#jogamp @ irc.freenode.net - 20130816 05:06:07 (UTC)
20130816 05:06:07 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130815050607.html
20130816 05:06:07 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130816050607.html
20130816 06:50:51 * xranby (~xranby@anon) Quit (Ping timeout: 268 seconds)
20130816 06:57:19 * xranby (~xranby@anon) has joined #jogamp
20130816 07:09:20 * xranby (~xranby@anon) Quit (Ping timeout: 256 seconds)
20130816 07:22:30 * xranby (~xranby@anon) has joined #jogamp
20130816 07:37:40 * xranby (~xranby@anon) Quit (Ping timeout: 256 seconds)
20130816 07:52:26 * xranby (~xranby@anon) has joined #jogamp
20130816 12:53:52 * xranby (~xranby@anon) Quit (Ping timeout: 256 seconds)
20130816 13:07:48 * xranby (~xranby@anon) has joined #jogamp
20130816 18:19:09 <sgothel> c200045aa661cf82474c2b3c1db0ac69db40452a: GLMediaPlayer Multithreaded Decoding: GLMediaPlayer* (Part-4) - WIP
20130816 18:20:10 <sgothel> TODO is to have fine grain info _and_ control over the audio clock .. to be able to do proper seek, plus heavily A/V async frames in stream
20130816 18:20:36 <sgothel> shall I push the test data streams as mentioned somewhere ? hmm dunno about proper license
20130816 18:21:30 <sgothel> note: also have not tested w/ Android .. sure needs some regression fixes here, but I like to see all the real requirements w/ libav
20130816 18:22:14 <sgothel> if we assume to use a threaded renderer always: I guess impl. of other backends is much simpler (removal of current buggy OMX buffer handling, .. etc)
20130816 18:23:21 <sgothel> Yes, Platform.currentTimeMillis() solved a huge issue - I haven't considered .. duh!
20130816 18:35:22 <sgothel> https://www.youtube.com/watch?v=szoOsG9137U - Five-minute-sync-test
20130816 18:42:15 <sgothel> Audio-Video-Sync-Test-Calibration-23.98fps-24fps https://www.youtube.com/watch?v=cGgf_dbDMsw
20130816 18:43:01 <sgothel> sound_in_sync_test https://www.youtube.com/watch?v=O-zIZkhXNLE
20130816 18:43:23 <sgothel> anybody finding these files as originals (w/ download URLs) .. much appreciated
20130816 19:03:42 <sgothel> quite interesting .. Five-minute-sync-test saturates the audio buffers, then decoding thread waits .. causing video sync troubles (textureCount = 4)
20130816 19:04:16 <sgothel> incr. the buffers, i.e. textureCount = 10 - solves the issue (or by incr. the audio/video frame ration)
20130816 19:04:47 <sgothel> would be good to 'snoop' the packets to gather the avg. ration, so we can use a proper value for caching audio frames/buffer ..
20130816 19:04:52 <sgothel> or do all dynamically
20130816 19:05:15 <sgothel> audio-buffers are cheaper than texture buffers ofc !
20130816 19:06:55 <sgothel> this is also good indicator that threaded decoded ahead of time is the only way to go ..
20130816 19:07:20 <sgothel> (while avoiding copying data .. etc)
20130816 19:38:03 * jk4 (~jk4@anon) Quit (Ping timeout: 245 seconds)
20130816 19:38:20 * jk4 (~jk4@anon) has joined #jogamp
20130817 05:06:08 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130817050608.html