Re: Talking piano

[Peter Meijer <feedback@xxxxxxxxxxxxxxxxxxx>]

I enjoyed this cool demonstration of what I consider to be additive synthesis
using repeated piano tones instead of pure sines in sounding a human speech
spectrogram. The human brain is extremely tolerant to speech distortions, as
was also nicely demonstrated long ago by 3-tone sinewave speech,

   http://www.scholarpedia.org/article/Sine-wave_speech

so why not 88-tone piano speech.

Apparently one can also get away with replacing the sine waves of additive
synthesis by piano tones while ignoring any phase relationships across the
spectrum (and accepting the newly added but now "meaningless" or "spurious"
phase relationships and amplitude envelope that arise from each piano tone).

And pianists: don't despair, I can imagine that a further simplification to
10-finger piano speech might also work well enough, by analogy to the 3-tone
sinewave speech results... :-)

Best regards,

Peter Meijer


Seeing with Sound - The vOICe
http://www.seeingwithsound.com