[jmsl] Audiosample playback & Synchronization

jmsl at music.columbia.edu jmsl at music.columbia.edu
Mon Jul 12 16:38:28 EDT 2004

Hi Nick. Thanks a lot for all the pointers. They made a big difference.

I have a couple more questions. I am thinking of doing some optimization of 
the playback. For example, where I do sample playing,
if (sample != null) {
         int jsynTimeToPlay = (int) JMSL.clock.timeToNative(timeToPlay); 
//tell component to play at right time?
         try {
                 //clear the previous samples out first...
                 //now play
                 sampleEnvelope.envelopePort.queue(jsynTimeToPlay, envData);
                 samplePlayer.samplePort.queue(jsynTimeToPlay, sample);
                 } catch (SynthException e) {
if the same sample is being played repeatedly as will happen in most 
instances, the playback frequency might not permit the sample to be played 
to exhaustion. Is it possible in JMSL to pause or stop the samplePlayer 
(and envelope), reset the sample frame index being played to 0 and restart 
the samplePlayer? Even more pertinent, would that be a more efficient 
implementation than simply clearing the samplePlayer every time?

Can you also give me a clear idea of what the MusicJob halt(), 
advanceCurrentTime() and finish() methods behave? I could not find any 
construct for halting (or pausing) an MJ at a specific time. Can a MusicJob 
be put on hold and restarted?


PS does the synth usage include jmsl usage? Since the latter relies on 
event-scheduling, having a lot of MusicJobs as I did in my testcases might 
be held responsible.

At 10:39 PM 7/10/2004, you wrote:
>Ok sounds like it's working out.
>I would also comment out the flash and see how things feel when you move 
>the gui around. Creating new FlashJobs every repeat is not really an 
>especially bad thing.  Try doing this without calling the actual flash 
>command and see how things behave.
>Curious what the speed of your machine is.  Sample playback is very 
>efficient in JSyn so I wonder what's driving it up to 12%
>jmsl at music.columbia.edu wrote:
>>Thanks Nick!!!
>>On adding the setAdvance, I noticed less GUI interference as per the test 
>>I mentioned. There was a marked improvement in the output with 
>>ParallelCollections but only for usages of less than 4%.
>>The reason I was re-launching each repeat was that I wanted the display 
>>to hold a flash color for some time, while the sample is being played. I 
>>have changed it to recreate a FlashJob each repeat but I will have to 
>>change that to reduce overhead in the full application.
>>Usage shows about 12%. We can consider that an average. My application 
>>needs to be scalable up to around 40% usage, if possible. Thanks again.
>>- Bosun

More information about the jmsl mailing list