fix(dictation): show Listening while the speech model loads, and link start toasts to Voice settings (#39136)

* fix(dictation): say Listening once the mic is recording, and translate start toasts with an Open Settings action * refactor(dictation): derive the indicator label from the listening flag
2026-11-15 02:17:31 +00:00 · 2026-11-07 00:21:33 +00:00
parent a3a0e68380
commit ea32ccb0ec
15 changed files with 140 additions and 52 deletions
@@ -1,4 +1,4 @@
-import { useEffect, useRef, useCallback } from 'react'
+import { useEffect, useRef, useCallback, useState } from 'react'
 import { useAppStore } from '@/store'
 import { useAudioCapture } from '@/hooks/use-audio-capture'
 import { toast } from 'sonner'
@@ -11,7 +11,11 @@ import {
 import { formatFinalTranscriptSegment } from '../../../../shared/dictation-transcript-spacing'
 import { recordStoppedSession, waitForStoppedSession } from './dictation-stopped-sessions'
 import { translate } from '@/i18n/i18n'
-import { showDictationStartErrorToast } from './dictation-start-error-toast'
+import {
+  showDictationDisabledToast,
+  showDictationNoModelToast,
+  showDictationStartErrorToast
+} from './dictation-start-error-toast'
 import { useHoldDictationGesture } from './use-hold-dictation-gesture'
 import { DICTATION_CONTROL_EVENT, type DictationControlAction } from './dictation-control-events'
 import { publishDictationMeter } from './dictation-meter-store'
@@ -46,6 +50,7 @@ export function DictationController() {
  // Push-to-talk restarts per utterance; report a missing preferred microphone once per preference.
  const micFallbackNotifiedForRef = useRef<string | null>(null)
  const stopDictationRef = useRef<(() => void) | null>(null)
+  const [captureLive, setCaptureLive] = useState(false)

  const drainStoppedSession = useCallback((sessionId: string) => {
    void waitForStoppedSession(sessionId, stoppedSessionIdsRef, stoppedResolversRef)
@@ -97,25 +102,12 @@ export function DictationController() {

    const modelId = settings?.voice?.sttModel
    if (!modelId) {
-      toast('No speech model selected. Download one in Settings > Voice.', {
-        // Held keys and repeated presses replace this toast instead of stacking it.
-        id: 'dictation-no-model',
-        action: {
-          label: translate(
-            'auto.components.dictation.DictationController.bb7f599ee7',
-            'Open Settings'
-          ),
-          onClick: () => {
-            useAppStore.getState().openSettingsTarget({ pane: 'voice', repoId: null })
-            useAppStore.getState().openSettingsPage()
-          }
-        }
-      })
+      showDictationNoModelToast()
      return
    }

    if (!settings?.voice?.enabled) {
-      toast('Voice dictation is disabled. Enable it in Settings > Voice.')
+      showDictationDisabledToast()
      return
    }

@@ -130,11 +122,12 @@ export function DictationController() {
    insertedFinalTranscriptRef.current = ''
    dictationStateRef.current = 'starting'
    setDictationState('starting')
+    setCaptureLive(false)

    let captureStarted = false

    try {
-      // Capture and buffer before slow worker startup so speech during Starting is retained.
+      // Capture and buffer before slow worker startup so speech during model load is retained.
      const preferredMicrophoneDeviceId = settings?.voice?.microphoneDeviceId ?? null
      const captureResult = await startCapture({
        bufferAudio: true,
@@ -181,6 +174,7 @@ export function DictationController() {
        insertionTargetRef.current = null
        return
      }
+      setCaptureLive(true)

      await window.api.speech.startDictation(modelId, undefined, sessionId)
      if (dictationRunRef.current !== runId) {
@@ -409,5 +403,5 @@ export function DictationController() {
    }
  }, [setPartialTranscript, setDictationState, stopCapture, discardBufferedAudio])

-  return <DictationIndicator />
+  return <DictationIndicator captureLive={captureLive} />
 }
@@ -74,7 +74,7 @@ describe('DictationIndicator', () => {
    }
    document.addEventListener(DICTATION_CONTROL_EVENT, listener)

