Wow, cap type makes a difference

[Toney]

I hear tonal differences that are quite obvious to me with same value different manufacturer caps.
Eventually reading all the traditional replies regarding values differences with manufacturers etc, I tried a couple of very simple 1 or 2  cap builds with a variety of Mylar caps matched the best I could find with my cap meter. No significant value difference but audible tonal change.
...
Really don't care if people say its impossible because I can hear it.

I fully, completely believe you. You absolutely, positively DO hear it.

But have you ever set up a test where neither you nor the person changing the caps didn't know which one you were listening to at any moment?  

Yes, yes, I know expectations alter perceptions.
In my case I was the only participant and like most musicians I trust my ears. 30+ years of playing guitar. What can I tell you?
I went with all this originally. I overcame what my ears where showing me. Ok, must be those value differences cap to cap I figured. I didn't really buy it based on what I was hearing Vs the general experience you pick up swapping out different value caps in builds over the years, but I went with it.
I bought a cap meter and, well , it was interesting seeing the variations but they were not so great with the modern Mylar types.
Once I had matched caps I socked and did it again with all of the various cap types I had - mostly different Mylar types: same results (to me).
It would be like trying to tell me, oh there are no, we have no parameters to measure any audible difference amp to amp from this fine batch here. They are all the same, it must be your perception.  
I would like to add the results surprised me a little. I thought not much of the el-cheapo Dick Smith Mylars but they really sounded good in my test circuits. The Panas were no surprise. The yellow box type sounded kinda crumby in the boost and fuzz I used to test.
Quite obvious differences again using different Cap types (matched).
Really my possible preconceptions and the results where different. I would not have expected the Dick Smith greenies to be much chop, but they sounded great and audibly different to other Mylars. The Panas also where consistently good. Big surprise on the yellow metal film box type, just average with that set up (I was actually expecting theses to sound good, kind of annoyed they did not as they are so convenient)
Mostly the differences where as discussed here for years, the Panas sounded good in most situations, Ceramics sounded "gritty" etc. The thing I learned though is YES, definitely there are tonal differences and yes those tonal differences are still obvious when caps are carefully matched.
Based on this eventually bought a biggish supply of the Panasonics once I heard they were being discontinued. I use them in most builds now except for the occasional fuzz which "likes" ceramics.
I don't need to believe or disbelieve what I hear. I just hear it. I am not interested in cork sniffery or any voodoo just going for practical usable circuits. This is music electronics!  I

[candidate]

These days I qualify components in terms of "potential to produce audible differences given the proper circumstances".

The green caps' and ceramic caps' readings are off the charts!

[R.G.]

Yes, yes, I know expectations alter perceptions.
In my case I was the only participant and like most musicians I trust my ears. 30+ years of playing guitar. What can I tell you?

That'll do fine. 

[Michael Weidenauer]

I also agree that perception may fool you, but I also hear differences.

Some years ago I did lots of BluesBreaker clones - all with 1% resistors and measured and selected caps, so the tolerance between unists should be low.
My test was with 7 different versions (1. cer-disc , 2. cer-resistor-look(like the orig.), 3. cheap MKT-box, 4. Wima MKT, 5. orange-drops 600V, 6. Mallory 150s and 7. old styroflex and paper-oil caps (the high end-stuff )) all with the same circuit (modded for more treble and bass),  fixed 1%resistors instead of gain and tone controls, all pedals in a chain.

The ceramic disc versions sounded fuzzier with the highest intermodulation distortion, the average MKTs sounded fatter and clearer but lacked some treble bite, orange drop type MKPs sounded clear but cold, big MKTs (those 600V types used in tube amps) smooth and clear. Old styroflex and paper-oil caps were my favourites (still are) sounding warm and clear with really low intermodulation.
There may be debates in the middle field, but I think the differences between the ceramics and the styroflex could be heard by every player.

P.S.:
80% of the caps I use are recyled caps from old TVs, telephones, VRCs,...
Really nice caps (and other parts) for free...

