> So even if MIDI could replicate the piano's sound
> (which would make the test more difficult) and people
> got questions wrong
> Secondly - and this was the problem with the Faulkner
> test as well - you cannot take elements out of their
> context.
Well, people *do* get questions wrong. I don't automatically keep track of the quiz results, so only know when people send me a message. From those the statistics for the "MIDI or virtuosi?" quiz ( http://www.ee.ucla.edu/~simkin/midi_or_virtuosi.html )
is as follows:
100% 3 83% 1 75% 1 67% 1 50% 2
Regarding "Machine translation or Faulkner?" quiz ( http://www.ee.ucla.edu/~simkin/sounds_like_faulkner.html ) you had problems with it, because you don't understand Faulkner (some people got 100% on that quiz).
Mikhail Simkin --- http://www.ee.ucla.edu/~simkin/