Suggestions for how to generate a click track sync'd to wave file


    Jul 18 2012 | 10:36 am
    I have a short wave file (length is 4 beats) that I need to play as a loop and that will be heard for about 16 bars at occasional points in the song. I'm going to have to feed our drummer with a click derived from this loop so that he can make sure that the band is always playing at the right speed for when the loop is audible.
    I know I can use sfplay~ to play the wave file and I could use the seek message to start and end it (the actual loop is a few samples less than the actual wavefile) and retrigger that seek message from the bang that's generated each time the wave file plays.
    The problem is that I need to generate a click that triggers every beat, not every 4 beats and I'm wondering how best to do that.
    The "obvious" solution that comes to my mind is to have four sfplay~ objects each of which has a seek message with 1 beat's worth of sample start/end time and then daisy chain them using the "end of play" bang that comes out (and using that bang to generate a click), but this seems quite kludgy.
    I'm wondering if anyone knows of a cleaner solution for this.
    Thanks, David

    • Jul 18 2012 | 10:38 am
      I'm not even sure the 'seek' message is good enough as it only seems to have a resolution of 1 millisecond and I would need to have sample-accurate playback to avoid clicks in between each chunk.
    • Jul 18 2012 | 11:17 am
      By the way, I looked at splitting up the original wave file into four parts which could then be loaded into buffers and played back using the [groove~] object but there does not seem to be any way to detect when [groove~] has finished playing one buffer so that the next one can be started.
    • Jul 18 2012 | 11:47 am
      How about recording your click track and then sync'n the playback of the two files.
    • Jul 18 2012 | 11:47 am
      (This is turning into a documentary on problem solving - but maybe it will be useful to someone else)
      Looks like I might be able to just have the entire loop in [groove] and then use snapshot to detect the current position and then trigger something every quarter of the way through.
    • Jul 18 2012 | 11:49 am
      caveat, i am not yet fully awake. here is what i use for click tracks when playing back from (groove).
    • Jul 18 2012 | 12:01 pm
      OK, problem solved --- I suspect this has been solved many times but here's my quick hack solution in case it's useful to anyone.
    • Jul 18 2012 | 12:09 pm
      Sorry, I didn't realize that others had replied and so I apologize for having ignored the posts (just noticed via email). @kthozoid, my sympathies, I've been up since 4:30am having realized last night that I was going to have to figure something out. Thanks for your suggestion, which I will have to study, it seems a lot more complicated than what I did so I presume it has other benefits that I don't yet understand.
    • Jul 18 2012 | 12:20 pm
      @dhj, no worries. indeed my solution is probably overcooked. it just has a few extra functionalities for metric sub-divisions and altering the click track sounds and that kind of thing.
    • Jul 18 2012 | 12:23 pm
      ok ---- I am not generating the click sound inside Max --- I will be sending MIDI events to a remote device that the drummer can hear. So I just needed to make sure that I could break up a long loop into the right number of even segments....not having much familiarity with the MSP side of Max, I wasn't aware of [groove] and more importantly [snapshot]. Once I discovered the latter, the problem became trivial for me.
    • Jul 18 2012 | 12:36 pm
      ok, then your solution will do just fine then. and here is a quick (sfplay) version just for the general purpose of completion.
    • Jul 20 2012 | 8:22 pm
      I just noticed that the groove stuff using a buffer seems to only have audio coming out one channel. The example for groove shows a single channel being split into two. How can I play a stereo sample?
      Here's my current patch
    • Jul 21 2012 | 9:13 am
      your buffer~ was mono.
    • Jul 21 2012 | 11:24 am
      Thanks, pid, I had tried that before (although I used a positive value for the duration (while being annoyed that I had to put the duration before the number of channels so glad to see one can put -1 there) and it didn't work. I'll check that again.
    • Jul 21 2012 | 12:34 pm
      This doesn't work --- unless I'm doing something really stupid (which is quite possible), there's something else wrong. I've attached the patches (and the stereo aiff file) if you're willing to take a quick look
    • Jul 21 2012 | 7:50 pm
      Works here.
    • Jul 21 2012 | 8:34 pm
      Hmmm, that's Max 6 --- I'm doing it with Max 5 right now --- I wonder if that's an issue. (But at least I know there's nothing wrong with the sample or my patcher)
    • Jul 21 2012 | 9:12 pm
      Try this...