Looking for a developer experienced in JUCE and FFT analysis to create an audio plug-in which derives bark coefficients from several separate audio streams simultaneously. A strategy for doing this in the most efficient way possible will need to be discussed. The bark coefficients will then need to be sent out via OSC, with an additional option to have them written to disk in the form of a text file.
The eventual goal of this project will be to build it into a sophisticated standalone looper, but in the meantime the FFT plug-in would be integrated into a Max/MSP project.
Some questions to start the discussion necessary:
- It’s a requirement that the several FFT analyses be conducted in the most efficient way possible. I assume this would mean doing them all in the same place (possibly multiprocessed?) rather than 1 or 2 per plug-in with multiple instances of the plug-in loaded into Max? This would require getting several audio streams into the same plug-in.
- As I understand it, an ordinary DAW would only recognize at most 2 audio inputs of any plug-in (perhaps 4 if side chaining), but would Max recognize any number declared by a plug-in loaded into the vst~ object?
- Would multiple plug-ins be able to access the same buffer in memory? If so, helper plug-ins might be able to write audio to any number of buffers as the FFT plug-in follows just behind, reading the data for analysis. An unconventional strategy, but could it essentially get several audio streams into a single plug-in?
- Does JUCE have the capacity to continually write to a text file efficiently (leaving the file open for reading/writing while doing so)?
Any interested parties should contact me directly or, better yet, help get the job moving forward by responding with your input to these questions. Thanks!