-    render(<DictationIndicator />)
+    render(<DictationIndicator captureLive />)
    fireEvent.click(screen.getByRole('button', { name: 'Stop dictation' }))

    document.removeEventListener(DICTATION_CONTROL_EVENT, listener)
@@ -10,7 +10,8 @@ import { dispatchDictationControl } from './dictation-control-events'
 import { DictationGrapes } from './DictationGrapes'
 import { useDictationMeter } from './dictation-meter-store'

-export function DictationIndicator() {
+// captureLive: the mic is recording (and buffering) while the speech engine is still starting.
+export function DictationIndicator({ captureLive }: { captureLive: boolean }) {
  const dictationState = useAppStore((state) => state.dictationState)
  const partialTranscript = useAppStore((state) => state.partialTranscript)
  const dictationMeter = useDictationMeter()
@@ -22,15 +23,15 @@ export function DictationIndicator() {
    return null
  }

-  const isListening = dictationState === 'listening'
+  const isLoadingModel = dictationState === 'starting' && captureLive
+  const isListening = dictationState === 'listening' || isLoadingModel
  const isClipping = isListening && dictationMeter.isClipping
  const isSpeaking = isListening && dictationMeter.isSpeaking && !isClipping
-  const lifecycleLabel =
-    dictationState === 'starting'
+  const lifecycleLabel = isListening
+    ? translate('auto.components.dictation.DictationIndicator.3de5a129e7', 'Listening')
+    : dictationState === 'starting'
      ? translate('auto.components.dictation.DictationIndicator.7f3660a7ba', 'Starting mic…')
-      : dictationState === 'stopping'
-        ? translate('auto.components.dictation.DictationIndicator.f082d0cb9d', 'Processing…')
-        : translate('auto.components.dictation.DictationIndicator.3de5a129e7', 'Listening')
+      : translate('auto.components.dictation.DictationIndicator.f082d0cb9d', 'Processing…')
  const label = isClipping
    ? translate('auto.components.dictation.DictationIndicator.4977162383', 'Too loud')
    : isSpeaking
@@ -62,11 +63,19 @@ export function DictationIndicator() {
        <DictationGrapes
          level={dictationMeter.level}
          active={dictationState !== 'stopping'}
-          transitioning={dictationState !== 'listening'}
+          transitioning={!isListening}
        />
        <span aria-hidden className="min-w-0 truncate font-medium">
          {label}
        </span>
+        {isLoadingModel ? (
+          <span aria-hidden className="min-w-0 truncate text-xs text-muted-foreground">
+            {translate(
+              'auto.components.dictation.DictationIndicator.loadingModel',
+              'Loading speech model…'
+            )}
+          </span>
+        ) : null}
        <span role="status" aria-live="polite" aria-atomic="true" className="sr-only">
          {announcedLabel}
        </span>
@@ -4,9 +4,35 @@ import { translate } from '@/i18n/i18n'
 import { compactIpcErrorMessage } from '@/lib/ipc-error'
 import { hostDictationErrorMessage } from '../../../../shared/host-dictation-error-messages'

-function openVoiceSettings(): void {
-  useAppStore.getState().openSettingsTarget({ pane: 'voice', repoId: null })
-  useAppStore.getState().openSettingsPage()
+function openVoiceSettingsAction(): { label: string; onClick: () => void } {
+  return {
+    label: translate('auto.components.dictation.DictationController.bb7f599ee7', 'Open Settings'),
+    onClick: () => {
+      useAppStore.getState().openSettingsTarget({ pane: 'voice', repoId: null })
+      useAppStore.getState().openSettingsPage()
+    }
+  }
+}
+
+export function showDictationNoModelToast(): void {
+  toast(
+    translate(
+      'auto.components.dictation.DictationController.noModel',
+      'No speech model selected. Download one in Settings > Voice.'
+    ),
+    // Held keys and repeated presses replace this toast instead of stacking it.
+    { id: 'dictation-no-model', action: openVoiceSettingsAction() }
+  )
+}
+
+export function showDictationDisabledToast(): void {
+  toast(
+    translate(
+      'auto.components.dictation.DictationController.disabled',
+      'Voice dictation is disabled. Enable it in Settings > Voice.'
+    ),
+    { action: openVoiceSettingsAction() }
+  )
 }

