auto-convergence on for the stereo self view (call) and the camera sample

CHANGELOG.md

@@ -5,6 +5,17 @@ entry points (`.`, `./three`) are frozen for 1.x, while the **scene subpaths** (
 `./splat`, `./model`) are a preview tier whose options may change in any release. Entries below say
 which tier they touch, because that is what tells you whether an upgrade can move your pixels.
 
+## Unreleased
+
+Touches the **preview tier** (`./call`) only: one default. No export or type changes.
+
+### `./call` — the stereo self view auto-converges too
+
+- The call's self view now follows the call's **`autoConverge`** option (default `true`) instead of
+  being hard-wired off, so your own stereo camera feed sits at the display plane like the remote
+  tiles. `autoConverge: false` turns it off for both. (`samples/camera` also starts with
+  auto-convergence on.)
+
 ## 1.37.1 — 2026-10-07
 
 ### Depth cursor: hybrid anchor keeps the line of sight, not the foot

docs/call.md

@@ -138,7 +138,7 @@ Disconnecting the element (`el.remove()`, a framework unmounting it) leaves the
 | `camera` | `'auto'` | `'auto'` opens the best camera through [`/camera`](camera.md): a stereo device when one is present (the DisplayXR Browser's "3D Camera", or a wide side-by-side device) else the default webcam; `'stereo'` prefers the pair and falls back; `'mono'` never probes; a `deviceId` string; a `MediaStream` you own; or a `StereoCamera` from `openCamera()` (left open when the call ends) |
 | `format` | `'mono'` | The format of a page-supplied `MediaStream` (`'sbs'` or `'mono'`). 3D-ness is never guessed from a stream |
 | `audio` | `true` | Microphone, with echo cancellation and noise suppression |
-| `autoConverge` | `true` | Measure the disparity of the point between a stereo peer's eyes and put that person at the display plane. No calibration needed |
+| `autoConverge` | `true` | Measure the disparity of the point between a stereo peer's eyes and put that person at the display plane. No calibration needed. Also governs your own stereo self view |
 | `mono3D` | `'auto'` | Lift 2D peers to 3D: `'auto'` · `'off'` · a `lift` function (see [`/call/full`](#callfull-vs-call--lift--which-import)) |
 | `liftOptions` | — | Extra `lift()` options, e.g. `{ models }` for where the depth model is served from, plus `max` (default 4): concurrent lifted tiles; further 2D peers stay flat |
 | `maxPeers` | `4` | Participants including you, clamped to 2–4 (full mesh) |

js/inline3d-call.js

@@ -1873,7 +1873,7 @@ class SelfTile {
       // will not weave (firstWoven → woven:false, web#131), and the badge must follow it.
       this.view = addCameraView(wall, this.canvas, cam, {
         mirror: true,
-        autoConverge: false,
+        autoConverge: !!this.call.o.autoConverge, // the call option (default on) governs the self view too
         aspect: this.call.o.tileAspect,
         onRouteChange: (route, st) => {
           if (this.cam !== cam) return; // a view this tile already replaced

samples/camera/app.js

@@ -75,7 +75,7 @@ say(
   inline3DAvailable() ? 'woven' : 'flat'
 );
 
-const view = addCameraView(wall, /** @type {HTMLCanvasElement} */ (document.getElementById('self')), cam, { mirror: true, autoConverge: false });
+const view = addCameraView(wall, /** @type {HTMLCanvasElement} */ (document.getElementById('self')), cam, { mirror: true, autoConverge: true });
 $('mirror').addEventListener('change', () => view.setMirror($('mirror').checked));
 $('converge').addEventListener('change', () => view.setAutoConverge($('converge').checked));
 $('depth').addEventListener('input', () => view.setDepth($('depth').valueAsNumber));

samples/camera/index.html

@@ -50,7 +50,7 @@ <h1>camera</h1>
     <button id="photo" type="button" disabled>Take 3D photo</button>
     <button id="rec" type="button" disabled>Record</button>
     <label><input id="mirror" type="checkbox" checked> Mirror</label>
-    <label><input id="converge" type="checkbox"> Auto-converge</label>
+    <label><input id="converge" type="checkbox" checked> Auto-converge</label>
     <label>Depth <input id="depth" type="range" min="-1" max="1" step="0.05" value="0"></label>
     <label><input id="mono" type="checkbox" checked> Also a 2D copy</label>
   </div>