MSU Lossless Video Codec
From Wikipedia, the free encyclopedia
Template:Citesource MSU Lossless video codec is a proprietary lossless (or sometimes "near-lossless") video codec written by the MSU Graphics & Media Lab Video Group. The algorithms used in the codec have not been publicly revealed.
This codec was primarily designed for maximum compression, so it is about 5 times slower in comparison with codecs like Huffyuv. Based on claims on the MSU web site and some anonymously-posted messages on Doom9 by someone calling themselves "Tommy Carrot", it delivers the highest compression ratio among well-known lossless video coding designs, and can be used as a reference point of maximum achievable lossless compression of video. The codec appears to be more suitable for offline processing, such as video archiving and as a storage format for intermediate results from video editting.
In some modes of operation it is not actually mathematically lossless, but rather is claimed to be "near lossless" or "visually lossless" as opposed to truly mathematically reversible. After preliminary tests using OpenCV's capture API, writing and re-reading the saved video frames, the codec is able to deliver lossless compression if no colour conversion and noise reduction options are used.
Characteristics:
- Input support RGB24, RGB32, YUY2, YUYV and YV12
- Using of I-, P- and something called PP-frames
- Near lossless compression with denoising (optional)
