are goldmines of information, but like the JAES and Stereophile, take with a grain of salt and apply some common sense...
 
Of course, T, an 'approved' double blind test will find the PCM-F1 virtually 'transparent'. That is the point.

Maybe it is, when levels are set correctly. When not, apparently the differences can be heard by Lipshitz and Vanderkooy, but not Tiefenbrun.:D

When you or anyone else pony up a reference or (mirable dictu) do an actual controlled listening test yourself, people can take your indignation a bit more seriously. Sneering and pseudo-philosophical discourses do not equal evidence.
 
Maybe it is, when levels are set correctly. When not, apparently the differences can be heard by Lipshitz and Vanderkooy,

According to Lipshitz and Vanderkooy everything sounds same when levels are set correctly. Their tests are pointless - to me. Even at levels set correctly the differences in A-B tests are clearly audible, in case we care about set-up, listening room, listening positions, high quality source material, amplifiers used, cables used, design of switch box etc. etc.

Lipshitz and Vanderkooy results are valid only and only for the tests they performed. No generalization is possible.
 
Well, T and PMA, looks like we have an 'unbeliever' in our midst. I would prefer to go forward with discussing 'unproven' differences, whether measurable or not, since a whole bunch of us seem to 'hear things' that can't be 'proven' to be there. I wish that it was easier to do so, at least on this thread. That is the whole point of my contributing here.
 
Last edited:
This reminds me of something that I hope to convey, here. I am always being asked, if not 'commanded' to participate in 'double blind' listening tests, even though I have tried them, previously, and found that I did 'worse' with them than several others. The ABX test, if fundamentally correct, without any chance of error, even the error of missing differences that really are there, precludes any more participation on this thread or any new development of audio electronics. Therefore, you are wasting your time reading my efforts in 'improving' audio electronics, as it must all sound the same, if certain rather minimal standards are met.
This is why I decided that ABX testing was not for me. Either I am 'crazy' or the ABX test makes different things sound virtually the same, during the test, and yet different in open listening, or even A-B comparisons. It is for you to chose.
 
Today I had an interesting feedback from my customer. Not audio, but measuring systems. The issue was input voltage divider (R//C) of the data acquisition system - transient recorder with fiber optic signal transmission. The problem was that step response has had unexpectedly wide settling error band. Not a typical overshoot, but rather long tail settling. We have been trying to find the real reason for years. Today he called me he has finally solved the problem. He has exchanged the polyester capacitor in the input compensated divider for the polypropylene cap - and the long tail settling disappeared. The issues are same as in audio ...
 
Today I had an interesting feedback from my customer. Not audio, but measuring systems. The issue was input voltage divider (R//C) of the data acquisition system - transient recorder with fiber optic signal transmission. The problem was that step response has had unexpectedly wide settling error band. Not a typical overshoot, but rather long tail settling. We have been trying to find the real reason for years. Today he called me he has finally solved the problem. He has exchanged the polyester capacitor in the input compensated divider for the polypropylene cap - and the long tail settling disappeared. The issues are same as in audio ...

Exactly. When you measure something that ain't right you try to figure out stuff that will make it work. Sometimes the solution will be provided by Princess Serendipity.

The issue with audio is that our measurement suite might be too limited. Otherwise, things in the transmission train seem pretty well covered.

vac
 
that's a lame example - instrumentation designers have known of DA error relative magnitude between dielectrics for a long human lifetime

BSTJ has "3-D" plots of polyester dielectric cap errors with frequency, temperature from characterizing them for telephone filters/eq

"real", "conventional" engineers do appreciate device differences exist, are quantifiable, give measureably different circuit results

in fact these parts selection hints are one of the reasons I follow the thread at all

but I certainly don't believe the "explainations" often offered when they are easily shown to be of absurd low magnitudes releative the the other audio reproduction errors
 
Of course, PMA, it is OBVIOUS in hindsight. '-) I had a brief fantasy of being a consultant for a firm with the same problem, and how 'obvious' it would have been to check for DA effects in any caps. Like you, I sometimes do this sort of thing as part of making a living. It is interesting that you call a 'tail' was described to me as 'echo' once in WW or HFN and I never heard the end of it, but it was obvious in the test that Walt Jung and Scott Wurcer developed to test DA in caps, and I used for Walt's and my paper on the subject.
 
Today I had an interesting feedback from my customer. Not audio, but measuring systems. The issue was input voltage divider (R//C) of the data acquisition system - transient recorder with fiber optic signal transmission. The problem was that step response has had unexpectedly wide settling error band. Not a typical overshoot, but rather long tail settling. We have been trying to find the real reason for years. Today he called me he has finally solved the problem. He has exchanged the polyester capacitor in the input compensated divider for the polypropylene cap - and the long tail settling disappeared. The issues are same as in audio ...

I guess he never read Bob Pease's article, very old news.

EDIT - Does we include you, you certainly know this?
 
Last edited:
Either I am 'crazy' or the ABX test makes different things sound virtually the same, during the test, and yet different in open listening, or even A-B comparisons. It is for you to chose.

Many other possibilities. And there's nothing magic about ABX, there are a lot of good ways of doing "ears only" tests. But we've been through that over and over. Some of us trust our ears, others rely on eyes and preconception.

“It is difficult to get a man to understand something when his salary depends on his not understanding it.” Upton Sinclair.
 
Don't worry PMA, Scott doesn't believe that it is important for 'audio' in any case. I beg to differ, as does Walt Jung. In fact, Walt sent Lipshitz et al the very WORST tantalums that he had determined by listening, and they could not sort them in an ABX test. What then, is the point? If anyone does not believe me, ask Walt Jung, yourself.
 
Yes, as SY implies, I am paid for finding 'differences' rather than maintaining that audio is virtually perfect...

Of course no-one maintains that "audio" is perfect, but hey, if that's what you need for an argument, go for it.

I assume that Walt Jung didn't sort those caps by ear alone either.:D

What then, is the point?

The point is that there are many things which cause audible differences. There are many things that don't. The way to tell one from the other is by ear, not by eye.
 
Status
Not open for further replies.