Bravo, you have finally learned to play your cylinders electrically, none of the barbaric tin-horn sound heard on your former transfers!
Alas - that was only my first impression. On close listening, it becomes clear that you haven't transferred your own cylinder at all, but have duplicated - with some de-clicking and EQ - the transfer of cylinder #5540 (not the 2450 pictured) as provided by the University of California, Santa Barbara, at their website (UCSB cylinder 10860)! The grotesquely slow speed perfectly identical to both, and clicks in exactly the same spots, give the misattribution away. The transfer was not done at 160rpm either, but at 140rpm, which is totally wrong to begin with. The final verse is written in B-natural, and the transfers (UCSB's and the present one) come out in G, a major third below! While I agree that the recording was sung transposed down to make it easier for the soprano and tenor, one full tone below score pitch (last verse in A major) is sufficient. To achieve this, the cylinder must turn at approx. 155rpm, or the present transfer played back at a sampling rate of 49000 instead of 44100. BTW: The recording starts at 160rpm, but slows down do 155 within the first 20 seconds or so. At Méphisto's words "peut-être il en est temps encore" a steady pitch is reached which the remains constant to the end.
French website Comme-laque, celluloid et vieilles cires contains two 'takes' it: the first - as my own copy 2450, and second take 29386. Both were transferred with 160 rpm. I think the frenchmen known right speed.
Yes, "the French" (more precisely my friend Henri Chamoux who runs that website you speak of) by-and-large do, although he made no attempt to compensate the changing speed in this particular recording either: At 160 rpm, it starts correctly but during Fournets' initial phrases the pitch creeps upwards because the recording machine slowed down somewhat. Anybody with an ear for music can hear that, and since a piano cannot suddenly go sharp in pitch during a performance, it must be a purely mechanical issue. "Your" transfer, which isn't from your cylinder at all but pirated from Santa Barbara University, however runs at 140 rpm which is very far off the mark! Somebody at UCSB transferred a whole bunch of Pathés at that wrong speed two or three years ago - their curator David Seubert already knows about it but apologizes that they currently have no funds to re-engineer these transfers. You write "160" while we hear "140".