Forum Discussion
New compression format for VR?
Does anyone know if a new compression format is in development for VR?
The reason I ask is both the left and right frames have a great deal of overlapping data. For example, the distant background in a scene would be identical, with only the foreground images having any distinguishable differences.
Having some sort of mirroring attribute to the data, based on sections of the video, could significantly reduce the file size, possibly upwards of 40%.
Or, maybe I just don't know what I'm talking about.
The reason I ask is both the left and right frames have a great deal of overlapping data. For example, the distant background in a scene would be identical, with only the foreground images having any distinguishable differences.
Having some sort of mirroring attribute to the data, based on sections of the video, could significantly reduce the file size, possibly upwards of 40%.
Or, maybe I just don't know what I'm talking about.
2 Replies
- I looked at this a bit a couple of years ago and at the time there was a standard but no encoder/decoder support yet. Maybe this is still true.
http://en.wikipedia.org/wiki/Multiview_Video_Coding
It does seem relatively straight forward though. I wonder if just doing a simple wrapper that splits the view, places them sequentially, and does a normal MPEG video compression would work almost as well? - Frame sequential stereo is what is used for 3D videos on the PlayStation Store for the PS3. Someone has reversed engineered the format for playback on the PS3 : https://www.doom9.org/showthread.php?t=170863
From what I remember the compression was much better than SBS but a little bit less than MVC.