P.P.S.:
There is a cap type you really shouldn't use: tantalums!
Not for technical reasons, but for humanity.
If you've ever seen a documentation what the need for coltan does to the people in Congo and the environment there, I think you'll agree.
http://en.wikipedia.org/wiki/Coltan

[R.G.]

I also agree that perception may fool you, but I also hear differences.

Some years ago I did lots of BluesBreaker clones - all with 1% resistors and measured and selected caps, so the tolerance between unists should be low.
My test was with 7 different versions ...
There may be debates in the middle field, but I think the differences between the ceramics and the styroflex could be heard by every player.

I fully, completely believe you. You absolutely, positively DO hear it.

But have you ever set up a test where neither you nor the person changing the caps didn't know which one you were listening to at any moment?

[Gus]

found the CB articles
http://www.scribd.com/doc/2610442/Capacitor-Sound

Note the voltage difference between plates, dielectric and construction.

[Michael Weidenauer]

I also agree that perception may fool you, but I also hear differences.

Some years ago I did lots of BluesBreaker clones - all with 1% resistors and measured and selected caps, so the tolerance between unists should be low.
My test was with 7 different versions ...
There may be debates in the middle field, but I think the differences between the ceramics and the styroflex could be heard by every player.

I fully, completely believe you. You absolutely, positively DO hear it.

But have you ever set up a test where neither you nor the person changing the caps didn't know which one you were listening to at any moment?

I've got an loop pedal that alternately switches two loops, without power-supply there is no indicator-led so I don't know which is on.
I did blind tests with that setup(one BB in each loop), randomly switching lots of times so I didn't know which which on, then playing and veryfying by pressing the bypass of the pedal that I had guessed.

As I said: When comparing several normal MKTs and MKPs you mostly can't  tell, but with cheap ceramics vs. big high quality film-caps there is a noticable difference - the intermodulation distortion was really high with bad ceramics and gets lower the better the caps.

[R.G.]

I've got an loop pedal that alternately switches two loops, without power-supply there is no indicator-led so I don't know which is on.
I did blind tests with that setup(one BB in each loop), randomly switching lots of times so I didn't know which which on, then playing and veryfying by pressing the bypass of the pedal that I had guessed.

I fully, completely believe you. You absolutely, positively DO hear it.

But have you ever set up a test where neither you nor the person changing the caps didn't know which one you were listening to at any moment?

As I said: When comparing several normal MKTs and MKPs you mostly can't  tell, but with cheap ceramics vs. big high quality film-caps there is a noticable difference - the intermodulation distortion was really high with bad ceramics and gets lower the better the caps.

There is no question that caps are diffferent inside. Otherwise, there would not be different kinds.

However, it is remarkably difficult to conduct a fair listening test by oneself.

It is remarkably difficult to conduct a fair listening test by oneself.

[Ronsonic]

However, it is remarkably difficult to conduct a fair listening test by oneself.

Nothing at all difficult about it. Compare and contrast and go with the results. Don't over-complicate it or believe your theories over your ears. Don't strain to hear a thing that might not be there and be comfortable saying "it don't matter" and then select on other criteria. There is no reason to believe that one component should be "better" than the other so there's no reason for the test listener to be biased. I really don't care whether the "chiclet" or the "box cap" wins, for example. I'll happily pull either from a drawer and solder it into place. I don't know that impartial = fair, but it's close enough. Not to wander too far off the reservation, but what is "fair" to a part? It's wire and plastic, "fair" has got nothing to do with it.

If you insist that only blind tests are worthwhile, and I'm not sure why that would be, it's easy enough to use an unmarked rotary switch or a stomp type switch to select components under test in a way that prevents one from knowing which is which until the test is done. Hmmm, in fact, I probably should have something like that handy and ready to go for auditioning parts. More convenient than the back&forth.

[R.G.]

Not to wander too far off the reservation, but what is "fair" to a part? It's wire and plastic, "fair" has got nothing to do with it.

Who said anything about fair to the parts? Parts is parts, they're inanimate objects to be used. I meant fair to YOU. Presumably if you conduct a test, you're doing it for a reason. There can be many reasons, including things like reaffirming what you already "know" to be true, whether it fits the facts or not, or aligning yourself with the guru of the moment.

