The author at A Revolutionary Critique of Robert Johnson suggests that Robert Johnson's recordings are being played too fast by about 20 percent. At an online forum discussing whether Robert Johnson's recordings were speeded up, it was suggested that the power line frequency might provide an accurate reference. At the time Robert Johnson made his recordings, power line frequency hadn't been standardized to 60 Hz yet, and 50 Hz was in use in some places. I've not been able to find any information on what the power line frequency was when (and where) the recordings were made.
These are plots of the low frequency portion of the spectrum of five tracks from "Robert Johnson: The complete Recordings" on CD. These are from the original release. I understand that they were rereleased with some changes, but haven't tracked down exactly what those changes were.
The plots below are for tracks 2, 4, and 5 from the first disk, and tracks 6 and 10 from the second disk. Notice the sharp peak at precisely 51.4 Hz in all of them. These are present in all the tracks I've looked at from that disk, always at the same amplitude. The peaks near 60 Hz move around in amplitude and in frequency, and are broader. Listening to just the near-60 Hz signal, the sound varies, but listening to the 51.4 Hz signal, it's a pretty constant tone, as power line noise would be. If the 51.4 Hz signal represents the electric power line frequency, and if that frequency was 50 Hz, then the Robert Johnson recordings are about 2.8 percent too fast, or about half a semitone (1/4 step).