To All Manufacturers Working on ACES 2 Support,
In September, we finalized and distributed the algorithm for ACES 2.0 with the request for you to integrate its implementation into your product plans. While some of you have made significant progress, others are still working on it and encountering challenges.
As with previous versions of ACES, the CTL serves as the “reference code,” but we expect implementers to translate it into other programming languages. However, ACES 2 is notably more computationally complex than version 1, and some implementers have raised concerns about the increased computational demands.
To address these concerns, the Output Transforms Architecture group has been exploring various potential optimizations and simplifications. While no decisions have been made yet, our goal is to chart a path forward that balances the needs of all implementers.
If optimizations are pursued, they may alter output pixels enough to necessitate updates to the reference code. Should this occur, we aim to implement these updates efficiently, ensuring they’re timely and fair to vendors who have already made significant progress with the current reference code. Ideally, any changes to pixel output for performance improvements would be done only once to minimize disruption.
Getting everyone together for a meeting is challenging, but we will start by repurposing our previously scheduled meeting time next Wednesday, December 11, at 1 PM Pacific, to kick off this discussion. If you’re able to attend, we strongly encourage you to join.
Additionally, we are exploring options for asynchronous communication to allow participants to share updates, ask questions, and provide input at their convenience. This platform would allow implementers to share ongoing progress or insights with others in the group (to the extent that you’re comfortable sharing with potential competitors).
We want all stakeholders to contribute to shaping the best path forward. Specifically, we’d like to understand:
- How far along are you with your implementation work?
- Have you encountered any performance issues?
- How and when would you prefer to stage updates to minimize impact and maximize agreement?
- What other tests or tools might be helpful for you as you work on implementation support?
If modifications to the base code are deemed necessary to achieve the desired performance gains, we seek your input on how to implement these changes in a way that minimizes challenges for those already supporting the current version.
We look forward to having your participation in collaborating on a solution that works for everyone.
If you have any private concerns or feedback to share, you may reach out to me.
Thank you.