ACES 2.0: What you need to know & what you need to change

We all love when things get an upgrade, but what happens when that upgrade means having to learn something new? Old habits die hard, so let's dive into what's up.

What changed in ACES 2.0

ACES 2.0 is not a cosmetic bump. The Output Transform has been rebuilt and parameterized, the tone scale is less aggressive, highlight roll-off is gentler, and gamut mapping is more robust and invertible. The practical wins: SDR/HDR outputs track each other more predictably, extreme colors clip less brutally, and custom display targets (nits, primaries) are first-class rather than hacks. That’s by design; the Academy’s 2.0 notes spell out reduced mid-tone contrast, gentler highlights, and improved uniformity, plus better inverse behavior for round-tripping VFX plates. 

The upshot in a facility: you’ll see fewer “why does the Dolby master not look like the 709” moments, and you can spec non-stock outputs (e.g., 600-nit P3-D65 for a boutique LED cinema) without forking the pipeline. Industry coverage has been clear that the 2.0 DRT was designed exactly to address hue skews and SDR/HDR inconsistencies that dogged 1.x—especially under narrow-band LED/laser lighting. 

Two more operational notes you’ll feel day one:

• Transform invertibility is improved, which matters when you’re reconciling graded elements with VFX comps or doing plate-aware looks. 
• Parameterization is now formal, not a vendor one-off. Resolve 20.1 exposes it via DCTL Output Transforms (details below). 

Simply: ACES 2.0 is calmer in the mids, saner in the highs, and more honest about saturated color.

The new DRT explained

In ACES 1.x we conceptually thought “RRT + ODT.” In 2.0, think of a modern DRT that couples tone scale and gamut mapping with perceptual goals across outputs. The differences you’ll see on a scope and a reference monitor:

• Tone curve: less mid-tone contrast (faces don’t get over-shaped as quickly) and a slower knee into highlight roll-off (practicals, speculars, and skies keep texture longer). 
• Gamut mapping: chroma compression behaves more uniformly; intense colors are steered toward a path-to-white rather than hard clipping at the gamut boundary. The ACES working group documented experiments with chroma compression that varies with color purity to preserve vividness without breaking highlight neutrality. 
• Cross-output consistency: the new design aims for a closer appearance match between SDR and HDR references; you’ll still manage nit targets, but the creative delta between a 100-nit 709 and a 1,000-nit PQ master is smaller than it was in 1.x, given equal intent. 

What does that change in the room? You’ll do less per-output look surgery. You can keep a single show LMT and expect it to travel more honestly across SDR/HDR variants.

Neon blues and “polarity flips”

Now the part that causes nervous coffee sips. ACES 2.0 fixes many old artifacts; it also exposes edge cases that you must recognize and triage.

Neon blues

Highly saturated narrow-band emitters (LED signage, stage lasers) can provoke DRT behavior that feels “non-literal”—skies separating strangely, neon compressing into unexpected hues, edge halos where a saturated emitter meets a white object. In 1.x, some of us leaned on Resolve’s “Neon Suppression” LMT or the old Resolve Neon LMT to tame this. In 2.0, those stopgaps are not recommended; they were originally aimed at ACES 1.x and tend to over-compress usable ranges. The ACESCentral guidance is explicit: avoid those legacy fixes in ACES 2.0 pipelines. 

Polarity flips

You’ll see this term in research threads describing cases where the luminance order of colored vs white elements in high exposure zones feels “inside-out” depending on mapping—e.g., colored dots that perceptually appear brighter than adjacent whites or vice versa after rendering. The 2.0 work included explorations into luminance–chrominance polarity to keep highlight behavior perceptually stable as color purity rises. Understanding the phenomenon matters because you’ll be asked why a brand’s saturated logo either “pops too much” or “dies” under different outputs. The threads are technical, but reading them will help you diagnose whether the DRT, the source spectra, or a creative LMT is the lever to pull. 

