The application is located in /var/cs325/bin/pV. It is an executable, so start it by simply typing 'pV'.

The pV application is simply a Tcl/Tk front-end to Professor Paul Koonce's phase vocoder package, which itself is based on F. Richard Moore's phase vocoder implementation.

This is a work in progress. Currently, only the PlainPV module front-end is supported in Paul Koonce's package, which provides the most basic phase vocoder functionality. Other spectral transformations are of course available using FFT-based analysis/resynthesis, and these will be supported in pV in the future.

As an example, let's time-stretch a sound by a factor of 2 while preserving its pitch. Open the pV program and select 'Open Input File' from the File menu. Open /var/cs325/colby/water.snd. Click the 'Play Input' button to hear the sound.

Under the Options menu, click 'Analysis Settings." Set the FFT Size to 1024, the window size to 512, the output data format to 'same as input', and the frames per second to 200. Then click 'Done'.

If nothing is entered in the Start and Stop fields, the program will default to processing the entire sound. These fields are primarily useful for spot-checking portions of a long sound without having to wait for the computer to process the entire sound.

Pull down the Output Channel menu and select 'All' so that the program processes both channels of this stereo file.

Click the 'Compute' button. Notice that Paul Koonce's program prints output statistics to the console window from which you launched pV. When resynthesis is completed, click the 'Play' button to hear the sound. You can click 'Play' during processing as well to hear the output file as the samples are being computed.

This is alpha software. I know there are plenty of bugs still. I am aware of many of them but would appreciate any feedback.

Colby