[miguel-heygen]

What

HDR and shader-transition renders now report encoder_wait_avg_ms like the streaming and segmented paths do. Before, it was absent for them.

Why

#6839 added the encoder's drain wait to render telemetry, but the layered-composite stage (captureHdrStage.ts) also writes through spawnStreamingEncoder and only kept durationMs from its result, and the orchestrator's HDR branch only copied encodeMs. So those renders had no encoder wait even though the encoder measured one.

How

runCaptureHdrStage returns the encoder's drainWaitMs as encoderDrainWaitMs, and the orchestrator's HDR branch records it in perfStages.encoderDrainWaitMs, the same as the other two paths.

Size

Under 100 lines on purpose: it closes the one capture path #6839 missed, with its test.

Related work

Follows #6839.

Test plan

• Unit tests: captureStreamingStage.test.ts runCaptureHdrStage forwards useGpu=%s to the HDR encoder now also checks the stage returns the fake encoder's drain wait.
• fails without the fix: on feat(telemetry): render_complete splits capture time into seek, page, screenshot and encoder wait #6839's head those three cases fail with undefined instead of 7.
• Passes 3 runs in a row; the whole file passes (30). oxfmt, oxlint, producer tsc, fallow audit and the comment ratchet clean, on Linux.
• Manual testing (not applicable: the HDR branch only copies one more field the encoder already measures; the encoder side ran in a real render in feat(telemetry): render_complete splits capture time into seek, page, screenshot and encoder wait #6839)
• Documentation updated (not applicable)
• Comments follow CONTRIBUTING.md "Comments"

[heygengenesis]

Verdict: APPROVE · 0 blocker · 0 important · 0 nits · head f5eef5f4. This closes the HDR gap from my #6839 review.

The named case (forwards useGpu=%s to the HDR encoder) fails 3 of 3 with captureHdrStage.ts reverted to #6839's head and passes at this head (30/30). Producer tsc is clean.

This PR is stacked on feat/render-capture-phase-timings, so CI, CodeQL and Windows don't run on it (they trigger only for main). They will gate it after the rebase onto main. The runs that did fire (regression, preview-regression, Player perf) are green.

The base branch was changed.

[github-actions]

Edit accuracy: accurate 271 (base branch 271), smooth 126 of those

The gate passes.
Smoothness is reported in the artifact, not gated. A case fails only if it fails 2 of 3 runs.

Quarantined, measured but not gated (0)

Unstable (2)

• seqplainkeys-none-center-r0-root-z100: tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo false, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true / tracking 0.06, pressJump 0, drop 0, reload 0, render 0.03, renderKey -, undo true, teleport true
• seqplainkeys-none-pct-r0-nested-z100: tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true / tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true / tracking 0.05, pressJump 0, drop 0, reload 0, render 0.02, renderKey -, undo true, teleport true

[heygengenesis]

Verdict: APPROVE · 0 resolved, 0 open, 0 new · f5eef5f4 → e8d87027. Rebased onto main with the same patch (only a hunk offset moved). The file passes 30/30, and the main-targeted CI, CodeQL and Windows checks are now green.