Repository navigation
Stacked on seedds/nxnative#1383: the game is told the device's HDR peak (1,600 nits by model), not a fixed 1,000 - #8
oliverames wants to merge 2 commits into
Conversation
…front_ios.m reads the hardware model; 1,600 nits on the modern panels), and DM_HDR_EDR_MAX overrides
QueryDisplayHdrStatus and the layer's mastering metadata now report the front end's model peak when it published one (1,600 nits for the iPhone 12 Pro and later), else the virtual 1,000. DM_HDR_EDR_MAX still caps it. Stacked on seedds#1383.
Codex Review Summary: ❌ FailedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
|
Folded into #9, which carries this change rebuilt on seedds#1176's HDR design (the HDR code this was stacked on, seedds#1383's, is a different implementation) together with the launcher rows and the packed-conversion probe. |
Stacked on seedds#1383 (not mergeable into
mainyet:mainhas no HDR presentation code, it lives in seedds#1383 / seedds#1249 / seedds#1176). Not fully tested. The base branchpr-1383-headis seedds#1383's head (e8ac7386); this PR adds two commits on it, so it can be rebased onto whichever HDR PR lands.8c646e2f(earlier):front_ios.mreads the hardware model and publishes the device's peak HDR brightness (1,600 nits for the iPhone 12 Pro and later, 1,200 for the iPhone 10 and 11, the M-series iPad XDR panels), and the layers' mastering metadata can top out at it when it is below the game's own 1,000;DM_HDR_EDR_MAX(a ratio of paper white) overrides.acaedf15: the game is also told that peak.QueryDisplayHdrStatusanswered a fixed 1,000, so a game tone-mapped to a 1,000-nit display on a panel that reaches 1,600; it now answersdm_hdr_mastering_peak()(the model peak when published, else 1,000), and the layer metadata carries the same value.DM_HDR_EDR_MAXstill caps it.Checked. Built into an iOS Release build and run on a physical iPhone 19,4 (potential EDR headroom 8.0, paper white 203, so about 1,624 nits): the log line reads
vi HDR: enabled (EDR display), virtual peak 1600 nits, and the owner reports the picture looks good in a race.Not checked. No side-by-side against the 1,000-nit value, no measurement of nits, no Mac run, and the unit tests were not run (the build tree has the PR's own docs-index failure at
e8ac7386,docs/agents/game-patches.mdout of date, unrelated to this change). A model missing from the table keeps 1,000. A brightness or heat cost of the higher EDR peak was not measured: the slow GPU state in the same runs is unexplained and HDR load has not been ruled out as a contributor (aDM_HDR=0run is planned).