The characteristics of plastic film foil caps are very similar. Polar dielectrics such as polyester have poorer high frequency response (although this is probably not an issue with guitar amps) when compared to say polypropylene, which is as near as dammit the perfect dialectric.
A couple of possible explanations for the difference in sounds of capacitors (assuming that the the difference is real):
1) The caps are different values. Caps tend to have quite poor tolerance, 10% is not uncommon, so a pair of 0.1 µF 10% caps might well be 20% apart in actual value.
2) Parasitic effects e.g. series inductance or capacitive coupling to other parts of the circuit or ground. Parasitic effects will depend on how the cap is wound, its orientation with respect to other components in the amps, and its size.
Many vintage guitar amps used cheap ceramic caps in the signal path, eg all Blackface Fender amps used ceramic bright caps, treble caps in the tone stack and coupling cap into the phase splitter.
"Audiophiles" tend to believe that the more expensive the part the better it is. Whether this is true in a guitar amp where an imperfect part might add a pleasing character to the sound well I don't anyone has a definitive answer for this.
By the way I see plenty of vintage amps where all the caps have been changed, probably as a way of up charging a repair by "improving" the caps. I don't change coupling caps in amps unless I want to change the value to revoice the amp, or if they have gone leaky. For what its worth I use ERO polypropylene 630 V caps.
Regarding carbon composition resistors, there is probably something in this. I have yet to do the definitive experiment (which will involve building two identical amps one with CC resistors and one with metal film. This should be completed in the next couple of months or so), however I was chatting to a amp designer from the US who was asked to develop an amp for a major US manufacturer They wanted three amps — one with metal film resistors, one with carbon film resistors and one with CC resistors. The first two sounded identical, however the CC load amps sounded as if all the top end had been rolled off to the extent that the designer thought that the amp had been wired incorrectly! However, he could provide no rational explanation for this, as the known defects in CC resistors (voltage dependent resistance, and shunt capacitance) couldn't really account for this. Anyhow, I will be repeating the experiment later soon and will report back.
Regardless I expect to really hear a difference with CC resistors you would need to change more than the slope resistor in the tone stack.
As for noise, the main culprits are the plate load resistors where CC resistors can be very problematic (I've just had to change all the pre-amp plate load resistors in a BF Fender Deluxe as they had gone noisy), so I don't think a CC resistor in the tone stack will have much of a detrimental effect on noise. However, I expect in a high gain amp CC resistors would prove to be too noisy.
