FLV 中Video Tag Data 结构之Sorenson H.263编码方式
quoted:http://en.wikipedia.org/wiki/Sorenson_Media
Sorenson Media was established in December 1995 to address market demand for rich online media content by developing innovative, cost-effective video encoding technology that significantly reduced bandwidth requirements while preserving the highest video quality. Originally called Sorenson Vision, the company developed technology licensed and ultimately acquired from Utah State University. The company first unveiled its “codec” (compression and decompression tool) at a developer’s preview at MacWorld Expo in January 1997.
quoted: http://en.wikipedia.org/wiki/Sorenson_codec
Sorenson codec may refer to either of three proprietary video codecs: Sorenson Video, Sorenson Video 3 or Sorenson Spark. Sorenson Video is also known asSorenson Video Codec, Sorenson Video Quantizer or SVQ. Sorenson Spark is also known as Sorenson H.263 or FLV1 (and it is sometimes incorrectly named asFlash Video (FLV), which is the name of Adobe Flash container format).
Both codecs were devised by Sorenson Media Inc. (formerly Sorenson Vision Inc.). Sorenson Video is used in Apple's QuickTime and Sorenson Spark in Adobe Flash(formerly Macromedia Flash).
--------------------------------------------------------------------------------
Sorenson Spark is an older codec but it is also a widely available and compatible one, when used in Flash Video.[10] According to an Adobe engineer Tinic Uro, it is considered as an incomplete implementation of H.263.[9][11]
-----------------------------------------------------------------------------
quoted: http://xidianzhangjun.blog.163.com/blog/static/11548877120109131010127/
Summary of Differences from H.263 (again Sorenson Spark)
The following H.263 features have been removed from the Sorenson H.263 video format:
GOB (group of blocks) layer
Split-screen indicator
Document camera indicator
Picture freeze release
Syntax-based arithmetic coding
PB-Frames
Continuous presence multipoint
Overlapped block motion compensation
The following non-H.263 features have been added to the Sorenson H.263 video format:
Disposable frames (difference frames with no future dependencies)
Arbitrary picture width and height up to 65535 pixels
Unrestricted motion vector support is always on
A deblocking flag is available to suggest the use of a deblocking filter
In order to support these differences, the Sorenson H.263 video format uses different headers
than H.263 at both the Picture Layer and the Macroblock Layer. The GOB Layer is absent.
Two versions of the Sorenson H.263 video format are defined. In version 0, the Block Layer is
identical to H.263. In version 1, escape codes in Transform Coefficients are encoded differently
than in H.263. There are no other differences between version 0 and version 1.