Back to Search Document details
23rd Meeting: by teleconference, July 2021 2021-06-30 23:43
AHG12: on the status of the ECM software
Authors: Y. Ye Alibaba Marta Karczewicz Qualcomm Y.-W. Huang MediaTek P. Yin Dolby D. Wang OPPO X. Wang Kwai J. Ström Ericsson F. Le Leannec InterDigital Li Zhang ByteDance Seung Hwan Kim LGE M. Hannuksela Nokia Ping Wu ZTE
Abstract
This contribution proposes to establish the ECM software branch as an independent software base decoupled from the VTM and accompanied by an algorithm description output document. It is asserted that such an arrangement would align the ECM work with the JEM work. In view of the proponents, the JEM work fostered significant progress in improving coding efficiency and broadening functionality beyond HEVC’s capability, and was instrumental to jumpstarting the successful development and finalization of the VVC standard. It is proposed that the ECM should follow a similar arrangement, such that it could serve to provide a common platform focused on coding technologies that could lead to a next generation standard beyond VVC down the road.
Meeting Notes
JVET-W0049 AHG12: on the status of the ECM software [Y. Ye (Alibaba), M. Karczewicz (Qualcomm), Y.-W. Huang (MediaTek), P. Yin (Dolby), D. Wang (OPPO), X. Wang (Kwai), J. Ström (Ericsson), F. Le Leannec (InterDigital), L. Zhang (Bytedance), S.-H. Kim (LGE), M. Hannuksela (Nokia), P. Wu (ZTE)]

This contribution proposes to establish the ECM software branch as an independent software base decoupled from the VTM and accompanied by an algorithm description output document. It is asserted that such an arrangement would follow a similar process for the ECM as was done in the past with the JEM work. In view of the proponents, the JEM work fostered significant progress in improving coding efficiency and broadening functionality beyond HEVC’s capability, and was instrumental to jumpstarting the successful development and finalization of the VVC standard. It is proposed that the ECM should follow a similar arrangement, such that it could serve to provide a common platform focused on coding technologies that could lead to a next generation standard beyond VVC down the road.

From the discussion:

It is noted that ECM or parts thereof might also end up in an extension of VVC. This cannot be foreseen at the current time.

The exploration activity is about exploring capabilities for higher compression. It is part of this activity to find out how much is possible with backward capability to VVC, without restricting to that.

Actually, ECM is already in a different repository. The change would be to establish ECM as a master branch.

ECM is based on VTM10. It is a matter of practicality not following each change in VTM. However, an important aspect would be that encoder improvements would likewise be applied to ECM and the VTM anchor compared against (which should be the most recent VTM), and obvious bug fixes found in VTM should also be fixed in ECM.

Basically, the current practice in the EE has already been somewhat like that.

It was agreed that this is an appropriate move. Also, the suggestion of establishing an algorithm description was agreed.

References:
JVET-U0048 JVET-U0100 JVET-U2024 JVET-V2017 JVET-V2024 JVET-W0102
Cited:
JVET-W0012 JVET-W0102
ARCHIVE v1 :
JVET-W0049-v1.docx
Decisions
It was agreed that this is an appropriate move. Also, the suggestion of establishing an algorithm description was agreed.
Citation