evenly dividing up a stream of floats

    Sep 18 2009 | 4:24 pm
    OK, I'm having a brain freeze today. I know I will slap my forehead on this one when someone gives me a clue!
    I have a stream of floats going from 0.0 to 1.0 (which represents a quarter note duration). I would like to send a bang at specific, evenly divided points within that stream. For example, just at 1.0 (quarter note), or just at 0.5 and 1.0 (eighth notes), or just at 0.25, 0.5, 0.75 and 1.0 (sixteenth notes).
    How do I cleanly do this? It seems like it should be a one or two object operation, but for the life of me I can't figure it out today!

    • Sep 18 2009 | 5:02 pm
      Have a look at [%] although be careful as dealing with floating point numbers can cause problems, you might want to scale it up by multiplying by a power of ten, depending on the resolution you want.
    • Sep 18 2009 | 5:13 pm
      Yup, that's it! My forehead is nice and sore right now!!
      Thanks a million.
    • Sep 18 2009 | 7:14 pm
      OK, so I spoke too soon. What I'm really trying to do is to take ReWire timing info via hostsync~, and to be able to fire off bangs at eigth, sixteenth, etc.. intervals.
      The patch below, although I think technically correct, yields extremely poor timing:
      I took your advice and scaled up by 100 and using ints instead. Overdrive is on.
      Any ideas?
    • Sep 18 2009 | 7:48 pm
      Have you tried hooking that outlet of [hostsync~] straight into a [print] object? It might be that the timing resolution here is causing the problem. If it is skipping a lot of numbers then you will never hit exactly on your divisions and the bangs won't get triggered. If this is what is causing the issue then you might be able to use [line] to send out all the numbers in between the values at a rate determined by the tempo of the track to keep it smooth. It's just a thought but good luck!
    • Sep 18 2009 | 8:01 pm
      You're right! Here's a sample of the output:
      print: 0.021657
      print: 0.044900
      print: 0.068142
      print: 0.091319
      print: 0.114561
      print: 0.137803
      print: 0.160980
      print: 0.184223
      print: 0.207465
      print: 0.230642
      print: 0.253884
      print: 0.277126
      print: 0.300303
      print: 0.323545
      print: 0.346788
      print: 0.369965
      print: 0.393207
      print: 0.416449
      print: 0.439626
      print: 0.462868
      print: 0.486045
      print: 0.509288
      print: 0.532530
      print: 0.555707
      print: 0.578949
      print: 0.602191
      print: 0.625368
      print: 0.648611
      print: 0.671853
      I'll think about the line approach you suggest, or perhaps I'll have to use the raw ticks or sample count output instead...
      Thanks again,
    • Sep 18 2009 | 9:06 pm
      I've never used hostsync~ before, but as far as I can tell, hostphasor~ would be much better for your purposes. It basically outputs the same thing, but as a signal which you can treat exactly like a phasor~. You will almost certainly get better timing this way. For example:
    • Sep 19 2009 | 12:32 am
      That's a good suggestion, thanks!