Real-Time Spectrogram Inversion (RTSI)

Python implementation of RTSI. The algorithm needs no iterations and no information about future spectrogram columns. It is ideal for true real-time ('streaming') audio conversion applications. It inverts the spectrogram column-by-column, each time using the overlapping previous frames to make a good estimate of the phase of the current frame. The algorithm is inspired by RTISI-LA (see paper).

Can be as simple as:

rtsi = RTSI(win_length=win_length, hop_length=hop_length)

spect       = rtsi.audio_to_spect(audio)
audio_recon = rtsi.spect_to_audio(spect)

See simple_demo.ipynb for more examples

See also

http://ltfat.github.io/phaseret/doc/gabor/rtisila.html

TO DO:

• enable multiple iterations
• enable lookahead