How to improve reaction time to audio input?

  • I am working on a dance performance where my video should react pretty tight to the noise(footsteps) dancers make. I have 8 quad distort actors controlled by frequencies of incoming audio. So far I have been testing with MBP microphone. The lag to reaction is too eminent - about 1sec.

    How can I modify the patch to improve the timing.


  • Dear Vanakaru,

    What is the resolution and codec of the video you're feeding to the Quad Distorts? I'll need to know that to try it for myself. 
    Best Wishes,

  • I noticed in your Smoother actor the smoothing parameter is 0.98 --- is this 1 sec delay you observe?


  • eight you are the winner. I never thought there would be correlation with smoothing and time. Thanks!

    [mark](  PNG 1200x1200 with alpha file size 20-30kb. You may be able to tell me what to do if I need smoothing closer to 1.

  • Dear Vanakaru,

    Yes, the bigger the smooth, the longer the delay. The effect is very similar to a capacitor in electronics... to smooth the signal, you must let it slowly accumulate. Thus, reaching a the current value will take longer if the smoothing is higher. This photo kind of tells the story:
    The yellow line is a low smoothing factor (capacitance in fact for this diagram) and the red line is a high smoothing factor. To get to the high point takes considerably longer because of the higher smoothing/capacitance, thus the delay.
    Hope that helps! (Good call eight!)
    Best Wishes,
  • Tech Staff

    Rather than smoothing, you might want to try Curvature.. You may be able to ramp your values to get an acceptable visual effect this way without a time loss.

  • Thank you for the input. I tried the suggestions. I find the best solution to use Smoother still and mostly with parameter closer to 1. Otherwise the outcome will be too erratic and nervous. But to get the perception of more accurate timing I shall make just one part of the picture reacting rapidly and the rest smooth.

    I was not able to get satisfactory results using Curvature.
  • Tech Staff

    If you know your destination value at the start of smoothing.. you could dynamically set the smoothing amount based on the difference between the destination value and the current smoother value.
    I have used this method before.. and it worked great.. basically, starting smoothing at .98 and scaling it down to .65 as the value get closer the destination value. This limits the delay, but provides a smooth start.. with a snappier end.

    Of course since you are working with audio triggers I don't know if you have the destination value in advance.

  • Beta Gold

    Hi vanakaru

    you might want to check that very low latency. if you know pure data or max all is due to an object called bonk.