WIDI performs music recognition in several steps.
First it produces some kind of spectra-like data; this is further called the "Wave->Pitch" conversion. At the present moment, there exist four different algorithms implementing this step; each of them has its own cons and pros.
Then it multiplies that spectral data by coefficients of "Equalizer", making some pitches louder and some quieter.
The third step is the actual notes retrieval from the spectra, it is called the "Pitch->MIDI" process and there are also two algorithms doing that.
The final step is producing the MIDI file structure, and you can choose some attributes of the output MIDI file.
|
