MSP audio CLICKS again - on fast sequencing (not in loopmode)

    xidance
    Jan 24 2012 | 1:42 am
    Hi,
    its an old issue and you find many postings, but none of the solutions works for my purpose. the grooveduck example is perfect for loops, but i do not get it working well with triggered playback. [trapezoid] is too slow i think, and line~ well, i just don´t get it done.
    the problem is bass tones, especially fast kick drums. when the same sample buffer is triggered a second time, before the first hit playback has cometo en end, it clicks. i tried fading down to zero before triggering again, but it is not really solid.
    please take some time to help (and please max5 patches only).
    • Rodrigo
      Jan 24 2012 | 8:51 am
      Take a gander over at this thread:
      https://cycling74.com/forums/i-hate-sample-accurate-coding-audio-rate-incompetence-and-help-needed-inside
      (there's some max6 stuff in the hanning/hamming buffer filling but even without that it works well).
    • xidance
      Jan 24 2012 | 10:12 am
      i´m sure this is a rich patch, but from my level of knowledge it seems hard to isolate the parts contributing to the click-issue.
    • Rodrigo
      Jan 24 2012 | 10:30 am
      If you look at a couple of the patches along the way it will be easier to understand. Theres even a good version using trapezoid in there. The problem with trapezoid is that it's relative to sample size, so on long files its a fade rather than a declick.
      The last few versions that Peter posted you can see some math used to figure out an absolute trapezoid fade which sounds great. If you just remove wave~ it ill work in max5.
    • Rodrigo
      Jan 24 2012 | 10:33 am
      You can also wrap it in a poly so rather than retriggering the same sound (which will be a problem for longer sounds anyways, like cymbals or 808 kicks etc...) you are triggering a second sound. Basically making it polyphonic.
      The tutorial (in the help) on poly is pretty easy to get your head around and would do just that.