A neat tidbit I found on the Last Word column in the NewScientist e-zine. In response to the question "Someone claimed they could identify a piece of music just by looking at the record. Is this true?". Peter Copland, who is the Technical Manager at the British Library National Sound Archive in London posted an excellent response:
It is certainly possible to identify a piece of music on an LP record just by looking at it and you can hone your skills by using scientific principles.

The first clue is the length of the track (or tracks). This can be estimated within a tolerance of 10% or so by a trained eye, which can also take into account different groove pitches. Errors are comparable to different performance times for the same piece of music, so they aren't significant.

The next clue comes from the dynamics of the music. This can be read more easily by shining a parallel beam of light onto the surface of the record. The light is diffused by an amount proportional to the root mean square velocity of the stylus which cut the sound waves in the groove. Thus the dynamic shape of music is easy to see.

The third clue is the spectral content of the music. By looking straight down onto the disc, the horizontal surface between the grooves (known as "land") is reflected back to the eye. Where low-pitched notes occur, the cutting stylus leaves comparatively long islands of land as it vibrates to and fro. By putting together these last two pieces of evidence, one can easily distinguish between, say, a solo piccolo and a solo double bass. So one can gain an idea of the instruments used and how it changes with time. In the days of 78 revolutions- per-minute records and early LPs when the range of recorded repertoire was much narrower, these three pieces of evidence could be sufficient to identify the music unambiguously.

More information can be found in a paper by G. Buchmann and E. Meyer: "Eine neue optische Messmethode für Grammophonplatten" (Electrische Nachrichten-Technik), 1930, vol 7, p 47. This is the original citation, but the mathematics of the principle were not described until E. Meyer did so in his book Electro-Acoustics (G. Bell and Sons Limited, 1939, p 7).

... and a side comment Peter later makes about reading CDs:

I can't say I've mastered the art, but it's easy to see the length of a recording (which starts at the centre), and this is a vital part of my work--because we copy analog sounds to CD discs, we often need to distinguish between blank and recorded CDs when we label them. I can often see when a CD has been assembled track-by-track instead of being "burnt" continuously, so I can find the track changes and check them. However, manufactured CDs are mastered continuously, so this technique will not help.