Shared Textures
A Shared Texture is a named GPU image produced in one part of the project and sampled elsewhere. It is the explicit bridge for results that need to cross layer, shader, particle, projector, or 3D-scene boundaries.
Publish and consume
Section titled “Publish and consume”- Add a Shared Texture Output to the producer graph and connect the result you want to expose.
- Give the output a clear name based on its role, such as
Retained directionsorCamera mask. - In the consumer, select that named Shared Texture as the source.
- Check the GPU Resources panel if the expected output is missing or has the wrong size.
Shared Texture output preserves the underlying field type. A vector field or exact ID map is not automatically flattened into display color just because its preview looks like an image.
For a height field feeding Thin Film or Height Refraction, send the scalar field directly to a Shared Texture Output. Routing it through a screen-color image can round away small height differences and create visible contour steps when the material computes slopes. In custom WGSL, write the field to the shared-texture output as well as any screen preview.
Why not use Previous Pass everywhere?
Section titled “Why not use Previous Pass everywhere?”Previous Pass only means the current image entering a card in its own stack. It cannot represent every cross-layer or direct 3D/projector dependency. Shared Textures make those dependencies explicit and schedulable.
Names and Rack access
Section titled “Names and Rack access”The intended shared-texture workflow exposes useful names in Rack and GPU Resources, supports preview, and allows drag-to-sample authoring. Treat those controls as available only after the shared-texture Rack change is integrated into the Studio build you are using.
See Layers and the stack and Feedback for the two common producer patterns.