(Jamoma Platform) » DSP

(There is an FFT branch in Git for an FFT Lib extension already)

We should schedule a day where several of us can work on this. Some of the questions we should try to answer:

• How do we control FFT size? Is it simply the vector size of the audio passed in? Or should it be fixed, and then we have to buffer the audio that comes in? If the former, then I guess we should create an object the buffers audio to provide the possibility of changing vector sizes?

• How do we control the offset of an FFT? For example, in Max's fft~ object there is a third argument for this. However, why should we do this in the FFT? isn't it something we could control outside of the FFT? How?

• For what is the second argument in Max's object useful? Do we need something like this?

• I (tim) am assuming that windowing, overlapping, and adding, all happens outside of the fft object.

• Are the FFT and IFFT objects the same object? Or are the different objects?

• In Max, a sync outlet provides the fft bin number. If we simply process a the FFT based on the size of the audio signal passed in then we don't have to worry about this because we will be receiving and sending complete frames of the FFT rather than sending them piece-meal.