H.264 encoder speed-up via joint algorithm/code-level optimization

Yu-Lun Lai; Yu-Yuan Tseng; Chia-Wen Lin; Zhi Zhou; Ming-Ting Sun

doi:10.1117/12.631570

24 June 2005 H.264 encoder speed-up via joint algorithm/code-level optimization

Yu-Lun Lai, Yu-Yuan Tseng, Chia-Wen Lin, Zhi Zhou, Ming-Ting Sun

Proceedings Volume 5960, Visual Communications and Image Processing 2005; 596038 (2005) https://doi.org/10.1117/12.631570
Event: Visual Communications and Image Processing 2005, 2005, Beijing, China

Abstract

The outstanding coding performance of H.264 comes with the cost of significantly higher complexity, making it too complex to be applied widely. This work aims at accelerating the H.264 encoder using joint algorithm/code-level optimization techniques so as to make it feasible to perform real-time encoding on a commercial personal computer. We propose a fast inter-mode decision scheme based on spatio-temporal information of neighboring macroblocks for the algorithm-level optimization. We use a commercial profiling tool to identify most time consuming modules and then apply several code-level optimization techniques, including frame-memory rearrangement, single-instruction-multipledata (SIMD) implementations based on the Intel MMX/SSE2 instruction sets. Search mode reordering and early termination for variable block-size motion estimation, are then applied to speed up these time-critical modules. The simulation results show that our proposed joint optimization H.264 encoder achieves a speed-up factor of up to 18 compared to the reference encoder without introducing serious quality degradation.

Citation Download Citation

Yu-Lun Lai, Yu-Yuan Tseng, Chia-Wen Lin, Zhi Zhou, and Ming-Ting Sun "H.264 encoder speed-up via joint algorithm/code-level optimization", Proc. SPIE 5960, Visual Communications and Image Processing 2005, 596038 (24 June 2005); https://doi.org/10.1117/12.631570

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available