Sound Quality Vs. Measurements

[fas42]

Originally Posted by dvv

So, I need not go outside my own house to verify that any topology, when applied knowingly, will produce good results, more often than not better than many a device slapped together according to current fashon.

Zigactly, too ...

largest value used should be polycarbonate. It seems nothing else on this planet can match its extraordinary speed.

I always get a bit twitchy when people use the word "speed" with audio stuff ... to me, they really mean low distortion when they say this. Good sound has tremendous drive, PrAT is another term for this, and it only really falls into place when the level of audible distortion is sufficiently low.

To me, people are confused about distortion: there's obvious distortion - a cheap radio overloading; compression, when an amp sounds like it's got no more to give; edgy, hifi style sound; or flat, grey, poorly detailed sound. These are all forms of distortion, all are not true to the source, what the recording has encoded in it. So a system that lacks "speed" is a distorting system, it's not reproducing what's on the recording correctly ...

Frank

[dvv]

Originally Posted by vacuphile

I can't see how a polystyreen cap with a low loss factor could benefit by paralleling it with e.g. a polycarbonate one with much higher dielectric losses. (of course using half values). Just doesn't make sense to me.

Vac, don't take it at face value, that was just an example. I suppose half the fun of tweaking is in finding the just right combination for you.

And the Good Lord knows you have your choice of materials.

Oh, BTW, I never said polystyrene - I said poyester, polyproypelene and polycarbonate.

Personally, when I need small values in the pF range, whenever possiblem, I go for silver mica, or, better yet but often hard to find, multilayer ceramic caps by Siemens. Damn expensive, but second to none.

[dvv]

Originally Posted by fas42

Zigactly, too ...


I always get a bit twitchy when people use the word "speed" with audio stuff ... to me, they really mean low distortion when they say this. Good sound has tremendous drive, PrAT is another term for this, and it only really falls into place when the level of audible distortion is sufficiently low.

To me, people are confused about distortion: there's obvious distortion - a cheap radio overloading; compression, when an amp sounds like it's got no more to give; edgy, hifi style sound; or flat, grey, poorly detailed sound. These are all forms of distortion, all are not true to the source, what the recording has encoded in it. So a system that lacks "speed" is a distorting system, it's not reproducing what's on the recording correctly ...

Frank

I know what you mean, Frank, but there is some merit in talking about speed.

Here's an example - what use your amp's capability to slew at say 40 V/uS, and much more important, say 10 A/uS, if your filter caps will knock this down to say half the nominal value?

This may well be more of a technical point of view, because even if they do limit the amp's speed, the final speed may well be at or above what we actually need in real life, but I for one would be seriously stressed if I busted my behind ro make an amp fast only to have the caps cut it down.

I have said this before, and I'll say it again - the ONE AND ONLY capacitor spec sheet which even mentioned this speed was that of Siemens Sikorel professional series, they clearly stated that it did 100 V/uS and 8A/uS. You don't really want to know their price. It's all EPCOS these days, as Siemens and Matsushita established EPCOS as a joint capital manufacturing company some years ago.

And while speed is not everything, and maths may show we only need like 5 V/uS, I agree with John on this point 100%, we should, just in case, aim for something like say 40 V/uS. This necessarily forces us to consider the caps in that light as well.

[fas42]

And, perfect timing. Courtesy of another audio forum I've found a YouTube clip, recently posted, which has captured extremely well the key failing of current setups. This clip is intended to demonstrate how impressive the "best" high end can be but mainly serves to do other ...

The clip is Hi-End System vs percussion band - YouTube. The latter plays first, is recorded and then you hear immediate playback. First problem is that the system playback is too soft in comparison, so you need to turn the volume up, to match levels. Which more acutely reveals that key problem I referred to ... and that is, poor rendition of the treble. The recording as compared to the live version badly lacks a sense of realistic dynamics, essentially because the treble content is not "right" - I have no trouble picking this on my ordinary PC monitors, which have very average treble by most standards. This is something that I've always focused on, to me this is absolutely key to getting audio to be "correct"!

I've captured the HD version of the clip, which has excellent sound by YouTube standards; and intend to do some comparative analysis of the two sections of the audio track, see what I come up with ...

Frank

[dvv]

The difference is obvious, I'd say.

The reproduced version has lost a lot of sheen and bite and overall sounds diluted.

