Markdown Version | Session Recording
Session Date/Time: 28 May 2024 19:00
CELLAR
Date: May 2024 Attendees: Jerome, Michael (Chair), Martin, Steve (Chair)
Summary
The May 2024 CELLAR meeting discussed a proposal to embed FFV1 compression within PNG, evaluating its technical feasibility and alignment with the working group's charter. Significant attention was also given to the ongoing RFC Editor queue issues for the Matroska document (O48), specifically regarding the definition of chapter codec IDs. Updates on other document statuses and future work, including chapter codecs and FFV1v4, were also covered.
Key Discussion Points
- Previous Meeting Minutes Review: The minutes from the previous meeting (April 2024, which was cancelled) were briefly reviewed and accepted without further comment.
- PNG/FFV1 Compression Proposal:
- A recent mailing list discussion proposed using FFV1 as a compression method within PNG.
- Technically, embedding FFV1 in PNG is considered feasible as it's akin to using different compression methods (like Zlib) within the PNG container. There's no technical limitation preventing this.
- The primary challenge lies in registering a new compression ID within the PNG specification, which has a complex history (RFC 2083, W3C, ISO JTC1/SC24).
- The question was raised whether this work falls directly within the CELLAR WG's charter, which primarily focuses on Matroska and FFV1.
- W3C's GitHub repository was identified as a potential official channel for PNG development and ID registration, making interaction potentially easier than with ISO.
- It was noted that if the WG decides to pursue this, coordination with the IAB's liaison managers (e.g., Martin Thompson for W3C) or the responsible Area Director (Orie Steel) might be necessary.
- The consensus was that while technically possible and beneficial for FFV1's ubiquity, it might be on the edge of CELLAR's current charter and would require a motivated sponsor for the development and liaison work.
- FFV1 Video Codec Version 4 Milestone:
- The milestone for FFV1 Video Codec Version 4 was discussed.
- There is currently insufficient community interest or motivation to actively work on FFV1v4 development.
- A sense of those present indicated a desire to keep this as a placeholder for future work rather than removing it entirely, acknowledging potential future interest.
- Matroska (O48) Document Status:
- The Matroska document remains in the RFC Editor queue (O48).
- A specific issue was raised by the RFC Editor regarding the definition of chapter codec IDs (values 0 and 1) in section 20.3, which conflicted with section 27.2 where they are marked as "reserved."
- The problem arises because the values are intended to be defined in a separate chapter codec document, which is not yet published.
- It was proposed that the Matroska document should strictly keep these values as "reserved" and defer their specific definition to the forthcoming chapter codec specification.
- The chairs discussed the process of responding to RFC Editor queries, balancing the need to review changes for correctness with the RFC Editor's control over the document's XML.
- Ongoing corrections to RFC Editor changes (e.g., hexadecimal ranges for IDs) were noted to ensure accuracy.
- Additional Matroska Specifications (Chapter Codecs, Control Track):
- Work is proceeding on "Chapter Codecs," with a Google Summer of Code student working on interactive movie features in Matroska that rely on this.
- It was noted that "Chapter Codecs" and "Control Track" documents currently lack official milestones in the Datatracker, despite active development interest in the former.
- The chairs recognized the need to properly register these as working documents with associated milestones.
- Referencing RFC 9559:
- It was mentioned that some updates to other Matroska-related documents are pending the publication of RFC 9559 to allow for correct and stable referencing of sections and paragraphs.
Decisions and Action Items
- ACTION (Jerome): Draft an email reply to the mailing list regarding the FFV1-in-PNG proposal, summarizing the technical feasibility, charter considerations, and the need for codec ID registration (e.g., with W3C/GitHub).
- ACTION (Jerome): Investigate the official process for registering new compression IDs or methods with the PNG specification/W3C.
- ACTION (Chairs): Engage with Area Director Orie Steel to discuss the CELLAR WG's charter scope regarding potential involvement in the FFV1-in-PNG proposal and external liaison needs.
- DECISION: The milestone for "FFV1 Video Codec Version 4" will remain as a placeholder in the Datatracker, with its date removed to reflect its current inactive status.
- ACTION (Michael): Send an email to the RFC Editor proposing to resolve the Matroska (O48) issue by ensuring chapter codec IDs 0 and 1 are explicitly labeled as "reserved" in section 20.3 of the Matroska document, deferring their full definition to the separate chapter codec document.
- ACTION (Chairs): Create and associate milestones for the "Chapter Codecs" and "Control Track" documents within the Datatracker to reflect their status as ongoing or planned work.
Next Steps
- Jerome to complete the actions related to the FFV1-in-PNG proposal and report back to the WG.
- Michael to follow up with the RFC Editor regarding the Matroska (O48) document's codec ID issue.
- Chairs to meet with AD Orie Steel to discuss charter boundaries and liaison support.
- The next CELLAR WG meeting is scheduled for June 25th.