Back to Search Document details
12th Meeting: Macao, October 2018 2018-10-04 05:02
CE5: Per-context CABAC initialization with single window (Test 5.1.7)
Authors: Amir Said Qualcomm Jie Dong Qualcomm H. Egilmez Qualcomm Yung-Hsuan Chao Qualcomm Marta Karczewicz Qualcomm Vadim Seregin Qualcomm
Abstract
This contribution reports the simulation results for Test CE5.1.7. In this test, the CABAC engine uses a single window to estimate the binary symbol probabilities, where the window parameters are different for different contexts and are stored in a pre-defined look-up table (LUT). Under the CE5 test conditions [1], the average coding gains are 0.07% AI, -0.03% RA, and -0.02% LDB. Compared with BMS 2.0.1 VTM configuration, the average gains are -0.85% AI, -0.54% RA, -0.63% LDB.
Meeting Notes
JVET-L0116 CE5: Per-context CABAC initialization with single window (Test 5.1.7) [A. Said, J. Dong, H. Egilmez, Y.-H. Chao, M. Karczewicz, V. Seregin (Qualcomm)]
References:
JVET-K0385 JVET-K1025
Cited:
JVET-L0025 JVET-L1025 JVET-M0025 JVET-M0413 JVET-M0759
ARCHIVE v2 :
JVET-L0116-CE5.1.7 (512x64 table) vs. BMS2.0.1_VTMCfg.xlsm JVET-L0116-CE5.1.7 (512x64 table) vs. CEM1.xlsm JVET-L0116-CE5.1.7 (8x8 table) vs. BMS2.0.1_VTMCfg.xlsm JVET-L0116-CE5.1.7 (8x8 table) vs. CEM1.xlsm JVET-L0116-v1.docx JVET-L0116-v2.docx
Decisions
JVET-L0116 CE5: Per-context CABAC initialization with single window (Test 5.1.7) [A. Said, J. Dong, H. Egilmez, Y.-H. Chao, M. Karczewicz, V. Seregin (Qualcomm)]
Citation