What it does
Generates a natural language speech based on piano keys.
How I built it
Text gets synthesized to speech. The speech gets divided into smaller pieces. Multiple peaks of the frequency spectrum of each piece gets collected. The ones that are above or below the frequency of the piano will be ignored. The remaining list of frequencies will get matched to the frequency list of 88 keys of the piano. The ordered sequence of piano keys corresponding to the synthesized speech will be merged together and played back.
Challenges I ran into
My first interaction with DSP, so practically everything about FFTs and Autocorrelations and windowing and whatnot was a challange!
What I learned
Some basic application of DSP.
What's next for Talking Piano
Polish it and publish it as open source.