fix: PT final upsample is phase-aware (bilinear + depth joint bilateral)#98
Merged
Conversation
The trace-to-full upsample mapped every output pixel to its CONTAINING trace texel by integer division and centred the kernel weights there - no fractional phase. The lighting therefore reconstructed as trace-texel-sized constant blocks, and because the trace grid rotates its sample phase every frame (so the EMA integrates all four sub- positions), the block boundaries CRAWLED whenever the camera moved. On top of renderScale 0.5 (trace = quarter of output res) that crawl was the residual pixelated-while-moving look indoors after the PT-9/9b disocclusion fixes. The taps now weight by tent distance from the CONTINUOUS trace-space position (texel centres aligned), turning the pass into a true bilinear-plus-depth joint bilateral: smooth gradients, stable under the per-frame phase rotation, same 3x3 tap cost. The thin-geometry epsilon fallback and the sky guard are unchanged, and the full-res-trace path still degenerates to modulate-and-write. Verified: both golden PT tests pass; the translating-dolly capture at renderScale 0.5 shows smooth wall gradients with no block crawl at unchanged fps. Claude-Session: https://claude.ai/code/session_018574BdCfSjdpNK3WLgx1hP
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughThe SVGF final pass replaces integer trace-texel mapping and B3-spline sampling with center-aligned sub-texel coordinates and tent weights. Depth-based weighting and surrounding bounds checks remain unchanged. ChangesSVGF final-pass resampling
Estimated code review effort: 2 (Simple) | ~10 minutes ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
82 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The trace-to-full upsample mapped every output pixel to its CONTAINING
trace texel by integer division and centred the kernel weights there -
no fractional phase. The lighting therefore reconstructed as
trace-texel-sized constant blocks, and because the trace grid rotates
its sample phase every frame (so the EMA integrates all four sub-
positions), the block boundaries CRAWLED whenever the camera moved.
On top of renderScale 0.5 (trace = quarter of output res) that crawl
was the residual pixelated-while-moving look indoors after the
PT-9/9b disocclusion fixes.
The taps now weight by tent distance from the CONTINUOUS trace-space
position (texel centres aligned), turning the pass into a true
bilinear-plus-depth joint bilateral: smooth gradients, stable under
the per-frame phase rotation, same 3x3 tap cost. The thin-geometry
epsilon fallback and the sky guard are unchanged, and the
full-res-trace path still degenerates to modulate-and-write.
Verified: both golden PT tests pass; the translating-dolly capture at
renderScale 0.5 shows smooth wall gradients with no block crawl at
unchanged fps.
Claude-Session: https://claude.ai/code/session_018574BdCfSjdpNK3WLgx1hP
Summary by CodeRabbit