The thing about bypassing is that if you try it in an RF circuit you get: fire.
What this tells you is that you might gain some clarity due to speed increase in pass through, but there is still a problem of trapped energies in the temporal domain, ie phase issues, in the paring of the caps.
Which also means that some of the energy is being trapped and released, out of phase, for a bit of ’temporal smear’.
The ear hears this and can sometimes relate to it as increased detail. But due to smear, it can also be obscuring signal clarity. Both happen hand-in-hand.
This is why bypass caps should be no more than 1/10th the value of the main film cap it is bypassing, and more like 100th the size.
This, to try and avoid resonant can modes in the pairing under high slewing high current.
Which is especially possible in audio circuit use, due to very complex and exceedingly wide bandwidth considerations, which no RF circuit would ever encounter. Not just likely (I was being polite), but that it will be there, in some way or another, in a given audio circuit application.
It’s a trade off game, with all that in mind and being listened for. It can be half increase in quality and half increase in smear for the impression of greater clarity but really, the exact opposite.
Bypass caps are most efficient when the impedance differential is the greatest, as in in conjunction with an electrolytic cap. The large impedance differential between the two caps...helps avoid high level can resonances in the paring. the can resonance will still occur ...but the benefit outweighs the detraction, in most cases of properly done film bypass of an electrolytic.
Where with a pair of film caps, the can resonance problem is virtually guaranteed to happen in some notable way. The counter argument is that the smear is really fast, and of less importance due to the very high speed aspect. You decide, with all that in mind and being listened for.
Thus, for audio and pure film caps in the given application, the best direction is generally the highest quality single cap you can get in there, and like the sound of. It might be darker to the quick analysis... but in the final analysis, more correctly scaled detail will emerge, in my experience. To get this part right is important so that mistakes in changes are not made elsewhere to compensate for mistakes made here. Then it’s just an insane correction circus with no end (as the mistake was missed and thought to be a real and correct backdrop).
Importantly, you can’t have this discussion I just brought up at DIY audio. To many opinions raging on at the red cape of the thing I just said. There would be no discussion, just attacks, for suggesting that the ear is as important as the biblical dogmatic math. Math which came from human minds (and math exists no where else- math is conceptual, it is not a reality) which were considering realities. But never mind that fundamental contradiction in engineering dogma.....
