Repository navigation
P0 First Sound: visible native START AUDIO, real post-fader PCM proof and direct-monitor isolation - #257
Conversation
…se external direct monitoring
Codex Review SummaryThis 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. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 77f2513c50
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const mixer=!routed?"MIXER · NOT READY":!d.monitor?"MIXER · WAIT ARM": | ||
| mixed?"POST-FADER PCM":"POST-FADER SILENT"; |
There was a problem hiding this comment.
Preserve mixer PCM evidence when only secondary output is armed
When OUTPUT1 is armed while physical output is muted (secondaryOutputArmed=true, monitor=false), this reports MIXER · WAIT ARM even when native channel peaks prove post-fader PCM. LiveOutputRouteGate::evaluate() explicitly permits rendering in this configuration, and LiveAudioEngine.h publishes that PCM to the secondary endpoint before clearing physical outputs. The new diagnostic therefore hides valid mixer evidence and directs an already-playing operator to arm another output unnecessarily. Base the mixer indicator on native PCM/render eligibility independently of physical ARM, and account for the secondary destination in the accompanying guidance.
AGENTS.md reference: AGENTS.md:L30-L30
Useful? React with 👍 / 👎.
…onitor and gated visible START
…ambiguous duplicate selector
P0 physical first sound — do not confuse capture, native route ACK, and an audible mixer
Field evidence (5 user screenshots): Windows Audio 48 kHz / 480; MAONO capture and patch CH07 stereo 1/2 is native LIVE ACK but both AUDIO SETUP and SIGNAL CHECK explicitly read OUTPUT SAFE-MUTED. The user hears audio anyway. A muted native hardware MAIN cannot have produced that sound. Possible independent MAONO direct monitoring, Windows Listen/playback, or another output route must be tested; never fake that PCM crossed the mixer just because capture is present.
Concrete fix
dryRoutingLab.channelPeaksscoped to actual patched channels (not raw input or browser demo VUs); MAIN stage uses nativepeakand ARM gate.__arMixStartLiveAudio, native route APPLY→ACK→ARM. Both share onestartAudioBlocked()guard for pending draft, disconnected/stale physical I/O, endpoint switch and native availability. Emergency MUTE is always allowed.web/mixer-io.jsand CSS into standalone 858 KB HTML; native/preview parity maintained. Compact CSS, no blur or animations, zero new audio callbacks/workers.Acceptance
python scripts/build_ui_css.py --check, Chromium UI contract/smoke, existing Windows native CI at exact PR SHA.Adjacent engine PRs #251 and #256 remain independent.