H263 encoder

The encoder takes as inputs the original video sequence in YUV format and the resolution of the video, which can be any of the standard resolutions or a custom one. The start and the end of frames to be encoded may be given. The size of the search window to detect the motion vectors can also be controlled. The quantization parameter of P encoded frames may be fixed to control the output bit rate. However, the encoder supports several types of bit rate control methods: variable and constant bit rates. Thus, some of them override the quantization parameter option. It is known that the common frame sequence of H263 encoding is IPPPPP..., where the only frame encoded as I is the first one and all the others are encoded as P frames. The size of I frame may be five times large as that of P frame, hence to provide smoothing of the output bit rate, the encoder has the ability to vary the quantization parameter of the first I frame separately. H263 behaves, by default, as a one layer codec, however, it supports an extra enhancement layer (with the ability of choosing its output bit rate and other options). The encoder takes as well the frame rate of the original video sequence as input. The output frame rate can be controlled by allowing the encoder to skip some images between each two encoded images. For example, if the input sequence is encoded at 30 frames/s, by skipping one image between each two encoded images, the output frame rate will be 15 frames/s. ITU recommendation for H263 [69] defines some basic operations that all the implementations should support. In addition, some other functionalities are defined separately in annexes as options. The encoder that we deal with supports the following optional functionalities. These functionalities may be activated or not as needed. When activated some other options may be given to further fine tune the operation of the encoder. Technical specifications and details about these options can be found in the H263 standard [69]. $\bullet$

• The use of unrestricted motion vector mode as defined in annex D.
• The use of syntax-based arithmetic coding as defined in annex E.
• The use of advanced prediction mode as defined in annex F.
• The use of PB-frames as defined in annex G. In this case, the ability of controlling the quantization parameter of the B frame is given.
• The use of improved PB-frames as defined in annex M.
• The use of advanced Intra coding mode as defined in annex I.
• The use of deblocking filter as defined in annex J.
• The ability of controlling the number of true B frames between P frames as defined in annex O. In this case, the quantization parameter of the true B frame may be varied.
• The ability to select SNR or spatial scalability mode as defined in annex O.
• When using the enhancement layer, in addition, the quantization parameter may be controlled as defined in annex O.
• The use of alternative inter vlc mode as defined in annex S.
• The use of modified quantization mode as defined in annex T.
• It is possible to make the picture sequence like that of MPEG codec, by forcing I frame every n Frames.

In page , we show a list of the command-line options of the encoder.