Technical discussions related to Audio Signal Processing (digital effects, acoustics, noise reduction, musical signal processing, etc).
Does anyone out there have experience or info in how to construct fractional-octave bandwidth filters using FFTs? It does not seem to be an accepted method, yet I cannot understand why. The typical (ANSI S1.11-1986 approved) method uses digital IIR filters in the time domain. I'm interested in a FFT implementation to enable processing of short (transient) signals that are obscured by the settling time of the IIRs. My application allows post-processing (i.e., non-real time) of the data. Issues in the FFT approach (conceptually, by summing power in adjacent bins) include window selection, zero-padding, how to handle power in partial bins, etc. Thanks