Audio orientated capacitors will win but I am not sure by how much. That leakage mentioned shouldn't have any significant effect, except maybe make turn on/off pops louder, though there may be circuits where the added DC offset can get blown out of proportion and early clipping occurs. Ignoring that, there shouldn't be any audible differences. I would still go with audio orientated or regular e-caps in signal path over polymers.

Interesting question, I recapped my SB Pro 2 a few years ago, I got good electrolytic caps from DigiKey but not specifically designed for audio (maybe a few, I don't remember exactly). I didn't really care because sound FX and music didn't have good quality back then anyway (compression, etc), specially in games. IMO, I don't think there would be a noticeable difference, but I can't say for sure.
Looking at Badcaps.net/forum/ I just found this:
https://www.badcaps.net/forum/showthread.php? … t=sound+blaster
In that particular case, it looks like Creative wasn't using the right cap values and in some cases the correct type of cap. I have read about similar cases before. So, there may be a difference, but the solution may not be as simple as replacing old caps with good audio ones.
I recommend you spend some time researching their forum, it's a bit of a pain but it may be worth it. It looks like I'll have to do the same too and revisit my SB Pro 2 and start looking at my SB 16.

You could throw money into the fire on "audio grade" caps if you wanted, but you'll get much the same results from some halfway decent quality new caps in general. I find audio-grade caps to be pretty much a gimmick: to my ears, Nichicon's 'cheap' VP/UVP generic bi-polarized caps sound all but identical to Fine Golds, Muses and Cerafines, when put in the audio signal path. And they are far cheaper than those 'audio' caps. But the Fine Golds and Cerafines sure look pretty, I'll admit.

Some folks swear they can hear a favorable difference between low-ESR Panasonic caps and other brands. I remain neutral on this personally, but I do use Panasonics sometimes on my cards if I get a good deal on them.

Thank you very much for the feedback so far. Replacing bad chosen caps by Creative with better ones makes sense when recapping, I'll have a look at this.

Sometimes you don't have a choice which type of capacitors to use when the space available matters. My AWE32 is populated with wet caps of a small diameter. I haven't found organic polymer caps at Mouser those diameter fits but Nichicon Fine Gold Muse fits (they are thinner and longer).

Talking about space brings me to the next "audio caps" question:
I'm going to recap my NEC XR385 wavetable module, which is a clone of the Yamaha DB50XG. Since it is put piggyback on the sound card the height of the replacement caps plays a big role. When I compare the pictures of both wavetable modules it seems that Yamaha used Tantalum caps. As far as I know their current should either be limited when switching on the PC or their max. used voltage should be derated to 50% of their nominal voltage. The largest 47 µF Tantalum fitting on the NEC board I found on Mouser has 20V. Since they're used for buffering 12V it derates just to 60% and not 50%
Do you think that it is a good idea to use Tantalum caps in this specific case or in audio circuits in general? Should I worry about the caps used on the Yamaha board?

Tantalums are not good in signal path and should be avoided whenever possible, only X and especially Y dielectric ceramics are worse than those. They are good for power filtering duties where the nonlinearities do not matter.

When I have a look at Mouser for SMD caps which might fit on the NEC board I just find Tantalum and MLCC caps besides the "original" wet caps. Which caps might Yamaha have used on their board? For me the caps on the Yamaha board look like Tantalum, what do you think?

Wikipedia says "A particular cause of confusion is that on surface mount tantalum capacitors the positive terminal is marked with a bar. Whereas on aluminium surface mount capacitors it is the negative terminal that is so marked. "
https://en.wikipedia.org/wiki/Tantalum_capacitor

Yes, those are tantalums on the DB50XG and yeah, they mark out the positive side and not the negative. They will blow up if you install them backwards in many situations 🤣.

Many situations? I thought that they go up in smoke and fire in ALL situations when installing tantalums backwards 🤣.

Yamaha scares me a little bit because the 47 µF caps seem to buffer +-12V and +5V. When I have a look at the voltages they obviously ignored the 50% voltage derating rule. I found in the Intel ATX power supply design guide that the ramp-up of the voltages (from 10% to 95%) during power-up should take 0.2 t0 20ms, see section 3.3 of:
https://www.intel.com/content/dam/www/public/ … -guide-june.pdf

Should I worry about this or is a ramp-up to 75% of the rated voltage within 0.2ms (talking about worst case) still fine for tantalums?

In signal path there is not enough current to cause destruction but across a power rail there is potential for destruction. Modern tantalums have fuses in them to prevent explosive failures and are less fragile in general.

I wouldn't worry about the derating aspect too much, inductance of the traces to the cap limit inrush a lot and PSUs have soft start anyway that is slow enough that this should not be a problem at all.