 export function showDictationStartErrorToast(message: string): void {
@@ -21,17 +47,21 @@ export function showDictationStartErrorToast(message: string): void {
      )
    )
  } else if (message.includes('not ready')) {
-    toast('Speech model not ready. Download it in Settings > Voice.')
+    toast(
+      translate(
+        'auto.components.dictation.DictationController.notReady',
+        'Speech model not ready. Download it in Settings > Voice.'
+      ),
+      { action: openVoiceSettingsAction() }
+    )
  } else if (message.includes('Unknown model')) {
-    toast('Selected model is no longer available. Please choose another in Settings > Voice.', {
-      action: {
-        label: translate(
-          'auto.components.dictation.DictationController.bb7f599ee7',
-          'Open Settings'
-        ),
-        onClick: openVoiceSettings
-      }
-    })
+    toast(
+      translate(
+        'auto.components.dictation.DictationController.unknownModel',
+        'Selected model is no longer available. Please choose another in Settings > Voice.'
+      ),
+      { action: openVoiceSettingsAction() }
+    )
  } else {
    toast.error(
      translate(
@@ -16917,6 +16917,10 @@
          "55127a3706": "Diktat fehlgeschlagen: {{value0}}",
          "bb7f599ee7": "Einstellungen öffnen",
          "2d5b9fabf9": "Mikrofonzugriff verweigert. Zugriff in den Systemeinstellungen erlauben und anschließend Orca neu starten.",
+          "noModel": "Kein Sprachmodell ausgewählt. Laden Sie eines unter Einstellungen > Spracheingabe herunter.",
+          "disabled": "Spracheingabe ist deaktiviert. Aktivieren Sie sie unter Einstellungen > Spracheingabe.",
+          "notReady": "Sprachmodell nicht bereit. Laden Sie es unter Einstellungen > Spracheingabe herunter.",
+          "unknownModel": "Das ausgewählte Modell ist nicht mehr verfügbar. Wählen Sie unter Einstellungen > Spracheingabe ein anderes aus.",
          "5d2c3e7ae3": "Keine Sprache erkannt.",
          "micFallback": "Ausgewähltes Mikrofon nicht verfügbar. Systemstandard wird verwendet.",
          "micDisconnected": "Mikrofon getrennt. Diktat gestoppt."
@@ -16927,7 +16931,8 @@
          "f082d0cb9d": "Wird verarbeitet…",
          "3de5a129e7": "Hört zu",
          "4977162383": "Zu laut",
-          "25f2b7a6a5": "Sprache erkannt"
+          "25f2b7a6a5": "Sprache erkannt",
+          "loadingModel": "Sprachmodell wird geladen…"
        }
      },
      "dashboard": {
@@ -17341,6 +17341,10 @@
          "55127a3706": "Dictation failed: {{value0}}",
          "bb7f599ee7": "Open Settings",
          "2d5b9fabf9": "Microphone access denied. Grant access in system settings, then restart Orca.",
+          "noModel": "No speech model selected. Download one in Settings > Voice.",
+          "disabled": "Voice dictation is disabled. Enable it in Settings > Voice.",
+          "notReady": "Speech model not ready. Download it in Settings > Voice.",
+          "unknownModel": "Selected model is no longer available. Please choose another in Settings > Voice.",
          "5d2c3e7ae3": "No speech detected.",
          "micFallback": "Selected microphone unavailable. Using system default.",
