Video image compression using unequal weights
Techniques for video image compression in a video system involve providing a sequence of frames including picture regions, and determining first unequal weights, in which the first unequal weights are used to determine pixel luminance values of a picture region of at least one of the frames. The techniques involve determining second unequal weights, where the second unequal weights are used to determine pixel chroma values of the picture region of at least the one frame. The first unequal weights can include AC weights, and the second unequal weights can include DC weights. A representation of pixel luminance values can differ from a representation of pixel chroma values. The techniques for determining the first unequal weights can involve accessing representative transformations for the pixel luminance values, and the techniques for determining the second unequal weights can involve accessing representative transformations for the pixel chroma values.
1. A decoding method using a processor, the method comprising:
accessing a sequence of predicted and bidirectional predicted frames each comprising pixel values arranged in macroblocks,
wherein at least one macroblock within a bidirectional predicted frame is determined using direct mode prediction based on motion vectors from two or more referenceable frames,
wherein at least one such motion vector is scaled by a frame scale fraction, the frame scale fraction comprising a numerator and a denominator, wherein the numerator of the frame scale fraction is proportional to a first frame distance from the bidirectional predicted frame to one of the referenceable frames, and
wherein two or more of the referenceable frames are subsequent to the bidirectional predicted frame in display order.
2. The method of claim 1 , wherein one of the referenceable frames is a predicted frame.
3. The method of claim 1 , wherein the frame scale fraction is smaller than −1.
4. The method of claim 1 , wherein the frame scale fraction is larger than 1.
5. The method of claim 1 , wherein the denominator of the frame scale fraction is proportional to a second frame distance between a first referenceable frame and a second referenceable frame.