I’m interested in modeling an analog filter which I know responds differently at different frequencies. (The first one I want to try is a replica of the filter from the EDP Wasp)
My idea is to run white noise through the filter at various frequencies, sampling each one for some duration of time and saving a Fourier transform snapshot, then interpolating the results for filter frequencies in between the sampled frequencies.
Then, I can make a patch where presumably I can somehow apply these "snapshots" on the signal. This part of the step I am not sure exactly what I need to do. Perhaps I can perform an FFT on the incoming signal and multiply each bin against the corresponding bin in the snapshot, then run an IFFT on the result?
Are there any papers or references that explain a similar approach (not necessarily in Max)? Are there better approaches for what I want to do?
Edit: Obviously this approach would only apply for frequency, and I would need a 2D map of snapshots to model frequency and filter bandwidth (Q). But for now I just want to consider varying frequency with a constant bandwidth value.
Edit 2: I haven’t had any formal training with regards to DSP so bear with me if I’m not using terms correctly :¬(
Edit 3: Urgh, I just realized that different amplitudes of the same signal will likely respond differently at a given filter frequency for a nonlinear filter, adding another necessary dimension to my analysis. There are probably other dimensions that I haven’t considered yet as well. This is starting to look like a pipe dream….