Part of the Listen to This Audio Guide
An ad for the Victrola, 1915.
Here's the almost indecipherable 1889 cylinder recording of Brahms playing his First Hungarian Dance (p. 57):
For reference, here is a modern version, with Iván Fischer conducting the Budapest Festival Orchestra:
Jonathan Berger has closely analyzed about the Brahms cylinder and summarized his conclusions here. You can listen here to his MIDI reconstruction of Brahms's playing, in which he has projected the basic data of the recording onto the entire piece.
In 1915, Edison made a promotional film entitled The Voice of the Violin. At the climax, you can see one of the famous Tone Tests, at which audiences were allegedly unable to tell the difference between live performers and machines (p. 58):
The Library of Congress has a site devoted to the Edison company and its recordings. Vast quantities of early recordings can be found at the Cylinder Archive, UC Santa Barbara's Cylinder Preservation and Digitization Project, Tinfoil.com, and the Edison National Historic Site.
Decade after decade, advertisers have claimed parity between live music and the latest technological reproduction. Edison himself can hear "no difference" between his Diamond Disc phonograph and the real thing:
This famous Memorex ad claims that Count Basie can't hear the difference between Ella Fitzgerald live and on tape:
Recordings seemed to change how music was played, whether in the studio or live onstage. Early recordings show more expressive freedom and less conscious technical precision. The great Belgian violinist and composer Eugène Ysaÿe, in a 1912 recording of Vieuxtemps's Rondino, weaves his way around the tempo given by the piano (p. 63):
From the Cylinder Archive.
When Caruso recorded the famous weeping aria "Una furtiva lagrima" — on three occasions, between 1902 and 1911 — he created a cadenza that was soon "locked in" as the global standard, replicated note for note in the work of hundreds of subsequent tenors (pp. 63-64). This is Caruso's 1904 version, with the cadenza beginning at 4:45:
As the conductor-scholar Will Crutchfield notes, tenors once used this cadenza as an opportunity for personal invention, as in this 1909 recording by Giuseppe Anselmi: