| Paper: | WP-L3.5 |
| Session: | Video Coding III |
| Time: | Wednesday, September 19, 16:10 - 16:30 |
| Presentation: |
Lecture
|
| Title: |
LOW-DRIFT FIXED-POINT 8X8 IDCT APPROXIMATION WITH 8-BIT TRANSFORM FACTORS |
| Authors: |
Yuriy Reznik; QUALCOMM Incorporated | | |
| | De Hsu; QUALCOMM Incorporated | | |
| | Prasanjit Panda; QUALCOMM Incorporated | | |
| | Brijesh Pillai; QUALCOMM Incorporated | | |
| Abstract: |
This paper presents an efficient algorithm for computing the Inverse Discrete Cosine Transform (IDCT) for image and video coding applications. This algorithm was derived by converting an 8-point IDCT factorization of C. Loeffler, A. Ligtenberg, and G. S. Moschytz (LLM) into a scaled form, leaving only 8 multiplications by irrational factors inside the transform. The key advantage of such a modification is that these factors can be sufficiently accurately represented by 8-bit integer values, resulting in a very small dynamic range of intermediate values inside the transform. The 8-point scaled transform can be implemented either by using 8 multiplications, 26 additions and 6 shifts or (in a multiplier-less fashion) by using 44 additions and 18 shifts. This implementation complies with both IEEE 1180 standard and the new MPEG IDCT precision specification (ISO/IEC 23002-1), and shows remarkably low drift in decoding of H.263, MPEG-2, and MPEG-4 SP bitstreams produced by reference encoders (employing double-precision floating-point DCTs and IDCTs). |
