Repository navigation
P1I: premium 10CH console drives real native faders/PAN/MUTE/SOLO without VST3 rebuild - #127
Conversation
… without VST3 rebuild
…ficients for P1H compatibility
…ficients for P1H compatibility
…/MUTE/SOLO event contract
…/MUTE/SOLO event contract
…nd live DSP fader controls
…E/SOLO closure, zero RT heap and epoch fencing
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. |
…or native 10CH control token
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b12b4059f3
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".
| { | ||
| auto lease=publisher.acquire(); | ||
| assert(lease.get(4)&&lease.get(4)->prepared->identity==42); | ||
| assert(near(lease.get(4)->prepared->params.at(3).left.load(),.5f)); |
There was a problem hiding this comment.
Assert the panned coefficient rather than the unpanned gain
The update above sets gain to 0.5, pan to 0.2, and mono to false, so RealtimeStripControls::set() correctly computes the left coefficient as 0.5 * (1 - 0.2) = 0.4. Expecting 0.5 makes the new test abort deterministically; compiling and running it locally reproduces the failure. Since this target is registered with CTest and included in .github/workflows/p1-windows.yml, the native test gate cannot pass even after compilation succeeds. Change the expected coefficient to 0.4 or use centered pan.
AGENTS.md reference: AGENTS.md:L26-L26
Useful? React with 👍 / 👎.
| // P1I: native routing token rejects UI gestures from stale device graphs. | ||
| const liveToken=window.__armixRouteToken; | ||
| if(ch>=0&&ch<10&&liveToken?.epoch>0&&liveToken?.revision>0){ | ||
| event.routeEpoch=liveToken.epoch;event.routeRevision=liveToken.revision; | ||
| } |
There was a problem hiding this comment.
Synchronize the standalone script with the changed app source
This changes the canonical application script without updating its embedded copy in AR-MIX-Live-FX-Host-Standalone-v0.3.html. Consequently, python tests/gui_baseline_lock_test.py fails at its script-parity assertion, blocking the GUI workflow, which explicitly runs that test in .github/workflows/ui-smoke.yml. Update the standalone script alongside web/app.js to preserve the required distributable parity.
AGENTS.md reference: AGENTS.md:L15-L15
Useful? React with 👍 / 👎.
| for(let i=0;i<Math.min(10,lab.channelPeaks.length);++i) | ||
| paintMeter(document.querySelector('[data-meter="'+i+'"]'), | ||
| percent(Number(lab.channelPeaks[i])||0)); |
There was a problem hiding this comment.
Address the stereo meter elements on channels 7–10
With Routing LAB active, channels 7–10 receive no meter updates: channels initializes those strips with stereo: true, and makeStrip() creates meter keys such as 6L and 6R, not 6. This selector therefore returns null for all four strips, and paintMeter() silently skips them even when their native audio is playing. Resolve the actual mono/stereo meter elements when applying telemetry; if separate L/R readings are intended, expose corresponding native measurements rather than leaving both meters at zero.
Useful? React with 👍 / 👎.
…serve arbitrary gain
P1I: 10-channel premium console now drives real per-strip native DSP (not a second LAB GUI)
Baseline reconciled: latest
main@758ec23e959305115d7fc4e23f7c9f73ddad43c8after P1H; Windows, 43/43 CTests and Chromium green. Preserves GUI v0.8 latest CH/MASTER strip geometry, BTN-180, Orbit, CH1 native FX rack, backup, 10CH hardware masks, SEND/BUS graph, independent VST3 ownership and zero-FX diagnostic #15.Functional milestone
RealtimeStripControlsprecomputes pan/fader coefficients on the message thread; callback reads atomics once per block and interpolates gains and mute/solo PRE-send gates over bounded 64 samples; SOLO closure preserves real DAG ancestors and descendants. No VST3 reconstruction, no callback detachment, locks, allocations or device restart during ordinary fader/pan/mute/solo actions.docs/P1I_UNIFIED_10CH_REALTIME_CONSOLE_20261028.md.Honest limits: still opt-in LAB, BUS consumes a strip, one VST3 per CH, no independent AUX/CUE hardware, PDC, full MixSession SSOT/route persistence, or hardware #15 crackle FAT; vendor plugin RT guarantees cannot be proven by synthetic tests. This is integration M1, not claim of Adobe Audition/Studio One equivalence.
Merge gate: exact-head Windows native executable, all CTests including new
P1ITenChannelLiveFaderPanMuteSoloAndStableFx, packaged artifact, Chromium; verify main HEAD hasn't moved. Never merge stale #115/#117 wholesale.