pepepper/fluxer
1
0 Fork 0
mirror of https://github.com/fluxerapp/fluxer synced 2026-10-19 05:55:34 +09:00
Code Issues Packages Projects Releases Wiki Activity

fix(voice): stabilize voice input and noise suppression (#3087)

Browse Source
This commit is contained in:
Hampus authored and GitHub committed 2026-10-01 14:02:12 +02:00
1 parent 7c9564bcad
commit e3fcedbec5
98 files changed
+5610 -3433
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

fluxer_app/src/features/user/components/modals/tabs/UserVoiceTab.tsx
+28 -3
View File
@@ -33,11 +33,13 @@ import {useMediaPermission} from '@app/features/user/components/modals/tabs/hook
import styles from '@app/features/user/components/modals/tabs/UserVoiceTab.module.css';
import * as VoiceSettingsCommands from '@app/features/voice/commands/VoiceSettingsCommands';
import MediaEngine from '@app/features/voice/engine/MediaEngineFacade';
import {supportsVoiceOutputDeviceSelection} from '@app/features/voice/engine/VoiceSharedAudioContext';
import type VoiceSettings from '@app/features/voice/state/VoiceSettings';
import {
type ExternalAudioProcessorMatch,
findExternalProcessorForDevice,
} from '@app/features/voice/utils/ExternalAudioProcessor';
import {prefetchDeepFilterAssets} from '@app/features/voice/utils/noise_suppression/DeepFilter';
import type {VoiceNoiseSuppressionBackend} from '@app/features/voice/utils/noise_suppression/NoiseSuppressionBackends';
import {
getNoiseSuppressionChoiceValues,
@@ -48,6 +50,7 @@ import {
} from '@app/features/voice/utils/noise_suppression/NoiseSuppressionChoices';
import {
getNoiseSuppressionChoiceLabel,
getNoiseSuppressionFallbackMessage,
STEREO_MICROPHONE_DESCRIPTION_DESCRIPTOR,
STEREO_MICROPHONE_DESCRIPTOR,
} from '@app/features/voice/utils/noise_suppression/NoiseSuppressionLabels';
@@ -245,7 +248,10 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
const defaultPttCombo = getDefaultKeybind('voice_push_to_talk', i18n);
const inputHasLabels = hasDeviceLabels(inputDevices);
const effectiveInputDeviceId = resolveEffectiveDeviceId(inputDeviceId, inputDevices) ?? 'default';
const effectiveOutputDeviceId = resolveEffectiveDeviceId(outputDeviceId, outputDevices) ?? 'default';
const canSelectOutputDevice = supportsVoiceOutputDeviceSelection();
const effectiveOutputDeviceId = canSelectOutputDevice
? (resolveEffectiveDeviceId(outputDeviceId, outputDevices) ?? 'default')
: 'default';
const activeInputDevice = inputDevices.find((d) => d.deviceId === effectiveInputDeviceId) ?? null;
const activeInputLabel = activeInputDevice?.label || null;
const voiceProcessingMode = voiceSettings.getVoiceProcessingModeForDeviceLabel(activeInputLabel);
@@ -256,6 +262,9 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
useEffect(() => {
if (pttReleaseDelay !== selectedPttReleaseDelay) Keybind.setPushToTalkReleaseDelay(selectedPttReleaseDelay);
}, [pttReleaseDelay, selectedPttReleaseDelay]);
useEffect(() => {
prefetchDeepFilterAssets();
}, [voiceProcessingMode]);
const handleInputDeviceChange = (value: string) => {
VoiceSettingsCommands.update({inputDeviceId: value});
};
@@ -267,8 +276,13 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
[deviceState, i18n.locale],
);
const outputDeviceOptions = useMemo(
() => buildSettingsDeviceOptions(deviceState, 'audiooutput', i18n),
[deviceState, i18n.locale],
() =>
buildSettingsDeviceOptions(
canSelectOutputDevice ? deviceState : {...deviceState, outputDevices: []},
'audiooutput',
i18n,
),
[canSelectOutputDevice, deviceState, i18n.locale],
);
const resetSliderLabel = i18n._(RESET_SLIDER_TO_DEFAULT_VALUE_DESCRIPTOR);
const profileOptions: Array<RadioOption<VoiceProcessingMode>> = [
@@ -289,6 +303,7 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
},
];
const noiseSuppressionChoice = getSelectedNoiseSuppressionChoice();
const noiseSuppressionFallbackMessage = getNoiseSuppressionFallbackMessage(i18n);
const noiseSuppressionOptions: Array<ComboboxOption<VoiceNoiseSuppressionBackend>> =
getNoiseSuppressionChoiceValues().map((backend) => ({
value: backend,
@@ -496,6 +511,7 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
)}
<CompactComboboxRow<VoiceNoiseSuppressionBackend>
label={i18n._(VOICE_NOISE_SUPPRESSION_DESCRIPTOR)}
description={noiseSuppressionFallbackMessage}
value={noiseSuppressionChoice}
options={noiseSuppressionOptions}
onChange={setNoiseSuppressionChoice}
@@ -569,8 +585,14 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
/>
<CompactComboboxRow
label={i18n._(VOICE_OUTPUT_DEVICE_DESCRIPTOR)}
description={
!canSelectOutputDevice ? (
<Trans>Voice uses your system output device in this browser. Change it in your system settings.</Trans>
) : null
}
value={effectiveOutputDeviceId}
options={outputDeviceOptions}
disabled={!canSelectOutputDevice}
onChange={(value) => VoiceSettingsCommands.update({outputDeviceId: value})}
controlWidth="wide"
menuMinWidth={280}
@@ -660,6 +682,9 @@ export const VoiceTab: React.FC<VoiceTabProps> = observer(({voiceSettings, autoR
aria-label={i18n._(SELECT_VOICE_PROCESSING_DESCRIPTOR)}
data-flx="user.voice-tab.radio-group.voice-processing-mode-change"
/>
{voiceProcessingMode === 'voice' && noiseSuppressionFallbackMessage && (
<p className={styles.pttSettingDescription}>{noiseSuppressionFallbackMessage}</p>
)}
{voiceProcessingMode === 'voice' && (
<div className={styles.profileSubSection} data-flx="user.voice-tab.profile-sub-section">
{renderPttControls()}