How do we use video codecs in mult-channel case?

zhao sun

hi,

     For multi-channel application case, how do we use the video codec provided by TI for C6678.
     1. If the number of channels is larger than that of the cores(say, if I need do 30 channels
CIF video encoding), how do we distribute the encoding jobs between these cores?
     2. How do we save the encoding context if we need do channel switch for each core.
     3. What is the layout of DSP's memory in that case. Does TI provide the solution for this
application?

     Thank you very much!

Sunzhao

over 13 years ago

0 Paula Carrillo over 13 years ago

TI__Mastermind 38820 points

Hi Sunzhao,

One option is to distribute the channels "evenly" between the cores. Each channel would have a codec instance. A good way to start is checking our Multi-channel MCSDK Video demos.

http://www.ti.com/tool/demovideo-multicore

http://processors.wiki.ti.com/index.php/MCSDK_VIDEO_2.0_Getting_Started_Guide

http://processors.wiki.ti.com/index.php/MCSDK_VIDEO_2.0_Demo_Guide

In case that the processing power of one core is not enough to run complete Encoder (or decoder) more than one core can be employed to complete the Encode/Decode. Codec partitioning uses Master-Slave(s) strategy and could be a slice-based approach (e.g., H.264 BP Encoder) or pipeline-based approach (e.g., H.264 HP Decoder) or an intermediate method. The functions necessary for multicore synchronization, memory allocation, and semaphores are supplied to the codec and the implementation of these functions is done in the framework/application.

Some documentation about Multicore programing and Keystone multicore development kit could be found at: http://www.ti.com/dsp/docs/dspsplash.tsp?contentId=145760

Thank you,

Paula

Processors

Processors forum

How do we use video codecs in mult-channel case?