          "micDisconnected": "Microphone disconnected. Dictation stopped."
@@ -17351,7 +17355,8 @@
          "f082d0cb9d": "Processing…",
          "3de5a129e7": "Listening",
          "4977162383": "Too loud",
-          "25f2b7a6a5": "Speaking"
+          "25f2b7a6a5": "Speaking",
+          "loadingModel": "Loading speech model…"
        }
      },
      "dashboard": {
@@ -16646,6 +16646,10 @@
          "55127a3706": "Dictado fallido: {{value0}}",
          "bb7f599ee7": "Abrir configuración",
          "2d5b9fabf9": "Acceso al micrófono denegado. Conceda acceso en la configuración del sistema, luego reinicie Orca.",
+          "noModel": "No hay ningún modelo de voz seleccionado. Descargue uno en Configuración > Voz.",
+          "disabled": "El dictado por voz está desactivado. Actívelo en Configuración > Voz.",
+          "notReady": "El modelo de voz no está listo. Descárguelo en Configuración > Voz.",
+          "unknownModel": "El modelo seleccionado ya no está disponible. Elija otro en Configuración > Voz.",
          "5d2c3e7ae3": "No se detectó ninguna voz.",
          "micFallback": "Micrófono seleccionado no disponible. Se usará el predeterminado del sistema.",
          "micDisconnected": "Micrófono desconectado. Dictado detenido."
@@ -16656,7 +16660,8 @@
          "f082d0cb9d": "Procesando…",
          "3de5a129e7": "Escuchando",
          "4977162383": "Volumen demasiado alto",
-          "25f2b7a6a5": "Hablando"
+          "25f2b7a6a5": "Hablando",
+          "loadingModel": "Cargando modelo de voz…"
        }
      },
      "dashboard": {
@@ -16870,6 +16870,10 @@
          "55127a3706": "Échec de la dictée : {{value0}}",
          "bb7f599ee7": "Ouvrir les paramètres",
          "2d5b9fabf9": "Accès au micro refusé. Accordez l'autorisation dans les réglages système, puis redémarrez Orca.",
+          "noModel": "Aucun modèle vocal sélectionné. Téléchargez-en un dans Paramètres > Voix.",
+          "disabled": "La dictée vocale est désactivée. Activez-la dans Paramètres > Voix.",
+          "notReady": "Le modèle vocal n'est pas prêt. Téléchargez-le dans Paramètres > Voix.",
+          "unknownModel": "Le modèle sélectionné n'est plus disponible. Choisissez-en un autre dans Paramètres > Voix.",
          "5d2c3e7ae3": "Aucune parole détectée.",
          "micFallback": "Micro sélectionné indisponible. Utilisation du micro par défaut du système.",
          "micDisconnected": "Micro déconnecté. Dictée arrêtée."
@@ -16880,7 +16884,8 @@
          "f082d0cb9d": "Traitement…",
          "3de5a129e7": "Écoute en cours",
          "4977162383": "Trop fort",
-          "25f2b7a6a5": "Vous parlez"
+          "25f2b7a6a5": "Vous parlez",
+          "loadingModel": "Chargement du modèle vocal…"
        }
      },
      "dashboard": {
@@ -16244,6 +16244,10 @@
          "55127a3706": "ディクテーションが失敗しました: {{value0}}",
          "bb7f599ee7": "設定を開く",
          "2d5b9fabf9": "マイクへのアクセスが拒否されました。システム設定でアクセスを許可し、Orca を再起動。",
+          "noModel": "音声モデルが選択されていません。設定 > 音声 からダウンロードしてください。",
+          "disabled": "音声ディクテーションは無効です。設定 > 音声 で有効にしてください。",
+          "notReady": "音声モデルの準備ができていません。設定 > 音声 からダウンロードしてください。",
+          "unknownModel": "選択したモデルは使用できなくなりました。設定 > 音声 で別のモデルを選択してください。",
