Tuesday, February 8, 2011

SIF1

Up to date video codec SIF is widely used by the customers. Despite of its ?youth? it has proved itself in the best way during tests. Codec with the SIF-converting algorithm was created in 2000 and afterwards was named SIF1 codec. This codec is being continuously improved and currently there are no errors in video compression programs. It is available in all programs for video operations and it can be easily installed in such software environment as DivX. Such parameter as decision value is extremely important for the video codec. The more is the value of this parameter the lower is the image quality due to high level of compression. That?s why the best value for this parameter is 30. Regarding the shots, the codec fixes the grounds automatically and inserts key shots exactly as needed. According to the test results with the decision value of 30 the codec is compressing 798kb of video in 4:29 min:sec under the processor upload <100% during the viewing.
 
The peculiarities of this codec are as follows:
 
- low sensitivity to noise of incoming image;
- low rate of decompression artifacts;
- high flexibility of compression algorithms, in addition there is an opportunity to choose the compression parameter even for the image size of 2x2pcs;
- there are special requirements to the processor, advisable frequency for the comfortable viewing of the films coded by this codec is more that 1500 MHz;
- the basic set of commands MMX is set in the codec;
- the size of compressed image has to be 16-fold horizontally and vertically.
 
One of the lacks of SIF- converting is high requirement to the capacity of system memory and to the processor cash size. Continuously improving methods of algorithm optimization still are not able to eliminate this problem.
 
Each updated SIF1 version underscores the efficiency of the dissective part of Core6. The algorithm of the digitizer operation is being changed all the time. The vividness of the decoded image and total efficiency of the compression are being improved, and the operation in low resolution is getting better. Psycho-visual model is greatly improved in the latest version. In the course of time the developer will implement SSE2 support, which affects the processing speed. SIF1 and its improved converting algorithm can easily take the lead among the other programs, and moreover it can displace DivX and XVid in the nearest future.