fix(render): jitter the volumetric march per pixel (IGN); file #421 for the shadow transform buffer (Campaign VM VM5)

Fixed step-count volumetric ray marches (Low/Medium presets, e.g. 40 steps)
sample the same t values at every pixel, so a shadow-cascade edge crossing
lands at the same march step across a whole neighbourhood of pixels and
reads as visible concentric banding around light shafts. Offset each
pixel's march phase with interleaved gradient noise (Jimenez 2014,
fract(52.9829189 * fract(dot(gl_FragCoord.xy, vec2(0.06711056,
0.00583715))))) instead of the fixed 0.5 sub-step center: t = (step + ign)
/ steps. This decorrelates the step boundaries across pixels, turning the
banding into fine per-pixel noise that the existing quarter/half-res
upsample already averages away. Pack-on only — atmospheric_volumetric is
never used on the retail default path; extinction, colour, and early-outs
are unchanged.

Regenerated SPIR-V via tools/compile-shaders.ps1: only
atmospheric_volumetric.frag.spv and shaders.manifest.json's source hash
for that stage changed.

Also files #421 (docs/ISSUES.md) for Campaign AR review finding F7: the
directional-shadow pass's DirectionalShadowTransformBufferSet duplicates
the main pass's N.5 instance SSBO instead of binding it, to be resolved
once the planned GPU-culling step for shadow cascades lands. Updates the
VM5 ledger row and outcome note in
docs/plans/2026-08-22-visualmaster-campaign.md.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

docs/ISSUES.md

@@ -24,6 +24,33 @@ What does NOT go here:
 - Every session: scan OPEN issues at start; promote/close anything we touched during the session before ending.
 - Promoting to a Phase: mark as `DONE (promoted to Phase X)` + commit SHA where the Phase entry landed.
 
+## #421 — Directional-shadow pass uploads its own transform buffer instead of binding the main pass's instance SSBO
+
+**Status:** OPEN — filed 2026-08-22 at Campaign VM VM5 (post-M7 / with GPU culling).
+**Component:** rendering / render packs (Campaign AR Tier 2)
+
+**Description:** `DirectionalShadowTransformBufferSet`
+(`src/AcDream.App/Rendering/DirectionalShadowTransformBufferSet.cs`)
+re-composes every caster's world matrix from the same
+`MeshRef.PartTransform x LocalToWorld` inputs the main pass uses (same
+`WbDrawDispatcher.ComposePartWorldMatrix`, so there is no second pose —
+Campaign AR constraint 5 is honoured in spirit) and uploads them into a
+second GPU buffer each frame for the animated subset. This is where the
+65,536-matrix binding ceiling regression lived (fixed in Campaign AR
+Stage 1). Cost is measured small (0.11–0.21 ms CPU incremental on the
+reference matrix) but it is a duplicate of the N.5 instance SSBO the main
+pass already binds.
+
+**What to do:** when the planned GPU-culling step for shadow cascades
+lands (Campaign AR plan, "If that is too expensive, the next permitted
+step is GPU culling"), bind the main pass's instance transform SSBO for
+the caster pass instead of maintaining `DirectionalShadowTransformBufferSet`,
+and delete the set. Not before — the shared buffer's lifetime/flight
+rules differ (main pass ring vs retained caster topology) and unifying
+them without the culling redesign would just move the ceiling.
+Reference: Campaign AR review F7,
+`docs/research/2026-08-22-campaign-ar-review.md`.
+
 ## #420 — Client crashes on the character-select screen (`UiButton.OnDraw` null media-state key)
 
 **Status:** ✅ FIXED 2026-08-19, root cause proven by a reverting test.