Reading peaks from a waveform


    Oct 28 2010 | 11:42 am
    Hi there,
    I've been working on this retarded patch the whole morning and I did all I need except for this little problem.
    Here is a simplified version of the patch I'm working on.
    1 / Just replace a sample into the buffer
    2 / Toggle the [sig~]
    3 / Change the number to move wherever you want on the sample and bang the [peakamp~].
    From what I learned from the max help at this point the [peakamp~] should give me the amplitude value but most of the times it seems incorrect.
    Do you have any idea about how to read peaks at specific points?
    Thanks as always.

    • Oct 28 2010 | 11:50 am
      Does this help? You're using a millisecond value to check the peak values, and this won't be sample-accurate, so will miss some peaks.
    • Oct 28 2010 | 12:07 pm
      Thanks Tim,
      suddenly I thought that your solution was working correctly but I then figured out that I was having the same problems. It looks like a max problem but I don't know how to explain it.
      Try to put the attached file into the buffer for example. There are some empty parts in the loop which are read from the [peek~] and the [peakamp~] as well giving really strange values that looks non sense to me.
      Anyway thanks for remembering me about the [peek~]. I didn't like to use a [groove~] for that.
    • Oct 28 2010 | 12:22 pm
      The link to your .aif isn't working. I have to go make some soup before I implode, but I'll have a little think about it while I cook :) You're not driving the [groove~ > peekamp~] and [peek~] with the same integer object are you? That would produce completely different results as one is ms and the other is samples.
    • Oct 28 2010 | 12:43 pm
      Hey I was wrong!
      Your last message opened my eyes in some way!
      I was using the same number to move the line on the [waveform~] and to control the [peek~] as well. But in order to match the visualization and the [peek~] output I had to multiply by 44.1 the number before going into the [peek~].
      I've attached the final one.
      Thanks Tim
    • Oct 28 2010 | 12:45 pm
      By the way, have a nice lunch... or dinner.
    • Oct 28 2010 | 7:25 pm
      for peak by sample just use minimum and maximum ...
    • Oct 28 2010 | 8:16 pm
      I don't think degezo wants the "peak" value of the buffer, but rather the rectified amplitude value of a variable time point within the buffer.
      (so in my comment object, I shouldn't have put "peak sample value" because it isnt ;) )