Audio Descriptors / Analysis - Offline

    frank desben
    Mar 05 2019 | 12:10 pm
    Dear all,
    I have been working with numerous audio descriptors (zsa, ircam, ...) - but always in realtime. What I would like to do is to analyze different samples or sections of samples - so basically a buffer. But not playing the content and then analyzing it in realtime - but as an offline procedure - and through this then as a bulk process.
    What's the best way to do this in Max?
    Thank you in advance! Frank
    • frank desben
      Mar 07 2019 | 8:58 am
      I guess there is MUBU, correct? Are there any other ways? Also perhaps in JS or via Node or via Jitter? Mubu also seems to crash very often for me...
    • Pedro Santos
      Mar 07 2019 | 6:55 pm
      You can do always do any MSP processing in nonrealtime choosing the "NonRealTime" driver option in Audio Setup. In this mode, the computer processes the samples as fast as it can
    • Nico Daleman
      Mar 07 2019 | 7:18 pm
      Nice, I was wondering something similar. Seems like MuBu is the way to go within Max. I've been using EAnalysis. But is not so straight forward and it's kind of outdated. I think most people use python / matlab for visualizations, specially if you don't need realtime/playback. Also I'm following up for a more qualified response. Cheers!
    • frank desben
      Mar 08 2019 | 1:00 pm
      "NonRealTime" driver is a good idea. Ideally I would like to be able to switch between analysis and other processes without having to stop the audio driver.
    • Pedro Santos
      Mar 08 2019 | 5:37 pm
      You could use the "adstatus" object and automate the following procedure: 1- switch driver to nonrealtime 2- do the processing 3- switch the driver back to the main driver (CoreAudio, ASIO, etc.)