ACES 1.1 now available


(Alex Forsythe) #1

To the ACES Community,

Today we have released ACES 1.1, a minor update to the ACES system, which includes a series of new features. This is a “minor” system release, and therefore the version number is incremented to 1.1. This means the update adds a number of transforms but does not change the look or modify the existing core transforms beyond addressing reported bugs and/or inconsequential formatting/whitespace changes. We will also continue to refer to ACES as “ACES version 1” in conversation and documentation.

A series of new standard dynamic range (SDR) ODTs are included in this release. These ODTs were added at the request of the ACES Community based on their production needs. Examples include: P3 ODTs for devices with a D65 calibration white point, simulation of a D65 white point on a P3 device with a DCI calibration white point, and limiting of output image colorimetry to Rec.709 when using a P3D65 calibrated device. Additional limiting ODTs were also added for Rec.2020 to restrict the image colorimetry to Rec.709 and P3. A DCDM ODT with limiting to P3D65 were added to compliment the existing DCDM ODT with limiting to P3D60. These transforms provide support for additional use cases not included in previous ACES releases.

ACES 1.1 also includes the first release of a series of Output Transforms that combine the RRT and an ODT into a single transform. The new Output Transforms replace the previous HDR ODTs. The new Output Transforms are based on a unified, parametric output function. These individual Output Transforms pass a series of parameters to the underlying output function to improve the consistency of the image processing operations. Examples of the parameters that that are specified in the Output Transforms include: display primaries, display white point, display max luminance, display min luminance, luminance reproduction of mid-gray, limiting primaries (if any), surround, display EOTF, etc. In the future, this will make it trivial to generate Output Transforms for non-standard devices.

Output Transforms using the underlying parametric output function are only provided for HDR devices in dark surround environments at this time, although the transforms are suitable for use in both dark and dim surround environments (they’ve been used this way on several major motion pictures with a corresponding “trim pass” to adjust contrast and saturation). Modifying the surround parameter inside the individual HDR Output Transforms has no effect when the EOTF is set to either ST-2084(PQ) or HLG. A functional dim surround compensation for HDR devices will be addressed in a future ACES release.

ACES 1.1 also improves documentation of the ACES system, derivation of its components, and historical context around system design choices. An example of this is TB-2018-001. This document provides details of how the ACES white point chromaticity coordinates are derived and why they were chosen. The documentation includes reference python code of the derivation and a iPython notebook to show the calculation of values included in the document.

The CTL file names and values of of some of the existing ACES transforms have been modified for clarity. In each case, the output code values from the transform remains the same. Academy S-2014-002 states “Any transform updates that do not change the output of that transform shall not require the Transform Identifier to be incremented – e.g. whitespace changes, modifications to code comments, etc.” For this reason, the values have not been changed. Implementors should take care to update their user interfaces to display the new values of where applicable and to alert end-users of the name changes in their product documentation.

A detailed list of changes can be found in the README and CHANGELOG on github. We hope you find the improvements useful and look forward to your feedback!

ACES Leadership


ACES 1.1 Discussion - Tell us what you think!
(Steve Tobenkin) #2

We’d love your feedback too. Here is the discussion thread for the ACES 1.1 release: ACES 1.1 Discussion - Tell us what you think!