Now THAT is a sound card! :)

[willawake]

Re: Now THAT is a sound card!

[QUOTE=Logistics]Likely, something similar to these jokers(and yes people really buy into this stuff): http://www.belt.demon.co.uk/product/ref/ref.html

hahahaha it is possible to make big bucks with that crap

[gastorgrab]

Re: Now THAT is a sound card!

http://www.belt.demon.co.uk/product/ref/ref.html

Won't my foil helmet filter out most of that interference anyway?


.

[WeStSiDePLaYa]

Re: Now THAT is a sound card!

Why spend on a sound card like this? Use SPDIF out and get an external DAC, MUCH more potential.

[Logistics]

Re: Now THAT is a sound card!

It's what I have and regardless of all the excitement over things like external DAC's and the X-Fi, I don't believe the difference in quality could warrant the difference in price.

[999999999]

Re: Now THAT is a sound card!

You'll notice they used some of those film caps as bypass caps, which is a huge plus. Ceramics are known to cause audible distortion when used as bypass caps even in the power supply stage of an amplifier. This is nice, but the Nichicon VX is still just a general purpose cap. Can anyone tell which line those brown ones are from? And I like those normal type resistors they used. Could switch them out for 0.1% Cardas. RCA Jacks = Superior! ...but then I have no use for surround.

No, ceramic caps are not known to cause audible distortion as bypass, decoupling caps. Some people say they sound different though, because they are faster than film caps (more accurate except at temperature extremes). However, there is more to it, if dealing with an output stage between V+ and Gnd that is decoupled instead of rail-to-rail with a negative supply swing on the AC output, you will have less separation if decoupling direct to +- rails with a ceramic than if decoupling separately to +V-Gnd and -V-Gnd with ceramics. Even so, the difference here can be small and some even argue that an addt'l decoupling of +V to -V still has enough merit to outweight the detriments of cross-channel modulations.

The vast majority of the caps on that card are a waste, as the primary improvement would come from film coupling on the input and output instead of the 'lytics used.

Replacing the main power supply cap should only matter if driving a low Z load. That's like causing a problem and then trying to band-aid it (I realize, you did not write about this, mine is a summary reply).

Replacing the power cap to the output opamp has no effect whatsoever on the low frequency output, UNTIL the opamp is in suboptimal class A/B nonlinear mode and is already delivering more current than it should. In other words, don't ever allow a situation where it would matter, the input impedance to the next stage should completely obsolve the cap of any improvement.

If your DSP is not oscillating, there is no factor for replacing the cap to counter it. Capacitors are not magic, you can't just replace them then assume there is any real difference except for real measurable difference as in the coupling cap type ('lytic vs film or other type).

The X-Fi is definitely not worth the price difference in an absolute sense as it doesn't have higher sound quality except in some gaming scenarios. What it has is a sound processor, which is great for gaming and something any serious audiophile will disable as much as possible in non-gaming use, ASAP. If you have serious all-in-one card requirements to game, it might be a good compromise if the system's CPU is struggling.

A sound card chipset does not use a lot of current. Nor does an opamp output stage. The primary goals should be operating the output stage in class A, powering it linearly, decoupling it with film or ceramic nearby, and outputting into a very high impedance load instead of driving headphones directly.

To some people the easy answer here is a headamp, instead of trying to tweak an opamp into driving such a low Z load that it is far more distorted than the manufacturer spec'd. Capacitor improvements can help in areas but only so much and then the rest of the circuit has to be changed.

I own and have modded several Envy based cards. The output coupling cap conversion to film makes the largest difference by far. Anything else is far less important than what the next stage is that the sound card is driving (instead of headphones directly). Same goes for the X-Fi and most sound cards.

[Logistics]

Re: Now THAT is a sound card!

No, ceramic caps are not known to cause audible distortion as bypass, decoupling caps.

Walter Jung and Richard Marsh disagree with you.

Also, nice writeup. Yes, I'm forcing my Live! to drive a load a little lower than it is supposed to. Do you have any opinion as to what I should do about my output caps? They were originally 10uF's and I replaced them with 0.1uF generic films. It sounds alot better, but capacitor quality aside, what will likely be the difference in moving to a 10uF film?

I decided to use the original front channel to goof around and change the 10uF lytics to 1000uF and there really didn't seem to be any more bass or other changes, but the volume was much higher for a given setting. However, I am beginning to wonder if this is simply due to Windows audio mixing, because even though on my project system with the SB16, there was much more power behind the music and the bass was better, the low-frequency cut-off isn't necessarily better. I recently found out about that plugin for Winamp that allows you to do Kernel streaming and on my SbLive! system it sounds much better. I haven't tried it on the SB16 system, yet.

[999999999]

Re: Now THAT is a sound card!

Walter Jung and Richard Marsh disagree with you.

