|
From: Hans S. <han...@gm...> - 2009-05-22 01:01:45
|
> I should point out that a spectrogram is basically just a windowed DFT and that many packages (e.g. Matlab, Octave, Scilab) are capable of producing these in a numeric form that can be further processed --does this go any way towards meeting your requirement for this? Pretty much, thank you. I've set up a script by now that takes the PNGs outputed by SoX and reassambles them (basically like this: mplayer converts to wav, sox parses header and returns length, and is then invoked multiple times with trim, ImageMagick strips descriptions rotates and resizes outputting raw greyscale, which are concatenated and then fed as raw 1 pixel height greyscale stream into mencoder which resizes and produces a 64x64 video with huffyuv, which is then fed into ffmpeg2theora) and it works, but is by no means a clean solution. As I said, I'm new to audio processing and terms like "open source spectrum visualization" turns up astonishingly few useful results. I'll take a look at these programs, thank you! --hans |