Determining high, medium, low frequency range of sample audio.

    Jul 22 2011 | 6:12 pm
    How can I root out the high, medium, and low frequencies of an audio sample. I want to turn those three frequency ranges into Data and use it to manipulate videos. I know you can do something like this with Vizzie but I want to record both video and audio. Unfortunately the Followr (Vizzie) object doesn't connect to jit.vcr so I'm trying to make a custom patch that works similar to Followr but I could connect to jit.vcr for recording.
    (I'm still kind of new to this and programming in general)

    • Jul 23 2011 | 7:47 pm
      There are many ways. Basically you want to split the audio into N (in your case, 3) separate signals, each filtering out a particular range. There are various filters that chop away various frequency ranges, in your case you could do this with a lowpass (keep frequencies below X hertz), a band pass (keep frequencies around X hertz, with a band witdth of W), and a high pass (keep frequencies above X hertz). You could do this with 3 filtergraph~ + biquad~ objects; OR with a lowres~ and reson~ objects.
      Turning that into "Data" is a more complicated matter, depending on what type of data you want: pitch/amplitude/rhythmic attacks/etc. If you don't know how to code these, one place to start is the external object "analyzer" which combines various analysis routines to turn signals into discrete data.
    • Jul 23 2011 | 11:12 pm
      Have a look at [ffb~] for the audio splitting.