feat(calls): change noise suppression mid-call without rejoining
CI / Build & Quality Checks (push) Skipped
CI / Secret scan (gitleaks) (push) Skipped
CI / Playwright smoke (e2e) (push) Skipped
CI / Secret scan (gitleaks) (pull_request) Successful in 2m5s
CI / Build & Quality Checks (pull_request) Successful in 3m1s
CI / Trigger Desktop Build (pull_request) Skipped
CI / Docker image build & smoke test (pull_request) Skipped
CI / Playwright smoke (e2e) (pull_request) Successful in 17m2s
CI / Build & Quality Checks (push) Skipped
CI / Secret scan (gitleaks) (push) Skipped
CI / Playwright smoke (e2e) (push) Skipped
CI / Secret scan (gitleaks) (pull_request) Successful in 2m5s
CI / Build & Quality Checks (pull_request) Successful in 3m1s
CI / Trigger Desktop Build (pull_request) Skipped
CI / Docker image build & smoke test (pull_request) Skipped
CI / Playwright smoke (e2e) (pull_request) Successful in 17m2s
Changing Noise suppression (off / browser-native / ML), the ML model, the gate or "stack browser NS" in Settings now applies to the call you're in. useCallDenoise sends the fork's io.lotus.set_denoise whenever those settings change; CallControl join-gates it like set_quality and re-sends it after an EC reconnect (the remounted fork handler starts again from the join URL). The "requires rejoining the call" note is replaced. Bumps Element Call to 0.28.0-lotus.4, which adds io.lotus.set_denoise and fixes mic re-captures silently keeping the old noiseSuppression (the mic level meter's clone held the old track open), the meter following the old device after a mic switch, and voiceIsolation staying on in off/ML modes (element-call#66). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017PCb59uY3UKqDqt7B6hzd5
This commit is contained in:
1 parent
7d31b2b2f7
commit
504899aa14
8 files changed
+130
-6
No files matched your search
@@ -0,0 +1,31 @@
|
||||
import { DenoiseModelId, NoiseSuppressionMode } from '../state/settings';
|
||||
|
||||
/**
|
||||
* Payload for the fork's `io.lotus.set_denoise` action, which switches noise
|
||||
* suppression on the live mic mid-call (browser suppressor on/off and/or the
|
||||
* ML model) instead of needing a rejoin.
|
||||
*/
|
||||
export type LotusDenoisePayload = {
|
||||
mode: NoiseSuppressionMode;
|
||||
model: DenoiseModelId;
|
||||
nativeNS: boolean;
|
||||
gate: boolean;
|
||||
gateThreshold: number;
|
||||
};
|
||||
|
||||
export type DenoiseSettings = {
|
||||
callNoiseSuppression?: NoiseSuppressionMode;
|
||||
callDenoiseModel?: DenoiseModelId;
|
||||
callDenoiseNativeNS?: boolean;
|
||||
callDenoiseGate?: boolean;
|
||||
callDenoiseGateThreshold?: number;
|
||||
};
|
||||
|
||||
/** Same defaults as the call URL built at join (useCallStart/createCallEmbed). */
|
||||
export const buildDenoisePayload = (s: DenoiseSettings): LotusDenoisePayload => ({
|
||||
mode: s.callNoiseSuppression ?? 'browser',
|
||||
model: s.callDenoiseModel ?? 'rnnoise',
|
||||
nativeNS: s.callDenoiseNativeNS ?? true,
|
||||
gate: s.callDenoiseGate ?? false,
|
||||
gateThreshold: s.callDenoiseGateThreshold ?? -45,
|
||||
});
|
||||
Reference in new issue
Block a user