I must say that this literally left me speechless... what an amazing development!
I wonder why you don't use only ffmpeg to encode the 3d video. Googling a little bit I found that the most common procedure is: 1 - Converting the videos to a lossless image sequence (.png, for instance) 2 - Interleaving the images with some other tool (don't know if this can be done with ffmpeg) 3 - Re-encode the resulting sequence with ffmpeg I guess your goal is different because for anaglyph images there's not a real interleaving, but a hue/opacity change plus an overlay... Have you tried to do that in gstreamer? Since you've got a pretty powerful CA, it seems quite easy to "ink" the videos and making them semi-transparent (ffmpegcolorspace, maybe?), overlay one on the other, and encode the resulting video... Does this seem feasible or is it Sci-Fi? Now the gstreamer composer does sound like a good idea... I'm looking forward to hearing from your improvements on this work! Best regards, Rubén P.S: Markus, sorry for the repetition, in case you get this email twice --my fault! 2011/4/1 Markus Ketterl <[email protected]> > Hi list, > we here in Osnabrück, just want to announce that we managed to improve > Matterhorn so that we are now able to do 3D lecture recordings in selected > classrooms. Our physics department demanded that they would need 3D fro > their experiments so that the students get the full classroom experience. > Some of these experiments would be very hard to follow without the depth > information. > > In our setup we added a Blackmagic Intensity Pro with the some third party > gstreamer bindings to the capture agent. We had to take a quite fast > quad-core machine to encode the HDMI signal from the Panasonic HDC-SDT750 > camera.The recordings look just fine when we look at the raw material on the > capture agent. > > Currently we still have some small problems in the processing: > 1. we spilt the 3D-video-stream in two separate image sequences > 2. we mix the images together with imagemagick. We ink the separate > channels for anaglyph images (for red-green glasses). > 3. we re-create the videostream from the new images. > > The processing unfortunately takes "forever" and we really need to speed > this up. We are open for suggestions if you have further ideas which tools > could improve the process. > > On the engage side there was nothing to do. The students just need the > red-green glasses to view the videos. But we are currently considering to > add a 3D-skin for the player, so that the buttons seem to "fly" above the > rest of the interface. > > After some more testing we will add the 3D processing module to the contrib > section. > > Markus > _______________________________________________ > Matterhorn mailing list > [email protected] > http://lists.opencastproject.org/mailman/listinfo/matterhorn > > > To unsubscribe please email > [email protected] > _______________________________________________ >
_______________________________________________ Matterhorn mailing list [email protected] http://lists.opencastproject.org/mailman/listinfo/matterhorn To unsubscribe please email [email protected] _______________________________________________
