公开(公告)号：US20050262276A1

公开(公告)日：2005-11-24

申请号：US11126556

申请日：2005-05-11

申请人： Kismat Singh ,  Murali Muthukrishnan ,  Sriram Sethuraman ,  Sankaranarayanan Parameswaran ,  Bhavani Rao

发明人： Kismat Singh ,  Murali Muthukrishnan ,  Sriram Sethuraman ,  Sankaranarayanan Parameswaran ,  Bhavani Rao

IPC分类号： G06F13/28

CPC分类号： G06F13/28

摘要： A design method for implementing a high-memory algorithm for motion estimation and compensation uses a low internal memory processor and a DMA engine that interacts with the processor and the algorithm. The DMA takes care of large data transfers from an external memory to the processor internal memory and vice-versa, without using the CPU clock cycles. The design method is scalable and is suited to handle huge bandwidths without slowing down the processor. To prevent the processor from being idle during DMA, the processing is pipelined and staggered so that motion compensation is performed on an earlier block or data that is available, while DMA fetches the reference data for the current block. Several DMAs may be set up under an ISR if necessary. The invention has application in video decoders including those conforming to H.264, VC-1, and MPEG-4 ASP.

摘要翻译： 用于实现用于运动估计和补偿的高存储器算法的设计方法使用与处理器和算法交互的低内部存储器处理器和DMA引擎。 DMA不需要使用CPU时钟周期来处理从外部存储器到处理器内部存储器的大量数据传输，反之亦然。 该设计方法是可扩展的，适合处理巨大的带宽，而不会降低处理器的速度。 为了防止处理器在DMA期间处于空闲状态，处理流水线和交错，以便在较早的块或可用数据上执行运动补偿，而DMA获取当前块的参考数据。 如果需要，可以根据ISR设置几个DMA。 本发明适用于包括符合H.264，VC-1和MPEG-4 ASP的视频解码器。

公开(公告)号：US08189663B2

公开(公告)日：2012-05-29

申请号：US12004369

申请日：2007-12-20

申请人： Kismat Singh ,  Philip Graham White ,  Matthew Patrick Compton

发明人： Kismat Singh ,  Philip Graham White ,  Matthew Patrick Compton

IPC分类号： H04N7/12 ,  H04N11/02 ,  H04N11/04

CPC分类号： H04N19/48 ,  H04N19/126 ,  H04N19/14 ,  H04N19/149 ,  H04N19/15 ,  H04N19/176

摘要： A method of obtaining a quantization value for macroblocks of a picture for video compression includes estimating 11, for a plurality of quantization values, a size of the picture in terms of a sum of all quantized coded coefficients. A linear relationship between a size of a macroblock and Activity of the macroblocks is assumed 12, where the Activity of the macroblock is the number of coded coefficients in the macroblock before discarding any coefficients or quantizing the remaining coefficients. The macroblocks are divided 13 into groups by their Activity level such that substantially a same number of pre-quantization coefficients are in each group. An average Activity of each group is determined 14. A quantization value for each group is determined 15 from the assumed linear relationship between a number of bits per macroblock and Activity of the macroblocks and the average Activity of the group, as imposed by a predetermined bit rate.

摘要翻译： 获得用于视频压缩的图像的宏块的量化值的方法包括：对于多个量化值，估计11中的所有量化编码系数之和的图像大小。 假设宏块的大小和宏块的活动之间的线性关系为12，其中宏块的活动是在丢弃任何系数或量化剩余系数之前的宏块中的编码系数的数量。 宏块通过其活动级别被分成13组，使得在每组中基本上相同数量的前量化系数。 确定每个组的平均活动14.根据由宏块的位数和宏块的活动与组的平均活动度之间的假设线性关系确定每个组的量化值，如由预定位 率。

公开(公告)号：US20080165845A1

公开(公告)日：2008-07-10

申请号：US12004369

申请日：2007-12-20

申请人： Kismat Singh ,  Philip Graham White ,  Matthew Patrick Compton

发明人： Kismat Singh ,  Philip Graham White ,  Matthew Patrick Compton

IPC分类号： H04N7/26

CPC分类号： H04N19/48 ,  H04N19/126 ,  H04N19/14 ,  H04N19/149 ,  H04N19/15 ,  H04N19/176

摘要： A method of obtaining a quantisation value for macroblocks of a picture for video compression includes estimating 11, for a plurality of quantisation values, a size of the picture in terms of a sum of all quantised coded coefficients. A linear relationship between a size of a macroblock and Activity of the macroblocks is assumed 12, where the Activity of the macroblock is the number of coded coefficients in the macroblock before discarding any coefficients or quantising the remaining coefficients. The macroblocks are divided 13 into groups by their Activity level such that substantially a same number of pre-quantisation coefficients are in each group. An average Activity of each group is determined 14. A quantisation value for each group is determined 15 from the assumed linear relationship between a number of bits per macroblock and Activity of the macroblocks and the average Activity of the group, as imposed by a predetermined bit rate.

摘要翻译： 获得用于视频压缩的图像的宏块的量化值的方法包括：对于多个量化值，估计11，根据所有量化的编码系数的和来估计图像的大小。 假设宏块的大小和宏块的活动之间的线性关系为12，其中宏块的活动是在丢弃任何系数或量化剩余系数之前的宏块中的编码系数的数量。 宏块通过其活动级别被分成13组，使得每组中基本上相同数量的预量化系数。 确定每组的平均活动14。 从每个宏块的位数和宏块的活动与组的平均活动度之间的假设线性关系确定每个组的量化值，如预定比特率所施加的。