Abstract:
Scope of high quality videos is not just limited to entertainment industry, but they are used widely in E-learning and health care applications. To reduce the space and bandwidth requirements of these videos MPEG standards are widely used. As the task of encoding videos in MPEG standards is computationally intensive and time consuming, it cannot be achieved in real time. In this thesis, we present parallel implementation of video encoding algorithms by using economical processing model i.e. multicore processors.
In this work we have explained how the IBM Cell B. E. and NVidia CUDA architecture can be exploited to attain a fast video encoding system. In this thesis, we also explain various approaches that can be used to parallelize the MPEG encoder and address various problems faced during implementation.
The encoder discussed using Cell B. E. is a real time MPEG encoder for a frame size up to 384 * 288. The encoding rate of the encoder is above 26 frames per seconds
