Page 1 of 1
Experimental hardware decoding?
Posted: 10 Oct 2012 00:25
by arromdee
What hardware does this contain experimental decoding for?
Re: Experimental hardware decoding?
Posted: 10 Oct 2012 03:40
by edwardw
It depends largely on your device. But it should (in theory) work on any good and recent device - e.g. the Galaxies, HTC One X, etc
Re: Experimental hardware decoding?
Posted: 10 Oct 2012 18:04
by arromdee
The device is an Idea USA CT802 tablet which has a Cortex A8 (single core). Definitely a low-end tablet. The graphics chipset is a Vivante GC800, and supposedly has hardware decoding.
http://www.fatwallet.com/forums/hot-deals/1193620/
The in_to_tree 480p test plays with noticeable skipping. The higher resolutions skip to an unwatchable degree.
Does the hardware decoding support everything that has OpenGL or does it have to support chipsets specifically?
Re: Experimental hardware decoding?
Posted: 11 Oct 2012 00:48
by edwardw
The hardware decoding is based off IOMX, the openMAX API that the built-in media player uses. So it seems to vary a bit from device to device.
Did you enable hardware decoding in the preferences?
Re: Experimental hardware decoding?
Posted: 12 Oct 2012 03:36
by arromdee
I enabled it.
VLC still skips unwatchably on a file that the built-in player ("Super HD Player") plays without problems. The file is a 480p mp4. The built-in player also plays at full speed/no skipping the in_to_tree files, even the 576p and 720p ones, while VLC skips on the 480p and the 576p and 720p are unwatchable.
Re: Experimental hardware decoding?
Posted: 12 Oct 2012 03:38
by edwardw
If you want to, you can help VLC by getting the logcat log so that we can check to see if anything is wrong.
Re: Experimental hardware decoding?
Posted: 13 Oct 2012 16:05
by Goner
After enabling hardware accelerated decoding, the video and audio get out of synch (only tried .mp4 so far) and it gets worse when I skip forward or backward by dragging the slider.
I've got an Archos G9 80 tablet ( GPU : PowerVR SGX540). No logcat files are created when this happens ...
Re: Experimental hardware decoding?
Posted: 15 Oct 2012 08:02
by arromdee
Here's the result of logcat -d (cut at the start of VLC). All I did was turn on the tablet, use VLC to play in_to_tree_576p25_2000.mp4, then do the logcat.
Incidentally, not only does VLC skip unwatchably on files that play smooth in the built-in player, it seems that checking the hardware acceleration box in VLC has no effect whatsoever on any file; if it doesn't play smoothly using only software, it skips exactly the same way with or without hardware decoding enabled.
Re: Experimental hardware decoding?
Posted: 15 Oct 2012 13:37
by edwardw
Then it means that VLC doesn't support your device's hardware decoding.
Re: Experimental hardware decoding?
Posted: 15 Oct 2012 19:03
by Jean-Baptiste Kempf
We need all the logs...
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 00:07
by arromdee
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 00:44
by edwardw
D/VLC ( 682): [0x2f7520]: main decoder looking for decoder module: 28 candidates
D/VLC ( 682): [0x2f7520]: iomx decoder fmt in:h264, out:
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.aac.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.aac
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.amrnb.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.amrnb
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.amrwb.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.amrwb
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.g711.alaw.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.g711alaw
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.g711.mlaw.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.g711mlaw
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.h263.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: video_decoder.h263
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.h264.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: video_decoder.avc
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.mp3.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.mp3
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.mpeg4.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: video_decoder.mpeg4
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.vorbis.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: audio_decoder.vorbis
D/VLC ( 682): [0x2f7520]: iomx decoder component OMX.google.vpx.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder - role: video_decoder.vpx
D/VLC ( 682): [0x2f7520]: iomx decoder found 1 matching components for role video_decoder.avc
D/VLC ( 682): [0x2f7520]: iomx decoder - OMX.google.h264.decoder
D/VLC ( 682): [0x2f7520]: iomx decoder no component could be initialised
D/VLC ( 682): [0x2f7520]: avcodec decoder trying to use direct rendering
D/VLC ( 682): [0x2f7520]: avcodec decoder allowing 1 thread(s) for decoding
D/VLC ( 682): [0x2f7520]: avcodec decoder avcodec codec (H264 - MPEG-4 AVC (part 10)) started
D/VLC ( 682): [0x2f7520]: main decoder using decoder module "avcodec"
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 00:49
by Jean-Baptiste Kempf
None of those are hardware decoders...
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 00:53
by edwardw
... and therefore your device does not support hardware decoding.
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 02:45
by arromdee
Why then does the built-in player (which says "Super HD Player" but seems to be a renamed Gallery) play 720p without any skipping when VLC cannot do 480p?
(I tried a stock Gallery for ICS and it also played them full speed.)
Edit: The tablet contains an Allwinner A10 chip which includes the Mali-400 MP. The information about the Mali-400 MP is visible in a system info app on the actual tablet. The Mali MP_400
most certainly does have hardware acceleration.
Re: Experimental hardware decoding?
Posted: 16 Oct 2012 14:01
by Jean-Baptiste Kempf
But it does not expose it to OMX
Re: Experimental hardware decoding?
Posted: 17 Oct 2012 22:02
by Razor512
for me the hardware acceleration causes the audio to lose sync with the video, other than that it works smoothly with most formats.