Just to give me a better idea, do you have a link to one of the sites for the kit?

My interest in the 91 amp is entirely theoretical and historical, not practical. So this is all very much FWIW.

My impression is the big sources of noise are that the historical circuit, with the 310A driver tube, has like 40 dB of gain in the driver stage. This was for way back in the day, long before 2V source components. That really amplifies downstream noise and most people deal with it by changing the driver stage, often going with 6S[N|L]7 or C3g tubes and much less voltage gain. The other thing is the AC heater for the 300B tube, which is a whole topic by itself. Steve talks about it in the Sarah design thread. The obvious alternative would be a DC heater.

As for a choke, do you mean a filter choke in the power supply? If so that is relatively straight forward, there is an IR voltage drop across the choke so you need to compensate with a higher voltage power transformer if you want to maintain the same B+ as you would have with just a cap. That would also allow you to decrease the first filter cap to make it more rectifier friendly. But taken all together this adds up to a major redo of the original 91 circuit if you want to start from scratch.

This is all my impression from reading various online fora and a few books/magazines over the years, so again, very much FWIW.