Quick answer: Use Resolve Color Management when you want the input, working-space and output transforms handled once for the whole project. Use CST nodes when you'd rather see those transforms in the node tree and place them yourself.
Whichever route you pick, the job is the same: identify each camera input correctly, grade in a known working space such as DaVinci Wide Gamut/Intermediate, and feed every LUT the color space and gamma it was designed to receive. Don't combine RCM with matching input/output CSTs unless you're deliberately overriding part of the managed pipeline.
Think of a color-managed image as a controlled route: source encoding in, a timeline working space in the middle, a display or file out. Resolve Color Management runs that route at project level. Color Space Transform nodes put selected mappings right in the graph where you can see them.
The workflows and version-sensitive labels below follow the color management chapter of the DaVinci Resolve 21 Reference Manual for Resolve 21.1.
Most failures come from losing track of the signal. A clip gets tagged with the wrong log curve, it gets normalized twice, or it lands in a LUT that expects a different input.
So this guide builds the pipeline from the signal outward. Once the signal path is clear, you can change cameras, LUTs or deliverables without guessing. Keep the Resolve scopes guide handy and judge whether a transform preserves useful highlight, shadow and saturation information by the scopes instead of trusting the viewer alone.
RCM and CST Solve the Same Mapping Problem
| Question | RCM | CST workflow |
|---|---|---|
| Where are transforms controlled? | Project settings and clip input tags | Visible nodes, groups or timeline nodes |
| Mixed camera formats | Efficient when inputs are tagged correctly | Explicit but requires per-camera transform management |
| Changing SDR/HDR output | Central output setting | Change or branch the output transform |
| Debugging | Fewer visible nodes, more hidden state | Signal path is visible in the graph |
| Image quality | Accurate when configured correctly | Accurate when equivalent transforms are configured correctly |
Choose by project design, not because someone told you RCM or CSTs create better color. RCM is usually the easier pick for long edits with mixed cameras and several deliveries. CSTs earn their place when a colorist wants to inspect, move, group or branch the transforms explicitly.
You can build equivalent input, working and output mappings either way. The catch is that RCM Output DRT and CST tone/gamut-mapping settings must also match before comparing the rendered results tells you anything.
A habit worth breaking: enabling DaVinci YRGB Color Managed and then adding a camera-to-DWG CST at the start plus a DWG-to-Rec.709 CST at the end anyway. RCM may already be doing both jobs. That stack can normalize the source twice and apply the display transform twice.
Understand Input, Timeline and Output Color Space
- Input color space: the camera gamut and transfer function that encoded the clip, such as Sony S-Gamut3.Cine/S-Log3 or Blackmagic Wide Gamut/Film Gen 5.
- Timeline color space: the working environment in which color-aware tools and nodes operate. It's the space your grade actually lives in.
- Output color space: the target display or deliverable mapping, such as Rec.709 Gamma 2.4 for a controlled SDR viewing environment.
An input tag isn't a creative conversion. It tells RCM how to interpret the numbers already in the file. Get the tag wrong and contrast, saturation and hue go wrong before you've touched a wheel. Don't identify a clip by how flat it looks; check camera metadata, the production notes or the source-camera settings.
Automatic color management saves time when files carry reliable metadata, but I'd still inspect mixed-camera and transcoded media. The manual describes a cascade: RCM uses the camera's raw colorimetry, then embedded QuickTime or MXF color metadata, and only then the project's default Input Color Space. A transcoded file that lost its tags can quietly land on that default. In a managed project, right-click clips in the Media Pool or Color-page clip strip and assign the actual Input Color Space when needed.
What DaVinci Wide Gamut and Intermediate Actually Mean
DaVinci Wide Gamut is the color gamut. DaVinci Intermediate is its companion log-like transfer function. Together they form a large scene-referred working space designed to hold camera data and behave consistently across SDR and HDR grades. Blackmagic describes the gamut as larger than BT.2020, ARRI Wide Gamut and even ACES, so camera data isn't squeezed on the way in.
Working in DWG/Intermediate doesn't make every source HDR, it doesn't recover clipped sensor data, and it won't improve an 8-bit file. What it does is give correctly transformed sources a common, roomy environment. Balancing exposure and color, matching shots and controlling saturated colors is still your job.