Link the article(s) or posts you refer to, as I believe you may've misinterpreted them. I've read their articles and AFAIK, their only comments regarding this were about preserving signal quality through coupling, not decoupling. Further many semiconductor companies clearly specify use of ceramic or tantalums for power rail decoupling on their analog chips.

The noise on power rails is unwanted, there is no purpose to try to make unwanted power rail noise somehow linear or unchanged- the whole point was to change, minimize it.

[Logistics]

Re: Now THAT is a sound card!

Jung has his own site up somewhere in which this article is done nicely in PDF format. I thought I had saved the URL, but can't find it now. Here's a site which still has it up in HTML. Look under "Ceramic Capacitor Tests."

http://www.reliablecapacitors.com/pickcap.htm


As far as manufacturers specifying ceramics and tantalums, don't you feel this is simply for practicality/convenience sake?

[999999999]

Re: Now THAT is a sound card!

Again I urge you to reread the articles, hints like "Data in the form of the THD vs. frequency" emerge which are telltale signs this is a comparison of coupling, not decoupling caps.

No it's not for convenience sake that manufacturer's spec tantalum and ceramic, they guarantee their parts meet specs in defined implementations and if they wanted to spec especially linear film caps, they could have as easily stated it instead.

[Logistics]

Re: Now THAT is a sound card!

...it seems pretty clear to me.

"Our feeling is that they should simply be avoided anywhere near an audio signal path and probably just avoided altogether for audio. For example, some listening tests have indicated that they can produce audible distortion when used as supply bypasses, let alone coupling!"

[999999999]

Re: Now THAT is a sound card!

That "probably just avoided altogether for audio" is ambiguous at best, then taken out of context. Testing shows the differences in cap type apply to coupling scenarios.

[999999999]

Re: Now THAT is a sound card!

Further, "some listening tests" can make all kinds of snake oil seem beneficial, there is much more importance placed on controls in listening tests like a blind ABX. Remember that respected IC manufacturers are not only suggesting use of ceramic and tantalum for decoupling, they are guaranteeing the performance of their parts in these scenarios. If all it took was to suggest using a different cap composition (up until a few uF at least with film caps since they'd start to get large and costly at that point) to improve performance, in this highly competitive market they would quickly do so.

Outside of a blanket generalization, there are some roles where a ceramic may not be suited due to the potential for microphonics, but at this point we aren't even differentiating between a disc or MLC ceramic, and the article quoted from, published in 1980, is certainly not considering, using modern MLC capacitors that didn't exist at the time.

[gonzo0815]

Re: Now THAT is a sound card!

Well, i think it is a well known fact, that MLCC are very very non linear regarding frequency VS capacitance.
Sure, you can avoid it by using better material, but even then this is not far better.
The reason, why they are recommended is simply the price, nothing else.
The really good materials, are not suited to make capacitance close to one uF

Mass manufacturers are always seeking a complete solution, but it must be cheap.
So they go with the chip, which could be implemented in the cheapest way, suiting the needs of the unwashed mass (where i am not excluded this time, as my hearing is very bad).
The audiophiles are a minority, and in those high end amps i am afraid, no amplifier chip will ever be used, as those amps are build discrete.

[999999999]

Re: Now THAT is a sound card!

MLC are somewhat non-linear. If we include C0G/NP0 types, which are often only available in small (pF for example) values but in recent years have expanded ranges through about .47uF (IIRC), they're not bad.

This potential for nonlinearity alone, depending on type, is reason enough not to put a low-level audio signal through one, but it's not really a cost issue, because there aren't that many coupling caps in a typical analog audio stage, some surface mount films are reasonably enough priced in volume if either type had an acceptible range instead of going with the higher uF value 'lytics.

In decoupling, ceramics demonstrate they do fine in more demanding analog, high frequency decoupling uses far beyond audible frequency range. Yes they're cheaper but they also work fine as demonstrated in far too many devices to begin listing. Decoupling is about removal of noise, we don't care if power rail noise we want to be rid of stays linear or not, but (generally speaking) it is most desirable to reduce the magnitude at higher frequencies.

Audiophiles do sometimes build integrated (chip-) amps, when there isn't a need for a lot of power. Even the boring LM3886 gainclones can compare favorably to high-end, all discrete amps if only implemented well instead of being built to lowest cost possible. When one doesn't need especially high current output (transistor) pairs, attaining a short feedback loop via IC (and good layout) can improve sound quality.

[Operandi]

Re: Now THAT is a sound card!

Audiophiles do sometimes build integrated (chip-) amps, when there isn't a need for a lot of power. Even the boring LM3886 gainclones can compare favorably to high-end, all discrete amps if only implemented well instead of being built to lowest cost possible. When one doesn't need especially high current output (transistor) pairs, attaining a short feedback loop via IC (and good layout) can improve sound quality.

Yes chip amps can sound very good, here is one such example.