Compute Occluded Samples
Includes hidden/occluded deep samples when the holdout result is calculated.
← Fusion node reference
Deep CompositingAdvanced
dHoldout
Depth-aware occlusion / holdout
Removes or occludes Deep samples according to the depth of another Deep stream, producing physically meaningful holdouts.
dHoldout
DaVinci Resolve · Fusion
dHoldout solves occlusion in a Deep pipeline where a flat 2D mask would not be sufficient. It can remove samples that sit behind or within another depth-aware object while preserving the remaining depth structure.
Overview
What dHoldout does
dHoldout solves occlusion in a Deep pipeline where a flat 2D mask would not be sufficient. It can remove samples that sit behind or within another depth-aware object while preserving the remaining depth structure.
Mental model
Think of the holdout stream as an invisible object that cuts away Deep samples based on where it actually sits in Z space.
Common uses
Occlude CG elements behind foreground geometry
Create depth-aware mattes
Remove hidden Deep samples
Integrate rendered objects without manual roto
Prepare correct Deep intersections
Flow anatomy
Inputs & outputs
The port structure shows how this node fits into a Fusion graph.
Input
Deep stream that will be cut or occluded.
dHoldout
Fusion node
Output
Deep Image with the holdout applied while retaining usable depth samples.
Inspector
Important parameters
Start with these controls before reaching for more complicated workarounds.
Holdout Z Offset
Sets the Holdout Z Offset.
Sample
Lets you sample a point in the Viewer to guide the deep holdout setup.
Holdout Center X / Y
Sets the X/Y coordinates for Holdout Center.
Practical setup
A reliable workflow
- 1
Verify both branches use the same depth convention.
- 2
Connect the image to be cut and the Deep holdout object.
- 3
Inspect the result through DeepToImage.
- 4
Check semi-transparent edges and volumetric regions.
- 5
Continue Deep compositing only after the holdout behaves correctly.
Typical node chain
Deep image→
Deep holdout→
dHoldout→
Deep result
Examples
Where this node fits in real work
Foreground occluder
Use a Deep foreground object to hide a CG element behind it.
Deep CGDeep ForegrounddHoldoutDeepToImage
Depth-aware cleanup
Remove samples behind a holdout before merging the scene.
ImageToDeepdHoldoutdMergeDeepToImage
Troubleshooting
Common mistakes
Using a flat alpha mask when depth-aware occlusion is required
Mixing incompatible Z scales
Flattening the holdout branch before use
Expecting wrong depth values to be fixed by swapping inputs
Production notes
Practical tips
Inspect depth before troubleshooting edges.
Use clean Deep sources with consistent camera conventions.
Keep the output Deep until all depth-aware operations are complete.
Related nodes