One detail trips people up in a manual DaVinci YRGB plus CST workflow. Setting the project's Timeline Color Space to DWG/Intermediate doesn't perform the input or output conversion. It declares the working environment to color-space-aware tools and output metadata, while the explicit CST nodes still perform the image mappings.
Set Up Resolve Color Management
- Open Project Settings > Color Management.
- Set Color science to DaVinci YRGB Color Managed.
- Turn off Automatic color management when you want to see and control the input, timeline and output settings directly.
- Choose DaVinci Wide Gamut/Intermediate as the timeline working space.
- Set the output to the monitoring and delivery target you actually have.
- Save, then check and correct the Input Color Space for every source family.
- Grade in normal clip and group nodes without adding redundant normalization CSTs.
The output transform occurs after the managed grading pipeline. That makes it convenient to take a well-built scene-referred grade from an SDR output to a separately checked HDR output, but it doesn't remove the need for an output-specific trim and proper monitoring. The manual is direct about this: the automatic SDR-to-HDR mapping is a starting point, and every deliverable in a new color space gets its own trim pass.
RCM handles the mapping; it doesn't correct shots. A correct source tag has no way of knowing the camera was two stops under, the white balance was off or one angle used a different ND filter.
Continue with balancing and shot matching in the color grading guide. When you borrow someone else's RCM settings, check what they were delivering first. Video Village's filmic photography walkthrough for Resolve 21 from May 2026 sets the output to sRGB because the subject is still photos meant for screens and exported with sRGB tags. That's correct for their example, not a universal video preset, so choose the output that matches your calibrated monitoring and delivery target.
Build a Manual CST Sandwich
For explicit node-based management, leave the project's Color science on DaVinci YRGB and build a clean signal path yourself:
- Input CST: camera gamut/gamma to DaVinci Wide Gamut/DaVinci Intermediate.
- Balance: exposure and white-balance corrections.
- Shot match: bring the shots in the scene into line.
- Look: creative contrast, palette and texture, built in the known working space.
- Output CST: DWG/Intermediate to the monitoring/delivery space.
- Output-space finishing: only effects deliberately designed for that display space.
When many clips come from one camera, put the input transform at Group Pre-Clip level. Put a shared output transform at Group Post-Clip or Timeline level when the whole group or timeline shares one destination. Keep clip nodes for shot-specific work. Otherwise you end up with 100 slightly different copies of the same technical transform.
Blackmagic's Colorist Guide also documents a valid single-CST pattern: grade upstream in the source space and transform camera input directly to output near the end, in the Group Post-Clip or Timeline layer. The two-CST sandwich is easier when several camera families need a common working space. It's a convenience, not a law for every project.
Place LUTs by Their Input and Output Contract
Before you drop any LUT on a node, answer two questions:
- What gamut and gamma does the LUT expect?
- What gamut and gamma does it produce?
A camera technical LUT may expect one exact log profile and output Rec.709. A creative show LUT may expect and return DWG/Intermediate. A film-print emulation may expect Cineon Film Log and hand back a display-referred Rec.709 result. Resolve can't infer any of that from a generic .cube file, so the contract is on you.
| LUT type | Safe placement logic | Common error |
|---|---|---|
| Camera log to Rec.709 technical LUT | Use it as the intended normalization/output mapping, or do not use it | Adding it after RCM already mapped the same camera to Rec.709 |
| DWG-to-DWG creative LUT | Inside the DWG/Intermediate grading section | Feeding it camera log or display-referred Rec.709 |
| Rec.709 creative LUT | Feed it the specified Rec.709 variant and account for its output | Dropping it directly into a DWG node tree |
| Resolve film-print LUT | Prepare its documented Cineon Film Log input and design the output path around it | Treating it as a generic DWG look |
Blackmagic's Resolve 20 Colorist Guide spells this out for the bundled Rec709 Kodak 2383 film-look LUTs: they anticipate Cineon Film Log as the input transfer function and are intended for Rec.709 monitoring or delivery.