          "5d2c3e7ae3": "音声が検出されませんでした。",
          "micFallback": "選択したマイクが使用できません。システムのデフォルトを使用します。",
          "micDisconnected": "マイクが接続されていません。ディクテーションが停止しました。"
@@ -16254,7 +16258,8 @@
          "f082d0cb9d": "処理中…",
          "3de5a129e7": "聞き取り中",
          "4977162383": "音量が大きすぎます",
-          "25f2b7a6a5": "発話中"
+          "25f2b7a6a5": "発話中",
+          "loadingModel": "音声モデルを読み込み中…"
        }
      },
      "dashboard": {
@@ -16244,6 +16244,10 @@
          "55127a3706": "받아쓰기 실패: {{value0}}",
          "bb7f599ee7": "설정 열기",
          "2d5b9fabf9": "마이크 액세스가 거부되었습니다. 시스템 설정에서 액세스 권한을 부여한 다음 Orca를 다시 시작하세요.",
+          "noModel": "선택된 음성 모델이 없습니다. 설정 > 음성에서 다운로드하세요.",
+          "disabled": "음성 받아쓰기가 비활성화되어 있습니다. 설정 > 음성에서 활성화하세요.",
+          "notReady": "음성 모델이 준비되지 않았습니다. 설정 > 음성에서 다운로드하세요.",
+          "unknownModel": "선택한 모델을 더 이상 사용할 수 없습니다. 설정 > 음성에서 다른 모델을 선택하세요.",
          "5d2c3e7ae3": "음성이 감지되지 않았습니다.",
          "micFallback": "선택한 마이크를 사용할 수 없습니다. 시스템 기본값을 사용합니다.",
          "micDisconnected": "마이크 연결이 끊어졌습니다. 받아쓰기가 중지되었습니다."
@@ -16254,7 +16258,8 @@
          "f082d0cb9d": "처리 중…",
          "3de5a129e7": "듣는 중",
          "4977162383": "음량이 너무 큽니다",
-          "25f2b7a6a5": "말하는 중"
+          "25f2b7a6a5": "말하는 중",
+          "loadingModel": "음성 모델 불러오는 중…"
        }
      },
      "dashboard": {
@@ -19169,6 +19169,10 @@
          "55127a3706": "Falha no ditado: {{value0}}",
          "bb7f599ee7": "Abrir Configurações",
          "2d5b9fabf9": "Acesso ao microfone negado. Conceda acesso nas configurações do sistema e reinicie o Orca.",
+          "noModel": "Nenhum modelo de fala selecionado. Baixe um em Configurações > Voz.",
+          "disabled": "O ditado por voz está desativado. Ative-o em Configurações > Voz.",
+          "notReady": "O modelo de fala não está pronto. Baixe-o em Configurações > Voz.",
+          "unknownModel": "O modelo selecionado não está mais disponível. Escolha outro em Configurações > Voz.",
          "5d2c3e7ae3": "Nenhuma fala detectada.",
          "micFallback": "O microfone selecionado não está disponível. Usando o padrão do sistema.",
          "micDisconnected": "Microfone desconectado. O ditado foi interrompido."
@@ -19179,7 +19183,8 @@
          "f082d0cb9d": "Processando…",
          "3de5a129e7": "Ouvindo",
          "4977162383": "Muito alto",
-          "25f2b7a6a5": "Falando"
+          "25f2b7a6a5": "Falando",
+          "loadingModel": "Carregando modelo de fala…"
        }
      },
      "link": {
@@ -17415,6 +17415,10 @@
          "55127a3706": "Ошибка диктовки: {{value0}}",
          "bb7f599ee7": "Открыть настройки",
          "2d5b9fabf9": "Доступ к микрофону запрещён. Разрешите доступ в системных настройках и перезапустите Orca.",
+          "noModel": "Модель распознавания речи не выбрана. Скачайте её в разделе «Настройки > Голос».",
+          "disabled": "Голосовая диктовка отключена. Включите её в разделе «Настройки > Голос».",
+          "notReady": "Модель распознавания речи не готова. Скачайте её в разделе «Настройки > Голос».",
+          "unknownModel": "Выбранная модель больше недоступна. Выберите другую в разделе «Настройки > Голос».",
