Fredrik Ehnbom
Fredrik Ehnbom
FWIW, I don't have time to test anything right now but there's a field "divx_packed" in libavcodec/mpegvideo.h that might be possible to use to detect this.
Hmm, see I never commented on this, but I did try to get to the divx_packed property. However it was not available as the SMD codec was used for decoding...
Btw, this issue is referred to elsewhere as the [bad](http://ffmpeg.org/pipermail/ffmpeg-devel/2008-November/054385.html) b-frame [issue](http://www.avidemux.org/admWiki/doku.php?id=using:b-frames), and can be fixed in the video file by re-building the index (quick!) with correct frame type. I...
@KeyserSoze1 wrote: > I've been making changes to some of the audio interface and attempting to add some features that were originally in the boxee code. However I have been...
If you post a log, perhaps I could offer some insight based on the debug output
> I've increased the buffer length from 4k->8k > ... > 18:43:41 T:2991586192 WARNING: CIntelSMDRenderer::FlushAndSync - 0335 Frame pts is waaay behind the start time and wasn't marked as a...
Though reading the code now, I think the comparison should really be `(firstpts+add+2*DVD_TIME_BASE)/(double)DVD_TIME_BASE` as "start" will in the end be reset to the [post adjusted firstpts](https://github.com/quarnster/boxeebox-xbmc/blob/19f0c5804e7c5f4676f09bc84c690883ca6da7fa/xbmc/cores/VideoRenderers/IntelSMDRenderer.cpp#L346). I'd try this first...
The [if statement](https://github.com/quarnster/boxeebox-xbmc/blob/19f0c5804e7c5f4676f09bc84c690883ca6da7fa/xbmc/cores/VideoRenderers/IntelSMDRenderer.cpp#L379) resulting in the "wasn't marked for discontinuity" print is wrong too and also needs to be adjusted to test with `(firstpts+add)`. That one is however harmless. >...
Yeah, there's a [+90000/2 in there](https://github.com/quarnster/boxeebox-xbmc/blob/master/xbmc/cores/VideoRenderers/IntelSMDRenderer.cpp#L362), which would be almost exactly (62.469067-61.972933)*90000=44652. Don't recall the reason for it, I think it might be from earlier experiments in trying to get...
Glad I could help. Note that quite a bit in this whole process was me trial'n'erroring so if something looks strange it's quite possible it doesn't make sense and is...