simultaneous mixing


    May 24 2006 | 12:04 pm
    Hi there,
    I try to mix two qtmovies( i use single frames in the moment with the importmovie-message) and blend them( i use the jit.op.object)
    The problem is that I want two switch between the frames of two sources independently. But the jit.op-object just allows two trigger a blending when i bang the left inlet. That means the second source is depending always on the first.( alphablend inhabits the same obstacle)
    Any suggestions how i could blend two sources but at the same time switch between frames of the sources independently ?
    thanks,

    • May 24 2006 | 1:00 pm
      Have your second source generate a bang when it loads or
      receives a new matrix, and route the bang message to the
      jit.op object. In general, sending a bang message to the
      left outlet of the majority of Max objects will instruct them to
      perform an operation, or perform it *again.* Here is a really
      basic Max dataflow example that you can adapt for your
      needs....
      max v2;
    • May 25 2006 | 6:22 pm
      Hey, thanks for the fast reply!
      I tried it....but probably i am just to stupid.....
      Here is my tryoutpatch, maybe you see immediately whats wrong about it.
      It still happens that when i bang the importmovie-messages 3 + 4 that
      nothing happens....
      max v2;
    • May 25 2006 | 6:47 pm
      I don't think that it's about stupidity. I just think you need
      to spend a little time with the tutorials getting a good
      sense of how Max works. Unless I completely misunderstand
      what it is you want to do (load a static image, trigger the
      jit.op object when either one of two images is loaded), you
      really don't need quite a lot of this patch at all - just the
      jit.matrix objects (and the importmovie message) feeding
      the jit.op stuff with a few well-placed buttons for bangs. I would
      REALLY suggest spending some time woodshedding
      with the tutorials - the Jitter stuff so that you understanding
      about loading matrices, and the Max stuff so that dataflow
      and controlling the order in which things are done makes
      sense to you. The good news is that it ought to become
      sort of clear pretty quickly where the trouble lies.
    • May 27 2006 | 3:35 pm
      Hmm, alright I will have another look at the tutorials.
      The weird thing is just that i found out that when I want
      to change frames seperately( and simultaneously) in jit.glue,
      there is a message called "syncinlet", and if you give it the attribute "-1", it does exactly what i would like the op-object to do. Is there any message like that for the op-object?
      O.K.,whatever, thanks.
    • May 27 2006 | 7:03 pm
      No, that's a rather specialized case; I'd expect to
      find it for that object, given what it does.
      1. Here's the deal: Nearly everything in Max/MSP/Jitter
      uses what's called "lazy evaluation." In general, unless
      you have an object specifically created to do otherwise
      (either a complete external such as pak, or a mode of
      operation such as would be the case with jit.glue), you
      bang the leftmost inlet to trigger calculations.
      The example patch I sent you a few emails back
      demonstrates how you use buttons (which send
      bang messages) to explicitly trigger outlets with
      inputs other than the left inlet. That's how you do
      it in all but a very few cases.
      2. Similarly, you can nearly always set the value of
      a left inlet *without* triggering a calculation by using
      the "set" message.
      Two really really basic principles that will serve you
      well. I can certainly understand how you could stumble
      upon a random counterexample to the way that Max
      generally works and - if you don't know how Max works
      in the general sense, which is what the tutorials are
      supposed to give you a sense of - mistake that for
      "expected" behavior (You could as easily have decided
      that everything should work like the pak object, and
      we'd be having the same discussion). So now you know.