Method and system for transmitting video codec data between digital signal processors

Abstract

The invention discloses a method and a system for transmitting video codec data between digital signal processors. The method and the system can effectively solve the problems of low resource occupation and low efficiency caused by the fact that a plurality of DSPs share a shared bus and a shared storage device by transmitting interactive data between DSPs through GE interfaces of the DSPs, thereby improving the video data parallel processing capability and efficiency of a plurality of DSPs. Meanwhile, the shared bus and the shared storage device for data interaction in the prior art are obviated by using the GE interfaces of the DSPs for data interaction, so that data interaction for video processing can be achieved without any extra hardware resources, thereby improving the data interaction speed and saving the hardware cost.

Description

technical field [0001] The present invention relates to digital signal processing technology, in particular to a method and system for transmitting video codec data between digital signal processors (DSP, Digital Signal Processor). Background technique [0002] As people have higher and higher requirements for video quality, the importance of real-time encoding and decoding of high-definition (720p, 1080p) image formats has become increasingly apparent. However, at present, the mainstream video processing DSP in the industry cannot independently complete the full frame rate video encoding and decoding of high-definition resolution. Generally, multiple DSPs are used to encode and decode one frame of image in parallel. In this way, it is inevitable to exchange a large amount of data between multiple DSPs. [0003] For example, motion estimation algorithms in video coding. Motion estimation algorithm is an important step to eliminate time-domain redundancy in the encoding pr...

AI Technical Summary

Problems solved by technology

However, since multiple DSPs perform encoding and decoding algorithms in parallel, that is, the read and write operations of shared memory are concurrent.

Therefore, the problem of occupying shared resources will inevitably occur, and the shared bus and shared storage devices are unavailable, resulting in reduced access efficiency of the shared bus and shared storage devices, which is not conducive to the efficient execution of video encoding and decoding.

Especially when one DSP occupies the bus for too long and the requests of other DSPs cannot be responded to, it is easy to cause deadlock and cause the entire system to crash.

Images