How to triage in session

1. Is it acquisition or rendering? Check the plate in ACEScg linear before the DRT. If the oddity is already there, it’s camera/spectra, not rendering.
2. Is it gamut mapping or tone scale? Toggle gamut mapping modules (where exposed) and watch for hue shifts vs luminance compression.
3. Local, not global, fixes. Prefer shot-level tools (keys, constrained hue warpers) over pipeline-wide “neon suppression.” If you must constrain globally for a section (concert sequence), gate it with a show LMT and document the intent.

A practical tip: build a neon torture-test timeline (skies, sodium vapor street scenes, RGB LED signage, laser stage footage) and keep it as a facility regression test when you adopt minor DRT revisions or vendor builds. If that timeline survives, your show probably will too. Field reports and tests posted by practitioners show why this matters—neon-heavy images can behave very differently between 1.x and 2.0. 

Parametric DCTLs in Resolve (why you’ll actually use them)

Resolve 20.1 introduced parametric ACES 2.0 Output Transforms as DCTLs. This sounds academic until you need a deliverable your preset list doesn’t offer—say P3-D65 at 600 nits for a specific LED cinema or museum install. In 1.x, you would kludge this with LUT chains or custom ODTs; now you can instantiate a DCTL OT and set the parameters (primaries, white point, luminance target) inside the node, keeping your project ACES-managed and auditable. The ACESCentral announcement (and Nick Shaw’s example) calls this out explicitly. 

How to integrate without making a mess

• Project setup: Keep Color Management in ACES 2.0 (ACEScct timeline). Use standard IDTs. Do not stack unmanaged CSTs that defeat the pipeline unless you have a reason and notes.
• Output stage: Create a dedicated Output group post-clip node or timeline node where the DCTL Output Transform lives. Set the target primaries (e.g., P3-D65), white, and nit level per spec.
• QC: Verify with a calibrated display (pattern generator + probe), then round-trip: export a short EXR segment, re-ingest, and validate that inverse behavior behaves within tolerance for your VFX plates.
• Versioning: Name the node with the exact parameters (e.g., ACES2_DRT_P3D65_600nits_BT1886like_v0201) and snapshot the node tree. Treat this like a deliverable asset, not a tweak.

When to avoid parametric outputs

• If a distributor mandates a stock ODT (e.g., Rec.709 100-nit), stick to stock to avoid audit friction.
• If your client QC relies on their in-house LUT chain, giving them a parametric one-off breaks comparability.

Vendor support beyond Resolve

Facilities that live with onset look tools will care that ACES 2.0 is rolling into Silverstack/Livegrade, which means camera-to-cart look continuity matches the 2.0 DRT more closely than the 1.x era. That reduces “cart vs suite” arguments. 

Operator habits you must update

1. Stop trusting muscle memory on the mid-tones. Faces will initially feel “softer” compared to your 1.x habits. Don’t over-counter with a steep S-curve; ride lift/gamma with restraint and evaluate on a calibrated surround. The DRT is doing more of the graceful work; let it. 
2. Do not reach for legacy “neon fixes.” If signage bites, solve it locally or with a shot-scoped LMT. The older Resolve Neon LMT and RGC hacks are discouraged in 2.0. 
3. Match SDR/HDR by intent, not by eye-balling. Use the same creative LMT across outputs and let the DRT’s improved cross-output behavior reduce deltas. Validate with controlled viewing. 
4. Parameterize outputs deliberately. If you create a 600-nit P3 deliverable with a parametric DCTL, document it in the turnover and embed parameters in filenames and CM metadata where possible. 
5. Maintain a regression suite. Keep a standard set of torture plates (neon, saturated fabrics, lasers, LED walls, HDR skies). Run them whenever you update Resolve or DCTL packages. Practitioner threads repeatedly show surprises after updates; don’t be the next anecdote. 

Bottom line: ACES 2.0 is the most sensible the system has felt in years, but it asks you to retire old reflexes. Learn the DRT’s behavior, treat parametric outputs like deliverables—not toys—and keep neon torture tests handy. If you do that, you’ll spend less time nursing outputs and more time grading pictures.

Have any tips you think we missed? Discuss below!