Its worked example converts the incoming Rec.709 gamma to Cineon with a CST immediately before the LUT, and the guide openly notes that the gamma gets remapped several times along the way. That's fine for a lesson, less fine for a real project. If you don't want to manage that contract and its display-referred result explicitly, use a DWG-native look or a suitable film-look tool instead of treating the LUT as a generic DWG look.
Installation, previews and intensity controls are covered in the separate Resolve LUT guide. The rule here is narrower: the node before a LUT must produce what that LUT expects.
Choose Node Order Without Memorizing a Template
Ask three things of every node: what space enters, what operation happens, and what space leaves? A qualifier may key differently before and after a display transform. Grain or halation may be designed for scene-referred or display-referred values. A LUT has a fixed contract. Put each operation in the section it was designed for.
A practical wide-gamut tree looks like this:
Camera input to DWG/I | balance | matching | look | DWG/I to output | output-specific finish
In RCM, the first and last stages are managed outside the visible clip tree. In a CST workflow, you can see them. If a particular look really requires Rec.709 or Cineon, build a deliberate transform around that one operation rather than pretending the whole node tree changed spaces.
Diagnose Double Transforms and Wrong Tags
The log image becomes harsh before grading
Bypass the CST and LUT nodes one at a time. If RCM is enabled and a technical LUT or input CST adds the same normalization, remove the duplicate. While you're there, verify the clip's Input Color Space too.
Changing the RCM output makes the image break
There's probably a display-referred LUT or node sitting inside the scene-referred tree. Work out what that operation expects. A DWG-native look should survive an output change far more predictably than a Rec.709-baked LUT.
Different cameras do not match after RCM
Verify every input tag first. Then balance exposure and white balance. Color management normalizes encoding; it doesn't erase sensor, lens, lighting or exposure differences.
The manual CST image looks different from RCM
Compare the actual input gamut/gamma, output gamut/gamma, tone mapping, gamut mapping and timeline settings. "CST to Rec.709" isn't specific enough, because several Rec.709 gamma choices and mapping options can produce different results.
The LUT looks correct only after random CSTs
Stop and read the LUT documentation or filename. Rebuild its stated input and output on purpose. A chain that looks pleasing by accident can still clip, remap or double-transform values in ways that fall apart on the next shot.
Validate the Pipeline Before the Full Grade
- Pick representative clips from every camera and graphics source.
- Check the source tags and transform only those clips.
- Read the scopes for clipping and out-of-gamut behavior.
- Bypass the look for a moment and confirm the technical pipeline still holds together.
- Test the darkest, brightest and most saturated shots.
- Render a short file with the intended output tags and watch it on the target playback path.
- Only then ripple the pipeline across the project.
Monitoring is part of color management too. A correct Rec.709 mapping viewed on an uncalibrated display, or through a player with different color handling, can still look inconsistent. You need to know both the pipeline and the viewing path.
Frequently Asked Questions
Is RCM better quality than CST?
No. When both routes use equivalent input, timeline and output transforms with matching tone- and gamut-mapping choices, neither has an inherent quality advantage. RCM centralizes the pipeline; CSTs expose selected mappings in nodes. What actually decides it is operational consistency and deliberate settings.
Do I need CST nodes in a color-managed project?
Usually not for the same input-to-timeline and timeline-to-output transforms RCM already performs. Use a CST only for an intentional local conversion, a special effect or a LUT requirement, and bring the image back to the expected managed space afterward when necessary.
What is the difference between DWG and DaVinci Intermediate?
DaVinci Wide Gamut defines the color gamut, and DaVinci Intermediate defines the transfer function used with it. Together they make up the common DWG/Intermediate working space.
Where should a LUT go in the node tree?
Place it where the incoming gamut and gamma match the LUT's documented input, then plan for whatever the LUT outputs. No single position is correct for every LUT.
Why does my image look double-normalized?
RCM may be applying an input or output transform while a CST or technical LUT applies the same conversion again. Bypass each mapping stage in turn and keep only one intentional route.
Should I grade before or after the output CST?
In a scene-referred wide-gamut workflow, do most of the balancing, matching and look work before the output CST. Only operations deliberately designed for the output/display space belong after it.
From here, the DaVinci Resolve color grading hub takes you from pipeline setup into balancing, matching, LUT work and delivery checks without mixing those jobs together.