If you insist that only blind tests are worthwhile, and I'm not sure why that would be,

I don't insist that only blind tests are worthwhile. However, I do insist that tests which cannot be replicated by impartial third parties are worthless. They may be fun, they may make you feel good, but the data they turn out is inherently suspect.

it's easy enough to use an unmarked rotary switch or a stomp type switch to select components under test in a way that prevents one from knowing which is which until the test is done. Hmmm, in fact, I probably should have something like that handy and ready to go for auditioning parts. More convenient than the back&forth.

I use switching between parts and circuits for testing all the time. Sometimes the differences are massive and sometimes there is no possible preexisting bias. 

However, for any test where the differences are subtle, it is remarkably difficult to conduct a test which is fair to you (i.e. turns out data which is reproduceable by a disinterested observer, or can be measured by an impartial instrument in numbers) if you know which part/circuit/etc. is being tested, it's hard to be impartial. Sometimes even when you're TRYING to not pick the one you want before the test, you can't do it.

I suggest you google "Clever Hans" and read about Hans the horse that did arithmetic and the man who unmasked what was really going on. And what happened to him after he learned the real facts.

[Greg_G]

It is remarkably difficult to conduct a fair listening test by oneself.

Fair = Objective (in this instance).

Objective :
1. Uninfluenced by emotions or personal prejudices.
2. Based on observable phenomena; presented factually: an objective appraisal.

Our mind likes to convince us that we're being objective, but it's been proven many times that we rarely are objective... and we're simply fooling ourselves.

[Lurco]

You can fool some people some times - but you can fool yourself all the time!

[Gus]

Did anyone look at the link I posted? 

[DougH]

I perused it, Gus, but we talked about that on the phone.

I subbed a ceramic treble tone stack cap for a mica in one other amp I have. I recorded before and after sound clips without moving the mic, changing recording parameters etc. To be honest, I did not hear the differences I thought I heard in the room. There was a difference, but in the recording the mica sounded smoother than the cer whereas in the room it was opposite. Go figure. I will go back and re-listen when my hearing memory has run out.

With another couple amps I used to always think they sounded rough when I first plugged into them after a while (after my hearing memory of them died). Since swapping the micas for the cheapie cers they don't sound this way anymore. They just sound good and inspire me to play. Internal bias or hearing? Don't know and don't care. I don't think about caps when I'm playing, I just think about playing. Not peddling expensive mojo (the opposite, actually) just relating personal experience from something I experimented with. And this data is worth everything you paid for it.

I believe Gus's article shows there may be some valid reasons for this. Might make the thread more interesting if people actually read it and considered it. Or not- we can just opinionate too, but that's not very interesting.

[Ronsonic]

Who said anything about fair to the parts? Parts is parts, they're inanimate objects to be used. I meant fair to YOU. Presumably if you conduct a test, you're doing it for a reason. There can be many reasons, including things like reaffirming what you already "know" to be true, whether it fits the facts or not, or aligning yourself with the guru of the moment.

Okay, I was being a bit frivolous with the fair to the parts thing. But fair to myself is to my own satisfaction. I really mean it when I say I don't care which cap wins. Really. Strongly prefer if it not be some unobtainable striped NOS thing from 1963, but for cases like that I just don't even audition them. That's my biggest prejudice, the "In Stock" sign on the Mouser part description.

If you insist that only blind tests are worthwhile, and I'm not sure why that would be,

I don't insist that only blind tests are worthwhile. However, I do insist that tests which cannot be replicated by impartial third parties are worthless. They may be fun, they may make you feel good, but the data they turn out is inherently suspect. 

Impartial third parties are a tricky thing. You cannot know what they listen for, what they notice or don't.  Cannot tell you how many times I've sat with a client who was explaining what he didn't like about the way his amp sounded and it took some back and forth to explain / demonstrate exactly what we were each talking about and that's with a free flow of conversation and a guitar. Some things that are obvious to one listener can be unnoticed by another even if he does clearly hear it once it is pointed out. I've seen that happen a lot, and going both ways. Sorta like the lines on the old Trinitron, you don't see them until they're pointed out. You won't get to that in most tests where everyone is silent. You might consider that more valid, but since I've got to live with my stuff and they don't I'll go with what impresses me.

