test(hdr-regression): tighten Window C maxFrameFailures budget after Chunk 1 fix

Window C (direct <video> opacity tween) was previously listed as a known
failure with a maxFrameFailures budget of 30 to absorb expected drift
until Chunk 1 (opacity pipeline bugs) landed.

After the Chunk 1 fix, the regression test passes against the existing
golden with 0 failed frames, confirming Window C now renders correctly.
Regenerating the golden produces byte-identical output (HEVC encoding
is deterministic and the opacity fix doesn't perturb pixels at the
PSNR ≥ 28 checkpoint threshold).

Tighten maxFrameFailures from 30 → 5 to leave only a small budget for
encoder noise. Window F (transform + border-radius) remains pending
Chunk 4; its broken state is currently baked into the golden, so the
suite is green and Chunk 4's regen will catch any drift.

Update README.md to reflect Window C is fixed and the tightened budget.

Author: vanceingalls

packages/producer/tests/hdr-regression/README.md

@@ -8,16 +8,16 @@ windows back-to-back.
 
 ## What it covers
 
-| Window | Time          | Shape                                     | Expected            |
-| ------ | ------------- | ----------------------------------------- | ------------------- |
-| A      | 0.0 – 2.0 s   | Baseline HDR video + DOM overlay          | pass                |
-| B      | 2.0 – 4.5 s   | Wrapper opacity fade around HDR video     | pass                |
-| C      | 4.5 – 7.0 s   | Direct `<video>` opacity tween            | **known fail (C1)** |
-| D      | 7.0 – 9.5 s   | DOM → HDR → DOM z-order sandwich          | pass                |
-| E      | 9.5 – 12.0 s  | Two HDR videos side-by-side (same source) | pass                |
-| F      | 12.0 – 14.5 s | Transform + scale + border-radius         | **known fail (C4)** |
-| G      | 14.5 – 17.0 s | `object-fit: contain` letterbox           | pass                |
-| H      | 17.0 – 20.0 s | Shader transition (HDR video → HDR image) | pass                |
+| Window | Time          | Shape                                     | Expected                |
+| ------ | ------------- | ----------------------------------------- | ----------------------- |
+| A      | 0.0 – 2.0 s   | Baseline HDR video + DOM overlay          | pass                    |
+| B      | 2.0 – 4.5 s   | Wrapper opacity fade around HDR video     | pass                    |
+| C      | 4.5 – 7.0 s   | Direct `<video>` opacity tween            | pass (fixed by Chunk 1) |
+| D      | 7.0 – 9.5 s   | DOM → HDR → DOM z-order sandwich          | pass                    |
+| E      | 9.5 – 12.0 s  | Two HDR videos side-by-side (same source) | pass                    |
+| F      | 12.0 – 14.5 s | Transform + scale + border-radius         | **known fail (C4)**     |
+| G      | 14.5 – 17.0 s | `object-fit: contain` letterbox           | pass                    |
+| H      | 17.0 – 20.0 s | Shader transition (HDR video → HDR image) | pass                    |
 
 The test pins the contract that:
 
@@ -39,17 +39,18 @@ The test pins the contract that:
 
 ## Known failures
 
-Windows **C** (direct `<video>` opacity) and **F** (transform + border-radius
-on the video itself) are intentionally **expected to fail** until the
-corresponding follow-up chunks land:
-
-- **C** — fixed by chunk 1 (videoFrameInjector opacity-walk bugs).
-- **F** — fixed by chunk 4 (transform + clipping pipeline).
-
-`maxFrameFailures` is set high enough to absorb both windows. The intent is
-that the suite stays green while we ship the fixes, **and tightens
-automatically as soon as we regenerate goldens** (chunk 1 → drop C tolerance,
-chunk 4 → drop F tolerance, eventually reaching `maxFrameFailures: 0`).
+Window **F** (transform + border-radius on the video itself) is intentionally
+**expected to fail** until chunk 4 (transform + clipping pipeline) lands. Its
+broken state is currently baked into the golden, so the suite is green; when
+chunk 4 fixes the rendering path, the golden will be regenerated to match the
+correct output.
+
+Window **C** (direct `<video>` opacity) was previously known-failing — it is
+now fixed by chunk 1 (videoFrameInjector + screenshotService no longer clobber
+GSAP-controlled opacity), and the golden has been regenerated to match the
+correct output. `maxFrameFailures` was tightened from 30 → 5 to leave only a
+small budget for HEVC encoder noise; any drift larger than that will be
+caught immediately.
 
 ## Fixtures
 

packages/producer/tests/hdr-regression/meta.json

@@ -1,9 +1,9 @@
 {
   "name": "hdr-regression",
-  "description": "Comprehensive HDR10 (BT.2020 PQ) regression suite. Eight back-to-back 'windows' (A–H) lock down the layered HDR compositing pipeline against the most common composition shapes: baseline pass-through, wrapper-vs-direct opacity, z-order sandwiches, side-by-side multi-source, transform + border-radius, object-fit letterbox, and a shader transition between an HDR video and an HDR PQ image. Two windows (C, direct <video> opacity; F, transform + border-radius) are intentionally known-failing until follow-up chunks land — maxFrameFailures absorbs them so the suite stays green and any change in the failure footprint is caught immediately.",
+  "description": "Comprehensive HDR10 (BT.2020 PQ) regression suite. Eight back-to-back 'windows' (A–H) lock down the layered HDR compositing pipeline against the most common composition shapes: baseline pass-through, wrapper-vs-direct opacity, z-order sandwiches, side-by-side multi-source, transform + border-radius, object-fit letterbox, and a shader transition between an HDR video and an HDR PQ image. Window C (direct <video> opacity) is fixed (Chunk 1). Window F (transform + border-radius) is still pending Chunk 4 — its broken state is baked into the current golden, so the suite passes; when Chunk 4 lands, the golden will be regenerated and any drift will be caught immediately.",
   "tags": ["regression", "hdr"],
   "minPsnr": 28,
-  "maxFrameFailures": 30,
+  "maxFrameFailures": 5,
   "minAudioCorrelation": 0,
   "maxAudioLagWindows": 1,
   "renderConfig": {