This is where navigation should be.

DEMO_BLOCKPROC_PHASERET2 - STFT phase reconstruction algorithms comparison

Usage

demo_blockproc_phaseret2('gspi.wav')
demo_blockproc_phaseret2(...,'a',a)`

Description

For additional help call demo_blockproc_phaseret2 without arguments.

This file does real-time reconstruction from the STFT magnitude using 4 algorithms. No effect, just plain reconstruction.

The order of the algorithms (Algorithm slider) is the following::
0 - RTPGHI 1 - RTISI-LA 2 - GSRTISI-LA 3 - RTPGHI + GSRTISI-LA

Please note that you might experience some artifacts when switching between the algorithms or when changing their parameters. This does not have anything to do with the algorithms themselves.

demo_blockproc_phaseret(...,'a',a) allows setting the hop factor. Default value is 256.

../images/phaseret2_common.png ../images/phaseret2_rtpghi.png ../images/phaseret2_rtisila.png ../images/phaseret2_gsrtisila.png ../images/phaseret2_proposed.png

This code produces the following output:

DEMO_BLOCKPROC_PHASERET2:
To run the demo, use one of the following:

demo_blockproc_phaseret2('gspi.wav') to play gspi.wav (any wav file will do).
demo_blockproc_phaseret2('dialog') to choose the wav file via file chooser dialog GUI.
demo_blockproc_phaseret2(f,'fs',fs) to play from a column vector f using sampling frequency fs.
demo_blockproc_phaseret2('playrec') to record from a mic and play simultaneously.

Avalable input and output devices can be listed by |blockdevices|.
Particular device can be chosen by passing additional key-value pair 'devid',devid.
Output channels of the device cen be selected by additional key-value pair 'playch',[ch1,ch2].
Input channels of the device cen be selected by additional key-value pair 'recch',[ch1].

References:

Z. Průša and P. L. Søndergaard. Real-Time Spectrogram Inversion Using Phase Gradient Heap Integration. In Proc. Int. Conf. Digital Audio Effects (DAFx-16), Sep 2016.

Z. Průša. Towards High Quality Real-Time Signal Reconstruction from STFT Magnitude. Submitted, 2017. Preprint available at http://ltfat.github.io/notes/ltfatnote048.pdf.