unom/punktfunk
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases 20 Wiki Activity

fix(video): honor the signaled CSC matrix end-to-end + tvOS HDR presentation

Browse Source 
Clients derive Y'CbCr->RGB from the stream's SIGNALED matrix x range x depth
via shared csc rows (Rust csc_rows + Swift CscRows) instead of hardcoded
709/2020 - a BT.601-signaled stream (a Linux host's RGB-input NVENC) no longer
renders with a constant hue error. Host-side signaling made honest across
NVENC/VAAPI/openh264/GameStream and the session plan's chroma/bit-depth.
Decoded color-bar fixtures (601/709 x limited/full) pin the math in tests on
both cores.

Same presenter, tvOS HDR: tvOS has no Metal EDR API and a bare PQ colorspace
tag composites UNTONE-MAPPED (the "overblown" Apple TV report), so HDR now
splits on the display's live EDR headroom - PQ passthrough when the
per-session AVDisplayManager mode switch landed (a real HDR10 output
tone-maps itself), else an in-shader PQ->SDR tone-map (203-nit reference
white, extended-Reinhard 1000-nit knee, 2020->709) into the proven SDR layer
config. The 10-bit stream keeps its full decode depth either way.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Enrico Bühler
2026-07-10 16:58:11 +02:00
parent db49904c6d
commit 1fcf9e11ec
26 changed files with 2268 additions and 409 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/pf-presenter/shaders/nv12_csc.frag
+11 -1
View File
@@ -62,7 +62,17 @@ vec3 srgb_oetf(vec3 c) {
}
void main() {
vec3 yuv = vec3(texture(u_y, v_uv).r, texture(u_c, v_uv).rg);
// 4:2:0 chroma is left-cosited (H.273 type 0 — the default inference when unsignaled, and
// what the hosts produce), but sampling the half-res plane at the luma UV assumes CENTER
// siting — a ~0.5-luma-px rightward chroma shift on hard colored edges. Offset +0.25 chroma
// texels to re-align (the same correction the Apple/Windows clients apply). Self-disables
// when the plane widths match (a full-size 4:4:4 chroma plane needs no correction).
vec2 cuv = v_uv;
int cw = textureSize(u_c, 0).x;
if (cw < textureSize(u_y, 0).x) {
cuv.x += 0.25 / float(cw);
}
vec3 yuv = vec3(texture(u_y, v_uv).r, texture(u_c, cuv).rg);
vec3 rgb = vec3(
dot(pc.r0.xyz, yuv) + pc.r0.w,
dot(pc.r1.xyz, yuv) + pc.r1.w,