Immersive Gaussian Splat tour of grace cathedral, San Francisco
Details
- External ID
- 48984254
- Source
- HN
- Company
- —
- Product
- Immersive Gaussian Splat tour of grace cathedral, San Francisco
- Website domain
- vincentwoo.com
- Launched
- July 20, 2026
- Cohort
- —
- Upvotes
- 263
- Upvotes percentile
- 0.978494623655914
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Enrichment
- Theme
- creative coding and visual experiments
- Vertical
- Media & entertainment
- Function
- Content generation
- Audience
- B2C
- AI stance
- Not AI
- Project type
- Hobby / open-source project
- Normalized one-liner
- 3d gaussian splat tour of grace cathedral
- Manually corrected
- False
Could you build this?
Partial The web viewer itself is straightforward using open-source WebGL/WebGPU Gaussian Splatting viewers, but capturing and training the high-fidelity 3D reconstruction requires specialized photogrammetry workflows.
What it would actually take: The frontend is easily built using standard libraries like `@mkkellogg/gaussian-splats-3d` or PlayCanvas. The hard part is the capture and data pipeline: capturing thousands of high-resolution, controlled exposure photographs and drone scans of Grace Cathedral, running them through COLMAP for Structure-from-Motion, and optimizing the 3DGS model via heavy GPU training pipelines. Producing the clean, artifact-free dataset requires physical on-site access, photogrammetry hardware, and 3D computer vision expertise.
Discussion
20 comments analyzed.
Competitors mentioned: World Labs, Google Maps/Earth 3D views
Concerns raised: Firefox WebGPU support flaky on Windows, Complex patterns rendered as stretched, hazy ellipses, Lighting appears off, not capturing view-dependent specularity and reflections, Focal length perception feels incorrect
Feature requests: Anchor points on windows, Automation for tuning process, Procedure/tools documentation for creating city-scale captures
Competitors
Other products that read as similar to this one — 614 launches clear the similarity bar, closest 8 shown.
Attention rank: #18 of 615 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 261 days after the earliest competitor.
- Turning a Gaussian Splat into a videogame · hn · 2026-04-23 · 236 upvotes · similarity 0.63
- SVO Voxelization for Gaussian Splat Collisions · hn · 2026-02-18 · 5 upvotes · similarity 0.62
- gaussian-light-transport · github · 2026-09-11 · 14 upvotes · similarity 0.54
- Isopolis · hn · 2026-08-03 · 350 upvotes · similarity 0.54
- Lofi Cities · hn · 2026-09-27 · 314 upvotes · similarity 0.53
- stratum · github · 2026-09-18 · 11 upvotes · similarity 0.52
- hktex · github · 2026-09-09 · 97 upvotes · similarity 0.50
- seijaku · github · 2026-09-16 · 18 upvotes · similarity 0.50
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a content generation tool for Government yet.