Thu Nov 27 13:25:56 2008, comment #14:
As of revno 10365 the slides work, and seem to be synchronized with the audio.
Stop and start work for the slides, not for the audio (also seems to cover the problem of the audio starting automatically). Seek doesn't work for either.
|
Thu Nov 27 08:37:10 2008, comment #13:
At time of ::stop() call, no SoundInputStream is plugged yet
(as sound isn't realized yet) so ::stop() is a no-op.
I guess we may use the PlayHead class in Sound too..
|
Thu Nov 27 08:33:32 2008, comment #12:
Ah, now I see.
There's a Sound.stop() call.
Then Sound.start(0) is called again when you click the play button.
|
Thu Nov 27 08:29:15 2008, comment #11:
The sound that starts is:
http://downloads.bbc.co.uk/news/nol/shared/spl/hi/audio_slideshow/kenadamptw/content/audio_hi.mp3
It is loaded using Sound.loadSound(<url>, 1).
Second argument means 'streaming'.
When streaming, the sound starts while being loaded as per documentation. Might need some buffering, but we saw _soundbuftime is set to zero.
|
Thu Nov 27 08:27:51 2008, comment #10:
Actually:
UNIMPLEMENTED: MovieClip._soundbuftime=[number:0]
.. doesn't seem related to the start then
|
Thu Nov 27 08:24:27 2008, comment #9:
UNIMPLEMENTED: MovieClip._soundbuftime setting
^^ could be related to the premature start
|
Thu Nov 27 08:09:15 2008, comment #8:
in Gnash, sound starts immediately on load, with pp it takes an hit of 'play' button for that to happen.
|
Thu Nov 27 07:46:49 2008, comment #7:
That number was AV_NOPTS_VALUE, now taken as zero (still not correct, but reduces frames container management)
|
Thu Nov 27 07:22:56 2008, comment #6:
This is interesting:
DEBUG: 0-to-last element has timestamp 18344262162188942992 > 51878
That's with FFMPEG parsing the mp3.
The firsit timestamp is that huge number:
FFMPEG packet dts:-9223372036854775808, pts:-9223372036854775808
|
Wed Nov 26 21:11:29 2008, comment #5:
The movie contains traces, traces from pp and gnash don't match.
In Gnash, we get more 'loading ##' traces, pp stops at 6, gnash keeps going, no idea why.
Also, there's a mismatch in order between 'loading 4' and 'PRE-LOADED SECTION' (could be one of those action execution order things)
|
Tue Nov 18 16:20:43 2008, comment #4:
No more a blocker, as we handle the exception now...
|
Wed Oct 29 11:19:10 2008, comment #3:
marking to blocker for the unhandled exception
|
Wed Oct 29 11:09:45 2008, comment #2:
Implementing Sound.getBytesTotal() and Sound.getBytesLoaded() does make the SWF get off the loading screen.
The sound should not start playing immediately (it does), but rather wait for the play button to be pressed.
The slider at the bottom doesn't show the correct position (getPosition?), and the slideshow images don't change.
|
Wed Oct 29 10:59:50 2008, comment #1:
And I get
terminate called after throwing an instance of 'gnash::SoundException'
what(): Kann Aux-SDL-Audioausgabe nicht starten: No available audio device
when the audio device isn't available.
|
Wed Oct 29 10:55:33 2008, original submission:
The BBC slideshow SWF, an example here:
http://downloads.bbc.co.uk/news/nol/shared/spl/hi/audio_slideshow/kenadamptw/slideshow_629.swf
gets stuck on the loading frame, though the sound plays.
There are three UNIMPLEMENTEDs: Sound.getBytesLoaded(), Sound.getBytesTotal() and MovieClip._soundbuftime. The first two seem to be implemented in the Gnash Sound class, but not passed on to actionscript. MovieClip._soundbuftime just isn't implemented. "Fixing" the first two unfortunately doesn't fix the movie.
(this SWF is used a lot on the BBC site)
|