From bc72b3994e2f5b5dcc1eade4ce90814957324c52 Mon Sep 17 00:00:00 2001 From: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com> Date: Mon, 9 Nov 2026 17:50:23 +0000 Subject: [PATCH] feat(settings): redesign Voice settings around a setup card and a rated model list --- src/main/speech/model-catalog.ts | 24 -- .../speech-model-update-target.test.ts | 1 - .../settings-capability-section-renderers.tsx | 4 +- .../VoiceDictationLanguageSetting.fixture.tsx | 1 - .../VoiceDictationLanguageSetting.test.tsx | 25 -- .../voice/VoiceDictationLanguageSetting.tsx | 15 +- .../voice/VoiceDictationSettingsSection.tsx | 101 ------ .../settings/voice/VoiceInputSettings.tsx | 96 ++++++ .../settings/voice/VoiceMicrophoneSetting.tsx | 60 +--- .../settings/voice/VoicePane.test.tsx | 9 +- .../components/settings/voice/VoicePane.tsx | 228 ++++++------ .../settings/voice/VoiceSetupCard.fixture.tsx | 60 ++++ .../settings/voice/VoiceSetupCard.tsx | 177 ++++++++++ .../voice/VoiceSpeechModelQuality.tsx | 103 ++++++ .../settings/voice/VoiceSpeechModelRow.tsx | 191 ++++++++++ .../voice/VoiceSpeechModelSection.fixture.tsx | 128 +++++-- .../voice/VoiceSpeechModelSection.test.tsx | 41 +-- .../voice/VoiceSpeechModelSection.tsx | 325 ++++++------------ .../voice/speech-model-description.ts | 69 ---- .../settings/voice/speech-model-download.ts | 26 ++ .../voice/use-voice-microphone-devices.ts | 67 ++++ .../voice/voice-speech-model-rows.test.ts | 42 +++ .../settings/voice/voice-speech-model-rows.ts | 47 +++ src/renderer/src/i18n/locales/de.json | 81 +++-- src/renderer/src/i18n/locales/en.json | 82 +++-- src/renderer/src/i18n/locales/es.json | 81 +++-- src/renderer/src/i18n/locales/fr.json | 81 +++-- src/renderer/src/i18n/locales/ja.json | 81 +++-- src/renderer/src/i18n/locales/ko.json | 81 +++-- src/renderer/src/i18n/locales/pt.json | 81 +++-- src/renderer/src/i18n/locales/ru.json | 81 +++-- src/renderer/src/i18n/locales/tr.json | 81 +++-- src/renderer/src/i18n/locales/zh-Hant.json | 81 +++-- src/renderer/src/i18n/locales/zh.json | 81 +++-- src/shared/speech-types.ts | 1 - 35 files changed, 1833 insertions(+), 900 deletions(-) delete mode 100644 src/renderer/src/components/settings/voice/VoiceDictationSettingsSection.tsx create mode 100644 src/renderer/src/components/settings/voice/VoiceInputSettings.tsx create mode 100644 src/renderer/src/components/settings/voice/VoiceSetupCard.fixture.tsx create mode 100644 src/renderer/src/components/settings/voice/VoiceSetupCard.tsx create mode 100644 src/renderer/src/components/settings/voice/VoiceSpeechModelQuality.tsx create mode 100644 src/renderer/src/components/settings/voice/VoiceSpeechModelRow.tsx delete mode 100644 src/renderer/src/components/settings/voice/speech-model-description.ts create mode 100644 src/renderer/src/components/settings/voice/speech-model-download.ts create mode 100644 src/renderer/src/components/settings/voice/use-voice-microphone-devices.ts create mode 100644 src/renderer/src/components/settings/voice/voice-speech-model-rows.test.ts create mode 100644 src/renderer/src/components/settings/voice/voice-speech-model-rows.ts diff --git a/src/main/speech/model-catalog.ts b/src/main/speech/model-catalog.ts index 97ab776a901..258a129f940 100644 --- a/src/main/speech/model-catalog.ts +++ b/src/main/speech/model-catalog.ts @@ -18,8 +18,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'parakeet-tdt-0.6b-v3-int8', label: 'Parakeet TDT v3', - description: - 'Highest accuracy for 25 European languages. Punctuation, capitalization, and word-level timestamps.', type: 'transducer', provider: 'local', language: 'multilingual', @@ -35,8 +33,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'parakeet-tdt-0.6b-v2-int8', label: 'Parakeet TDT v2', - description: - 'English only. Faster than v3 with similar accuracy. Punctuation and capitalization.', type: 'transducer', provider: 'local', language: 'en', @@ -51,7 +47,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'zipformer-bilingual-zh-en', label: 'Zipformer Bilingual', - description: 'Chinese + English with code-switching. Low-latency real-time streaming.', type: 'transducer', provider: 'local', language: 'zh-en', @@ -65,8 +60,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'paraformer-bilingual-zh-en', label: 'Paraformer Bilingual', - description: - 'Chinese (Mandarin + dialects) + English. Strong on accented and regional Chinese.', type: 'paraformer', provider: 'local', language: 'zh-en', @@ -80,7 +73,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'zipformer-streaming-en-20m', label: 'Zipformer Streaming EN', - description: 'English only. Lightweight 20M-param model, good balance of speed and size.', type: 'transducer', provider: 'local', language: 'en', @@ -94,7 +86,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'zipformer-streaming-zh-14m', label: 'Zipformer Streaming ZH', - description: 'Chinese only. Ultra-lightweight 14M-param model, ideal for low-resource devices.', type: 'transducer', provider: 'local', language: 'zh', @@ -108,7 +99,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'zipformer-streaming-korean', label: 'Zipformer Streaming KO', - description: 'Korean only. Low-latency real-time streaming.', type: 'transducer', provider: 'local', language: 'ko', @@ -122,7 +112,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'parakeet-tdt-ctc-0.6b-ja-int8', label: 'Parakeet TDT-CTC JA', - description: 'Japanese only. Trained on 35k+ hours of natural