You could try Tom Butterworth's and my alternative video player: http://v002.info/downloads/beta/v002%20MoviePlayer%202.0%20Beta%203.zip
We use an out of process background player to get around 32/64 bit quicktime issues. That said, there will still be a delay in loading a new movie, but it wont (should not) bring QC to its knees. Also, thanks very much Chris for your reply. I will look into it :) On May 17, 2011, at 3:10 PM, Anatol Ulrich wrote: > On the topic of multithreading: I'm currently experiencing total composition > block when loading a new video file (via the "movie importer" patch). > This is quite ugly for the kind of live performance I'm doing ... I've tried > being semi-clever about it and render the next set of movies onto hidden > sprites, but I'm only shifting the blockage around. At some point in time it > *does* happen. > Is there anything I can do > - with the builtin set of patches > - with some 3rd party movie player I'm not aware of? > (If necessary I'd be willing to converty my footage to anything ffmpeg and > friends can spit out) > > Cheers, > Anatol > > > On 17.5.11 8:27 PM, Christopher Wright wrote: >>> I would like to move from a polling (CVDisplayLink) to a event driven >>> rendering model in my app. I would like to use GCD and named dispatch >>> queues, but I am concerned about QA 1538 (Multithreaded QCRenderer) : >>> http://developer.apple.com/library/mac/#qa/qa1538/_index.html >>> >>> Specifically, if I have a named dispatch queue, and alloc/init, and render >>> content in this queue only, is that safe and guideline following? My >>> understanding is, a queue can bounce around various thread pools, and wont, >>> technically, be on the same literal thread. Thus my question. >>> >>> What say ye? Yay or Nay? >> >> This will be a problem under the current implementation - internally some >> resources are tied to pthreads (via pthread_set_specific), and that doesn't >> work so well under GCD, where, as you say, queues can migrate threads. >> >> One thing you could do is set up a thread for the renderer that waits on a >> dispatch_semaphore, and then in the GCD block signal the semaphore when you >> want it to render. This will add some latency, but it'll remove the polling >> model without violating the thread guarantees QC currently expects. >> >> -- >> Christopher Wright >> [email protected] >> >> >> >> _______________________________________________ >> Do not post admin requests to the list. They will be ignored. >> Quartzcomposer-dev mailing list ([email protected]) >> Help/Unsubscribe/Update your Subscription: >> http://lists.apple.com/mailman/options/quartzcomposer-dev/quartzcomposer-dev%40mail.taugt.net >> >> This email sent to [email protected] >> > > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > Quartzcomposer-dev mailing list ([email protected]) > Help/Unsubscribe/Update your Subscription: > http://lists.apple.com/mailman/options/quartzcomposer-dev/doktorp%40mac.com > > This email sent to [email protected]
_______________________________________________ Do not post admin requests to the list. They will be ignored. Quartzcomposer-dev mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/quartzcomposer-dev/archive%40mail-archive.com This email sent to [email protected]
