Chroma quantization in video coding
摘要
說明書
The process starts when it receives (at 1710) a video picture. This video picture can be a raw image in an uncompressed video stream or a decoded picture from a compressed video bitstream. The process then identifies (at 1720) a region in the picture that shares common characteristics that make it suitable for the coding units in the region to share a common set of chroma QP offset values. For example, some embodiments identify a region that is originally coded in 4:2:0 format as a region in which higher QP values (and hence positive chroma QP offset) can be used to reduce bitrate. Conversely, some embodiments identify a region that is natively coded in 4:4:4 format as a region in which lower QP values (and hence lower or negative QP offset values) are needed to maintain quality.
The process then analyzes (at 1730) each region's spatial-temporal characteristics. In some embodiments, this analysis includes an analysis of the region's texture/variance/activity, format (4:4:4 or 4:2:0, etc.), noise, motion, bit-depth, or other characteristics that may affect the relationship between luma and chroma as well as between the two chroma components.
