VLC is not any better than mplayer. The same thing happens.
Originally Posted by VanillaMozilla
I also tried playing the video with theora_player_example and the playback is still borked.
That won't solve the problem. I already tried passing the fps option (-r) with ffmpeg using different values.
Originally Posted by no2498
I did some more testing and investigation into this issue...
When I mentioned the 15-18 fps I was wrong. It was recorded at 30 fps, however how it was really captured that is another issue because the playback shows that frames are missing (thus the playback looking like it's in fast forward mode).
$ mediainfo foo.ogg
ID : 1333266807 (0x4F780977)
Complete name : foo.ogg
Format : OGG
File size : 70.2 MiB
Duration : 2mn 28s
Overall bit rate : 3 966 Kbps
recordMyDesktop : 0.3.7.3
ID : 1770862840 (0x698D38F8)
Format : Theora
Duration : 2mn 28s
Bit rate : 3 628 Kbps
Nominal bit rate : 45.0 Kbps
Width : 1 280 pixels
Height : 1 024 pixels
Display aspect ratio : 5:4
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 0.092
Writing library : libTheora 3.2.1 (UTC 2008-10-20)
148 seconds * 30 frames = 4440 frames
$ theora_dump_video < foo.ogg > foo.dump
Encoded by Xiph.Org libTheora I 20081020 3 2 1
theora comment header:
Ogg logical stream 698d38f8 is Theora 1280x1024 30.00 fps video
Encoded frame content is 1280x1024 with 0x0 offset
So it confirms the 30 fps
Based on the playback of the capture (which includes a video being played with time) I could conclude that:
133 seconds of real world time were condensed into 101 seconds of desktop capture
So it's possible that these captures are foobared beyond repair.
Even if I could find a way to duplicate each frame that would produce the opposite effect. It would mean 133 seconds of real world being "slowed down" into 202 seconds of playback.
I would need to use decimal slowdown factor, which may not be even possible to get.
Any other ideas?