I checked and rechecked the above comment by several runs over my speakers (Luxman C-03 preamp > Marantz 170 DC power amp > speakers) and headphones (HP PC > Luxman C-03 preamp > Sennheiser HD 598 cans).

[a.wayne]

Any explanation on what John is suggesting below, how to confirm "leakage "...?

Originally Posted by john curl

Finally, some common sense about ground leakage! That is another 'tweak' that we did with the Parasound JC-1 when we designed it. As it has multiple transformers, we oriented each transformer to be lowest leakage when the power cable was connected normally. This is different from the HCA3500, that did NOT have properly oriented transformers. It was one of the first things that we changed.
Just another 'tweak' to get best performance. Interestingly enough, Hi Fi books from the '50's used to suggest the same thing. It was easier then, with just two leads, and no 3'rd wire ground, at the time.

[fas42]

Originally Posted by dvv

The difference is obvious, I'd say.

The reproduced version has lost a lot of sheen and bite and overall sounds diluted.

Of course. But, to me, to my ears, this encapsulates perfectly where normal audio gets it wrong; I listen to an expensive system, in the flesh, and they sound just like that system does via the clip, or worse. They should sound like the live version that comes up first; you have no problem picking up the much higher quality, realism of the musicians performing -- and this is a YouTube clip, for heavens sake! Supposedly a cheap and nasty way of putting up snippets of sound, but it still gets the message through!

So what's the problem? Well, we're hearing a recording of system playback at close to live levels, and it's not working in spite of all the expensive machinery thrown at the "problem". It's not the recording that is being played by that system that's at fault, because we can hear what's on that through a set of headphones, PC monitors, being the first half of the clip! That leaves the playback mechanism as the guilty party, but why is that so ...?

This is precisely the problem that I've been working on for the last couple of decades, and it's not an easy one to knock over. You mention "loss of sheen and bite", and that to me is another way of saying distortion of the high frequency component of the playback. Some may say that this is a generic problem of speakers at decent volumes, and I say this is wrong. Over and over again I've dealt with systems sounding like the second half of the clip, the speaker version, which I then tweak to perform, sound like the "live" section, the first half. The speakers stay the same, the amp, the source, etc, are all retained; but all the issues that drag down the quality, create a "diluted" playback, are sorted out ...

Frank

[fas42]

dvv, I am surprised ... projects for electronic volume control are extremely thin on the ground; these were what I found on a first scan: MiniVol PGA2320 Volume Control - error404's Audio DIY Endeavours and a thread mentioning others, Looking for a PGA2320 PCB. All use processors to do the fiddly stuff, but the first one, MiniVol, seems to be well done - the processor is put to sleep when not needed - , people have had success with it.

Personally, I have no hangups about digital volume controls, they give me the sound that I'm after. In that live vs. playback clip, the "diluted" quality is exactly what I find mechanical, analogue, gives; hence I steer well away from it ...

Frank

[benb]

Originally Posted by a.wayne

Any explanation on what John is suggesting below, how to confirm "leakage "...?

Fig. 1 here shows the problem:
http://www.jensen-transformers.com/an/an004.pdf

Maybe you could ask John, but it appears to be the AC power line current from the capacitance of the transformer windings and core that goes to the chassis ground and then (in a modern 3-wire power connection) to the ground pin at the power plug. Also, some power entry modules have capacitors between hot and ground and between neutral and ground (especially when connected to a switching power supply), and a small amount of power line current flows through these capacitors as well. Since wire (specifically the ground wire in the power cable) has greater than zero resistance, this causes a small AC voltage on the chassis relative to the ground in the power plug, and also relative to the "ground" on the chassis of some other equipment, even when plugged into the same outlet. This usually makes it into the signal path as shown in the above file.

Balanced connections are in general less affected by such ground currents, as well as being more immune to other kinds of interference. I recall an article in Audio Amater a few decades (!) ago titled something like "Is it time for balanced interconnects in consumer audio equipment?" I thought back then it would be expensive, but considering all the "high end" crap with unbalanced gold plated RCA connectors and expensive boutique interconnects, balanced connections would solve a lot of problems at minimal expense.

Here's a lot more - he's selling his own brand of transformers and balanced input buffers as solutions, but he describes the problems very well:
http://www.jensen-transformers.com/a...%20seminar.pdf
http://www.aes.org/sections/pnw/pnwr...lock_pnw05.pdf

[a.wayne]

Thanks Benb .......