A processing system. Not one colour profile.

ACES is neither a single gamut triangle nor a LUT that produces a “cinematic look”. It is a system coordinating images, working encodings and transforms. Input handling, task-specific encodings, creative changes and output rendering have separate roles.

Start with each component’s role. ACEScg identifies a linear working encoding; it does not specify the Output Transform used for viewing. ACEScct identifies a grading-friendly encoding, not an automatic conversion of arbitrary camera data.

ACES2065-1 uses AP0 primaries, ACES reference white and scene-referred linear values. It serves standardized interchange and archival purposes. AP0 and AP1 cannot be exchanged because their names look similar.

Some AP0 primaries are mathematical rather than physically realizable emitters. They define coordinates, not a proposed display. An archive also needs enough information about transforms and creative state: a wide gamut alone cannot reconstruct an approved appearance.

AP1 is the primary set used by ACEScg, ACEScct and ACEScc. They share ACES2065-1’s white: x = 0.32168, y = 0.33767. The AP0 ↔ AP1 matrix change therefore does not need a reference-white change.

Do not confuse changing primaries with taking a logarithm. First derive linear AP1 from AP0, then choose linear or nonlinear encoding for those coordinates. The lab separates those operations and numerically checks the return path.

ACEScg is scene-linear AP1 for computer graphics and compositing. Linear values are useful when operations should act on relative light contributions, including addition, scaling and appropriate blending.

One is not a mandatory ceiling for all data. Values above one can represent bright elements; transforms may also produce negatives. Operations and storage must support the intended range. Do not clip intermediate data merely to make a histogram fit before output rendering.

ACEScct keeps AP1 but changes its encoding. Above a linear value of 0.0078125 it uses a logarithmic branch; at and below the threshold it uses a linear toe. This makes its low-end behaviour different from ACEScc.

Linear AP1 = 0.18 encodes to approximately 0.413588 in ACEScct. Linear zero encodes to about 0.072906, not zero. These are encoding properties, not monitor black levels. ACEScc has separate near-zero rules and does not preserve negative linear inputs one-to-one.

An Input Transform maps a known source representation into ACES’s coordinated scene-referred representation. It must match actual camera decoding: primaries, log curve, version and capture mode matter.

Already rendered graphics pose a different problem from camera originals. Do not blindly assign a camera Input Transform. Establish whether the source represents scene information or an already rendered appearance, and what must be preserved.

A creative look and a technical transform answer different questions. A look changes appearance; a technical transform reconciles representation with context. A LUT also expects a particular input and does not become universal inside ACES.

Record the look’s placement and version. Applying a display-referred Rec.709 LUT directly to ACEScct is not its intended operation. Reducing Strength alone does not resolve an input mismatch.

Reference Gamut Compression, introduced with ACES 1.3, addresses problematic scene-referred colours. It is not simply a demand to fit every intermediate pixel into the final display triangle.

Do not conflate it with gamut mapping within an Output Transform. Their scope, input assumptions and purposes differ. When software offers several related operations, establish their placement and what already runs automatically before adding another instance.

ACES 1.x descriptions use the RRT and ODT framework. Its role is to render scene-referred imagery for a particular output. That is much more than writing a new gamut tag.

Preserve an older project’s exact version and output selection. A system update is not automatically a transparent replacement guaranteed to retain an approved grade. Migration requires reference shots and comparison under matching conditions.

In ACES 2, the Output Transform separates rendering from target display encoding. The Rendering Transform handles tonal and colour relationships; Display Encoding expresses the result for the selected output.

The documented pipeline uses JMh, tone mapping, chroma compression and gamut compression. This page’s lab does not replace it with a homemade curve: it computes only AP0/AP1 changes and ACES encodings. Evaluate actual output through a compatible ACES pipeline and verified viewing.

A reproducibility record includes ACES version, application implementation, input transforms, working encoding, creative operations and output. A project file version alone does not capture every external dependency.

Before updating, render references and save settings. Compare highlight details, saturated areas, neutrals and different cameras. Do not approve migration from one mid-shot or matching setting names.

In Resolve, first establish whether the project manages transforms centrally or you build them explicitly. Define correct inputs, working encoding and output. Available ACES variants and control names depend on application version.

Do not duplicate input or output transforms in nodes when the system already performs them. A flat image with output processing disabled is not necessarily damaged: you may be viewing working codes without display preparation. Identify where in the chain you inspect the signal.

For VFX exchange, agree on file encoding, transform direction and which appearance processing is baked in. Linear ACEScg, ACES2065-1 and a rendered review file serve different purposes. An “ACES” folder name is insufficient.

Test a no-grade round trip with coloured values, values above one and, where supported, negatives. Separate coordinate mismatches from storage losses, clipping and baked output rendering.

ACES and DaVinci Wide Gamut/Intermediate are not two saturation switches. ACES describes a system and encoding family; DWG/Intermediate defines a particular working gamut and transfer pair usable in Resolve’s managed workflow.

Choose based on interchange, production requirements, available transforms and reproducibility. Neither removes the need for correct input and output identification. Wide working coordinates cannot rescue incorrectly interpreted camera data.