Arkless Electronics
Trade: Amp design and repairs.
The valves at RF/IF probably don't matter much beyond, perhaps if the LO valve isn't steady. However the details of the IF filtering do matter as that can alter the FM pattern, and hence cause distortion, etc. The demodulator can also affect this.
For the above reason, the performance might also vary with signal level if that means the source impedances of the valves alters, in turn altering the matching to the staged filters.
Devils in many details.
Very much so yes and these details will matter vastly more than whether valves or SS are used at RF or IF. Microphonics in the LO valve would cause FM and be demodulated yes and could add a tiny amount of colouration... there again microphonics in any part of a LO whether SS or valve will do this.
My main point was that any thoughts of "oh it's a valved tuner so will have 'the valve sound'" are probably incorrect other than any "valve sound" added by an output amplifier or cathode follower, where used, and that in the case of a TL2 with decoder there are no valves in the audio signal path.
FWIW I once designed and built an FM tuner using wideband minimum phase ceramic IF filters and an unusually wideband Foster Seeley Discriminator (and 74LS TTL devices in the limiter) and it was the best sounding tuner I've ever heard! I had an inkling that maybe a good FSD would give better subjective sound quality than a quadrature detector, hence trying this out.
A McIntosh MR67 was about the best sounding commercial tuner I tried.