Temporal Denoiser Settings
Last updated
Last updated
Specifies the number of temporally accumulated frames. More samples lead to better noise reduction with no additional performance cost, while fewer samples make occlusion more reactive.
GTAO's temporal sample count ranges from 0 to 12 samples, while RTAO's range is from 8 to 16 samples due to the lower ray count.
Performance impact: none
Setting this value to 0 disables temporal denoising for GTAO. In this case, a different dither pattern, more suitable for non-temporal denoising, is automatically applied.
You can use Debug Mode "Temporal Disocclusion" to inspect reprojection. In this mode, freshly disoccluded areas, where no temporal history is available, are marked in red. Accumulation progress is shown in grey, and fully accumulated pixels are displayed in white.
Controls the strictness of temporal history rejection. Lower values accept all history, producing smoother output with less noise, but can cause ghosting and trailing near moving objects or during camera translation/rotation. Higher values reject potentially invalid history samples, but may result in noisier output.
Performance impact: insignificant
While GTAO applies motion rejection to the entire frame, RTAO only applies it to pixels whose rays hit moving objects. These areas can be inspected through the Motion Rejection Mask debug dropdown.
Specifies whether the difference in surface normals should be considered during temporal history reprojection. This option can mitigate reprojection artifacts, such as the one in the screenshot where color from the frontal plane of the cube "leaks" onto its newly revealed side during camera panning.
Setting this parameter to high values close to 1.0 can cause temporal instability on small and thin details, such as foliage. Use the Temporal Disocclusion debug view to inspect the impact of Normal Rejection on temporal reprojection.
Use Temporal Disocclusion debug view to inspect the impact of Normal Rejection on temporal reprojection.
Performance impact: moderate, activate only if necessary
Defines the overall strictness of temporal history rejection. Similar to Normal Rejection, setting this parameter to high values close to 1.0 can cause temporal instability on small and thin details, such as foliage. Using very low values close to 0.0 can lead to ghosting in certain scenarios.
Use the Temporal Disocclusion debug view to inspect the impact of Rejection Strength on temporal reprojection.
Defines a reprojection filter used for temporal history fetch. Bilinear is fast but introduces blur, while Lanczos is approximately three times slower but much sharper. This option affects only the sharpness of the reprojection, not its effectiveness.
The filter is applied only to the history, meaning the sharpening effect will be visible during motion. It isn't a general-purpose sharpening effect (like CAS) and will have no impact on the stationary image.
Performance impact: moderate, activate only if necessary