          "5d2c3e7ae3": "Речь не обнаружена.",
          "micFallback": "Выбранный микрофон недоступен. Используется системный микрофон по умолчанию.",
          "micDisconnected": "Микрофон отключён. Диктовка остановлена."
@@ -17425,7 +17429,8 @@
          "f082d0cb9d": "Обработка…",
          "3de5a129e7": "Прослушивание",
          "4977162383": "Слишком громко",
-          "25f2b7a6a5": "Речь"
+          "25f2b7a6a5": "Речь",
+          "loadingModel": "Загрузка модели речи…"
        }
      },
      "dashboard": {
@@ -16903,6 +16903,10 @@
          "55127a3706": "Dikte tamamlanamadı: {{value0}}",
          "bb7f599ee7": "Ayarları Aç",
          "2d5b9fabf9": "Mikrofon erişimi reddedildi. Sistem ayarlarından erişim verin, ardından Orca'yı yeniden başlatın.",
+          "noModel": "Konuşma modeli seçilmedi. Ayarlar > Ses bölümünden bir model indirin.",
+          "disabled": "Sesli dikte devre dışı. Ayarlar > Ses bölümünden etkinleştirin.",
+          "notReady": "Konuşma modeli hazır değil. Ayarlar > Ses bölümünden indirin.",
+          "unknownModel": "Seçili model artık kullanılamıyor. Ayarlar > Ses bölümünden başka bir model seçin.",
          "5d2c3e7ae3": "Konuşma algılanmadı.",
          "micFallback": "Seçili mikrofon kullanılamıyor. Sistem varsayılanı kullanılıyor.",
          "micDisconnected": "Mikrofon bağlantısı kesildi. Dikte durduruldu."
@@ -16913,7 +16917,8 @@
          "f082d0cb9d": "İşleniyor…",
          "3de5a129e7": "Dinliyor",
          "4977162383": "Çok yüksek sesli",
-          "25f2b7a6a5": "Konuşuyor"
+          "25f2b7a6a5": "Konuşuyor",
+          "loadingModel": "Konuşma modeli yükleniyor…"
        }
      },
      "dashboard": {
@@ -16692,6 +16692,10 @@
          "55127a3706": "聽寫失敗：{{value0}}",
          "bb7f599ee7": "開啟設定",
          "2d5b9fabf9": "麥克風存取被拒絕。在系統設定中授予存取權限，然後重新啟動 Orca。",
+          "noModel": "未選擇語音模型。請在「設定 > 語音」中下載一個。",
+          "disabled": "語音聽寫已停用。請在「設定 > 語音」中啟用。",
+          "notReady": "語音模型尚未就緒。請在「設定 > 語音」中下載。",
+          "unknownModel": "所選模型已無法使用。請在「設定 > 語音」中選擇其他模型。",
          "5d2c3e7ae3": "未偵測到語音。",
          "micFallback": "所選麥克風不可用。使用系統預設值。",
          "micDisconnected": "麥克風已中斷連線。聽寫已停止。"
@@ -16702,7 +16706,8 @@
          "f082d0cb9d": "正在處理…",
          "3de5a129e7": "正在聆聽",
          "4977162383": "聲音太大",
-          "25f2b7a6a5": "正在說話"
+          "25f2b7a6a5": "正在說話",
+          "loadingModel": "正在載入語音模型…"
        }
      },
      "dashboard": {
@@ -16244,6 +16244,10 @@
          "55127a3706": "听写失败：{{value0}}",
          "bb7f599ee7": "打开设置",
          "2d5b9fabf9": "麦克风访问被拒绝。在系统设置中授予访问权限，然后重新启动 Orca。",
+          "noModel": "未选择语音模型。请在“设置 > 语音”中下载一个。",
+          "disabled": "语音听写已禁用。请在“设置 > 语音”中启用。",
+          "notReady": "语音模型尚未就绪。请在“设置 > 语音”中下载。",
+          "unknownModel": "所选模型已不可用。请在“设置 > 语音”中选择其他模型。",
          "5d2c3e7ae3": "未检测到语音。",
          "micFallback": "所选麦克风不可用。使用系统默认值。",
          "micDisconnected": "麦克风已断开。听写停止了。"
@@ -16254,7 +16258,8 @@
          "f082d0cb9d": "正在处理…",
          "3de5a129e7": "正在聆听",
          "4977162383": "音量过大",
-          "25f2b7a6a5": "正在说话"
+          "25f2b7a6a5": "正在说话",
+          "loadingModel": "正在加载语音模型…"
        }
      },
      "dashboard": {