SheerVideo was initially developed Andreas Wittenstein at BitJazz Inc. in July 2002 by trimming all the fat from BitJazz's slower but more powerful codec, PhotoJazz.
Currently, SheerVideo contains specific codecs for real-time lossless encoding and decoding of RGB[A] data (for digitized film and CGI) as well as Y'CbCr[A] data (for video); with or without an alpha (A) channel for compositing; at a precision of 10 or 8 bits per channel; fully sampled (4:4:4) or 2:1 chroma-subsampled (4:2:2); and progressive-scan (p) or interlaced (i):
A unique feature of SheerVideo is the Synchromy lossless color-conversion engine which lets it convert between RGB and Y'CbCr color spaces with zero loss, or, when the output format lacks sufficient precision, with the maximum theoretically possible accuracy. This feature is useful for applying CGI tools to video footage, applying video tools to CGI and digitized film, and compositing video material with CGI and digitized film material. With the help of Synchromy, each of the specific Sheer codecs can input and output all supported uncompressed pixel formats - currently a couple dozen. The trade-off for this accuracy is that compression and decompression paths requiring color-space conversion are generally slower than real-time.
In addition, there is a generic Sheer codec which automatically chooses the appropriate Sheer format for the source pixel format:
For backwards compatibility, SheerVideo QT also retains a legacy codec for obsolete wide-range video data formats: