• Joined on 2026-05-22
lars commented on issue lars/splat-indoor-navigation#11 2026-08-11 16:08:09 +00:00
Decide: success gates and the measurement protocol

Add a sustained/thermal gate — peak frame rate is the wrong measurement.

Mobile-GS (ICLR 2026) measured 127 → 74 FPS thermal collapse at 0.83 W on a Snapdragon 8 Gen 3. For calibration…

lars commented on issue lars/splat-indoor-navigation#2 2026-08-11 16:08:09 +00:00
Research: on-device visual relocalization against a gaussian splat

Addendum 2 — the aliasing failure mode, now quantified, and it is worse than assumed

A third pass corroborates this ticket's verdict and puts numbers on the risk that was previously only…

lars pushed to main at lars/splat-indoor-navigation 2026-08-11 16:05:26 +00:00
b463af35d0 spark 2.1.0 splat viewer tab with OS-fused gyro look-around
lars commented on issue lars/splat-indoor-navigation#10 2026-08-11 16:03:08 +00:00
Decide: bake-off design — comparands, ground truth, decision rule

One more measurement this bake-off must make: recall, not just accuracy.

From #2's vendor addendum — every visual-positioning vendor publishes 6–8 cm accuracy, comfortably inside the 1 m…

lars commented on issue lars/splat-indoor-navigation#2 2026-08-11 16:03:08 +00:00
Research: on-device visual relocalization against a gaussian splat

Addendum — VPS is less closed off than the answer above concluded, and the real risk is recall, not accuracy

A late vendor-focused pass corrects several specifics and one framing.

###…

lars commented on issue lars/splat-indoor-navigation#10 2026-08-11 16:02:01 +00:00
Decide: bake-off design — comparands, ground truth, decision rule

Arm A is cheaper than assumed, and its error model is now known.

  • @reactvision/react-viro 2.57.5 gives ARKit pose in JS with zero SwiftonCameraTransformUpdate (≤60Hz),…
lars commented on issue lars/splat-indoor-navigation#3 2026-08-11 16:02:01 +00:00
Research: WebGL2 splat viewer performance inside iOS WKWebView

Addendum 4 — memory: you have two budgets, and one footgun that blanks the WebView permanently

Refines the 300–400MB figure in the original answer, from a WebKit-source pass.

**There…

lars commented on issue lars/splat-indoor-navigation#15 2026-08-11 16:02:01 +00:00
Decide: how the user's pose origin gets fixed, and re-fixed as it drifts

Largely pre-answered — every absolute-heading option is disqualified, so markers it is

A late research pass on ARKit VIO and RN bindings lands almost squarely on this ticket. Read before…

lars commented on issue lars/splat-indoor-navigation#7 2026-08-11 16:02:01 +00:00
Decide: the pose contract — shape, units, site frame

Hard constraint on the confidence field — ARKit gives you almost nothing to fill it with.

From the ARKit research pass:

  • Apple publishes no covariance and no accuracy field. The…
lars commented on issue lars/splat-indoor-navigation#14 2026-08-11 16:00:28 +00:00
Write the Session 1 spec (pose to splat rendering)

Architecture inputs, corrected and expanded (see #3 addendum 3, #6 addendum 2).

  • The bridge is not the bottleneck on iOS (~0.2–1.5ms one-way). Budget is dominated by **ARKit pose age…
lars commented on issue lars/splat-indoor-navigation#17 2026-08-11 16:00:28 +00:00
Decide: does orientation bypass the bridge via DeviceOrientationEvent in the WebView?

The calculus for this decision has weakened — read #3 addendum 3 before working it.

This ticket was created because the bridge looked expensive (~10ms per call). That figure is **Android-only…

lars commented on issue lars/splat-indoor-navigation#16 2026-08-11 16:00:28 +00:00
Measure the RN → WebView pose-bridge latency

Correction to my earlier note on this ticket: the 10ms floor is Android-only.

Measured iOS figures do exist (see #3 addendum 3): evaluateJavaScript round trip 0.39ms and `WKScriptMessage…

lars commented on issue lars/splat-indoor-navigation#6 2026-08-11 16:00:27 +00:00
Research: native / RN gaussian splat rendering without a WebView

Addendum 2 — correction: Babylon React Native IS a turnkey native splat path, cross-platform

The answer above says nothing cross-platform reads splats natively and that Babylon RN's…

lars commented on issue lars/splat-indoor-navigation#3 2026-08-11 16:00:27 +00:00
Research: WebGL2 splat viewer performance inside iOS WKWebView

Addendum 3 — correction: the iOS bridge is cheap. Addendum 2's 10ms floor is Android-only.

A third pass found the iOS measurements Addendum 2 said didn't exist. This reverses that…

lars closed issue lars/splat-indoor-navigation#5 2026-08-11 15:59:53 +00:00
Research: iPhone Pro LiDAR capture to a metric, gravity-aligned splat
lars commented on issue lars/splat-indoor-navigation#5 2026-08-11 15:59:52 +00:00
Research: iPhone Pro LiDAR capture to a metric, gravity-aligned splat

Answer

Research date: 2026-08-11. Every claim below is tagged [VERIFIED] (read directly in a vendor doc / repo / source file), [VENDOR CLAIM] (stated by the vendor but unverifiable…

lars opened issue lars/splat-indoor-navigation#17 2026-08-11 15:58:29 +00:00
Decide: does orientation bypass the bridge via DeviceOrientationEvent in the WebView?
lars commented on issue lars/splat-indoor-navigation#16 2026-08-11 15:58:29 +00:00
Measure the RN → WebView pose-bridge latency

Baseline and method, from #3's second addendum — this ticket is now partly answered and better specified.

Published Android numbers to measure against ([discussion #3669](https://github.com/r…

lars commented on issue lars/splat-indoor-navigation#3 2026-08-11 15:58:28 +00:00
Research: WebGL2 splat viewer performance inside iOS WKWebView

Addendum 2 — measured bridge costs, format refinement, and a bypass worth considering

A second late sub-agent pass landed after this issue closed. It read react-native-webview and RN source…

lars commented on issue lars/splat-indoor-navigation#6 2026-08-11 15:57:00 +00:00
Research: native / RN gaussian splat rendering without a WebView

Addendum — Apple ships first-party splat rendering on iOS 27

A late-reporting sub-agent pass landed after the answer above was posted. It does not change the verdict (WebView + Spark for…