Serge Auckland wrote:

I have never measured any increase in THD with a capacitor in circuit in
the amplifiers I have designed. This is true of all sorts of cap,
electrolytic, tantalum, polyester, polystyrene.

I've only seen it with small value electrolytics as coupling caps at low
frequencies and very high signal levels with low load impedance on the load end
of the cap ( *where the voltage across the cap starts to become significant* ).
Even so, the numbers are quite low. Roughly from memory with 10uF and a signal of
10V rms @ 20Hz into a 600 ohm load it was somewhere around 0.03%. Reduce the
signal level to something more typical, increase the load R, increase the
frequency and most of all using a larger cap ( to a more sensible 100uF ) all
reduce the figure. In fact with 100uF I simply couldn't measure *anything* except
with very high signal levels @ 20 Hz and even then it was barely out of the test
set noise floor.

In more typical real world use even the electrolytics produce no measurable THD (
AP THD floor is 0.0008% or -102 dB )

Graham