The real test is when someone else gets ahold of whatever it is you're making and tries it completely independently. Yeah, are people happy when they buy it. Which is another whole can of worms. My thinking is that this is so much a matter of taste all you can do is serve your own sensibilities with an acknowledgment of where you know you differ from the norm. A cook preparing dishes he dislikes can do passable work following a recipe but has no chance of producing anything brilliant. "This is what they'll like" rarely exceeds mediocrity.

Some folks can do it. Excellent commercial designers (we seem to know a few around here) can work in pedal designs outside their own uses and tastes and get fine products to people. Others don't much have that skill - for them it's a matter of serving themselves and finding like-minded others who agree.

it's easy enough to use an unmarked rotary switch or a stomp type switch to select components under test in a way that prevents one from knowing which is which until the test is done. Hmmm, in fact, I probably should have something like that handy and ready to go for auditioning parts. More convenient than the back&forth.

I use switching between parts and circuits for testing all the time. Sometimes the differences are massive and sometimes there is no possible preexisting bias. 

However, for any test where the differences are subtle, it is remarkably difficult to conduct a test which is fair to you (i.e. turns out data which is reproduceable by a disinterested observer, or can be measured by an impartial instrument in numbers) if you know which part/circuit/etc. is being tested, it's hard to be impartial. Sometimes even when you're TRYING to not pick the one you want before the test, you can't do it.

Given the lousy memory for sound that we humans have I'm impressed when the same subject can repeat a test, much less others. Oh, is it nice when measurements apply. Love it when that happens. Just don't have the instrumentation to tell me if that midrange isn't a bit too clotted when overdriven. So it's all that painful subjective stuff that gets the listening tests. There again, I don't care which cap works because my opinions and prejudices, good, bad and indifferent are focussed on the result. I have absolutely no theoretical framework on which to hang a bias about which cap is "better." Nope. Absolutely no idea whether one or another should be better there. Got a lot of theories and even some math to select a value. Even have some ideas about what to expect from a cap sonically, but that's why I try different ones because I know I don't know enough to make a call.

I suggest you google "Clever Hans" and read about Hans the horse that did arithmetic and the man who unmasked what was really going on. And what happened to him after he learned the real facts.

A true classic. Well aware of Hans and his amazing skills. Thanks for an interesting discussion of this and food for thought. Yeah, I've got opinions as stated, but am happy to review them.

[Eb7+9]

Did anyone look at the link I posted?  

partly, seems like some questionable/typical stuff going on in some of the setups - I'll say why if anybody's interested ...
btw, is there a way to host a copy of the document somewhere else so it's easier to download ??

Doug, the silver mica (250pF treble) cap was touted by Dan Torres when he was first starting out (same with Weber)
it stayed in my amp for a few hours ...
that idea served to question that guy's ears and intentions by very many people, similar to his other "mods" ...
same with Weber ...

George and Leo had it right the first time around

[Mark Hammer]

One of the chief limitations in such comparisons is that we do much better on materials we can verbally describe in detail.  It is a trivial matter for me to assess the accuracy of my recall of what I have seen by describing the elements of a static image, and then some time later describing what I think I remember and comparing that listing against what I said or wrote down at the time at the time.  When it comes to sound (which is NOT static), recalling the tone of the original as "toothy", "shimmery", "warm", or some other group of equivalently nonspecific descriptors affords too much room for NOT retaining the details, making it very difficult for anyone to compare the before and after with any degree of reliability.

[blueduck577]

This was just posted over at harmony central.  It's a video of a guy subbing in caps of different materials into his guitar's tone circuit:
http://www.youtube.com/watch?v=92G-jw4TqS4

I think it's interesting, but it makes me wonder:  are the differences between the different materials actually due to the materials or perhaps due to component tolerances and variance in the actual values?

[jrod]

It always been interesting to me how the "best" sounding parts are usually the coolest looking ones and/or the hardest to find!

[DougH]

Yeah, I know. These:


look so much cooler than these:
and are so hard to find...