"The best thing I've ever heard said of octaves was Edison's remark to
me that 'They are merely a nuisance and should not be played!' I was
making some records for him during the experimental stage of the disk
record, when he was trying to get an absolutely smooth _legato_ tone,
one that conformed to Loeffler's definition of it as 'no breaks' in the
tone. He had had Schubert's _Ave Maria_ recorded by Flesch, MacMillan
and others, and wanted me to play it for him. The records were all
played for me, and whenever he came to the octave passages Edison would
say: 'Listen to them! How badly they sound!' Yet the octaves were
absolutely in tune! 'Why do they sound so badly?' I inquired.
"Then Edison explained to me that according to the scientific theory of
vibration, the vibrations of the higher tone of the octaves should be
exactly twice those of the lower note. 'But here,' he continued, 'the
vibrations of the notes all vary.' 'Yet how can the player control his
fingers in the _vibrato_ beyond playing his octaves in perfect tune?' I
asked. 'Well, if he cannot do so,' said Edison, 'octaves are merely a
nuisance, and should not be played at all.' I experimented and found
that by simply pressing down the fingers and playing without any
_vibrato_, I could come pretty near securing the exact relation between
the vibrations of the upper and lower notes but--they sounded dreadful!
Of course, octaves sound well in _ensemble_, especially in the
orchestra, because each player plays but a single note. And tenths sound
even better than octaves when two people play them.