The experts agree that the hi-fi electronics are less a determinant factor
- the recording (it's cheap to buy a different, better recording)
- the listening room (affecting the sound considerably)
- the loudspeaker.
Still, there is a difference in using a transistor or valve amp.
Capable transistor amps usually employ a negative feedback per amplifier
section, reducing all faults practically below the level of audibility (that
doesn't necessarily mean that a transistor amp sounds best).
However, all amps have in common that capacitors for coupling and DC extraction
are being used between the individual amplifier sections.
These capacitors should obviously be of a quality that doesn't generate any
losses. The replacement of electrolytic capacitors by foil capacitors makes
a lot of sense.
Or when your (slightly aged) main amp employs 4000 - 10000 μF electrolytic
capacitors, it's quite useful e.g. to add a 100 μF MKP capacitor in parallel.
We mentioned already in our discussion about crossover tuning that electrolytic
capacitors employed in the signal path cause an audible deterioration of mid
and high frequencies.
The additional buffering of capacitors in the power supply of the power amp
usually doesn't seem to achieve much; but then, some experimenting might show
some amazing results.