speech. Punctuation included.', type: 'nemo-ctc', provider: 'local', language: 'ja', @@ -136,7 +125,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'whisper-tiny', label: 'Whisper Tiny', - description: '90+ languages. Lower accuracy than Parakeet but broadest language coverage.', type: 'whisper', provider: 'local', language: 'multilingual', @@ -152,8 +140,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'whisper-large-v3-turbo', label: 'Whisper Large v3 Turbo', - description: - 'Best accuracy for Japanese, Korean, and Chinese, and 90+ languages. Slower than Parakeet.', type: 'whisper', provider: 'local', language: 'multilingual', @@ -168,8 +154,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'sense-voice-zh-en-ja-ko-yue', label: 'SenseVoice', - description: - 'Chinese, English, Japanese, Korean, and Cantonese with automatic language detection.', type: 'senseVoice', provider: 'local', language: 'multilingual', @@ -185,8 +169,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'nemotron-3.5-asr-streaming-0.6b', label: 'Nemotron 3.5', - description: - 'English, Spanish, German and many more languages, with automatic language detection.', type: 'transducer', provider: 'local', language: 'multilingual', @@ -202,7 +184,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'openai-gpt-transcribe', label: 'GPT Transcribe', - description: 'Cloud transcription with high accuracy. Requires an OpenAI API key.', type: 'openai', provider: 'openai', language: 'multilingual', @@ -214,8 +195,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'openai-compatible', label: 'OpenAI-compatible', - description: - 'Any service with an OpenAI-style transcription API, such as Groq, Mistral or a local server.', type: 'openai', provider: 'openai-compatible', language: 'multilingual', @@ -227,8 +206,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'openai-gpt-4o-mini-transcribe', label: 'GPT-4o mini Transcribe', - description: - 'Cloud transcription with strong accuracy and low cost. Requires an OpenAI API key.', type: 'openai', provider: 'openai', language: 'multilingual', @@ -242,7 +219,6 @@ export const SPEECH_MODEL_CATALOG: SpeechModelManifest[] = [ { id: 'openai-gpt-4o-transcribe', label: 'GPT-4o Transcribe', - description: 'Cloud transcription with higher accuracy. Requires an OpenAI API key.', type: 'openai', provider: 'openai', language: 'multilingual', diff --git a/src/renderer/src/components/dictation/speech-model-update-target.test.ts b/src/renderer/src/components/dictation/speech-model-update-target.test.ts index c59e4e8bdd6..538bf45e94d 100644 --- a/src/renderer/src/components/dictation/speech-model-update-target.test.ts +++ b/src/renderer/src/components/dictation/speech-model-update-target.test.ts @@ -6,7 +6,6 @@ function model(id: string, extra: Partial = {}): SpeechMode return { id, label: id, - description: '', type: 'transducer', provider: 'local', language: 'en', diff --git a/src/renderer/src/components/settings/page/settings-capability-section-renderers.tsx b/src/renderer/src/components/settings/page/settings-capability-section-renderers.tsx index e3ec3097cca..5cddf1f09a0 100644 --- a/src/renderer/src/components/settings/page/settings-capability-section-renderers.tsx +++ b/src/renderer/src/components/settings/page/settings-capability-section-renderers.tsx @@ -177,8 +177,8 @@ export function renderDesktopCapabilitySettingsSections( id="voice" title={t('auto.components.settings.Settings.5063bb47a5', 'Voice')} description={t( - 'auto.components.settings.Settings.eb1176a14e', - 'Local speech-to-text dictation with on-device models.' + 'components.settings.voice.sectionDescription', + 'Talk instead of typing in any pane. Speech is recognized on this computer unless you pick a cloud model.' )} searchEntries={navigation.getSectionSearchEntries('voice')} > diff --git a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.fixture.tsx b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.fixture.tsx index 7e048404a80..9fde64dff1f 100644 --- a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.fixture.tsx +++ b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.fixture.tsx @@ -15,7 +15,6 @@ function model( return { id: label, label, - description: '', type, provider: 'local', language: 'multilingual', diff --git a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.test.tsx b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.test.tsx index 94d39441b7f..fea193fe3bf 100644 --- a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.test.tsx +++ b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.test.tsx @@ -35,7 +35,6 @@ it('keeps a selected language visible when the chosen model does not support it' const selectedModel: SpeechModelManifest = { id: 'limited', label: 'Limited model', - description: '', type: 'whisper', provider: 'local', language: 'multilingual', @@ -59,27 +58,3 @@ it('keeps a selected language visible when the chosen model does not support it' fireEvent.click(screen.getByRole('combobox', { name: 'Dictation Language' })) expect(screen.getByRole('option', { name: 'Korean' })).toBeTruthy() }) - -it('keeps the language control disabled for models that ignore language settings', () => { - const selectedModel: SpeechModelManifest = { - id: 'parakeet', - label: 'Parakeet', - description: '', - type: 'transducer', - provider: 'local', - language: 'multilingual', - languages: ['en', 'es'], - sampleRate: 16000, - streaming: false - } - render( - - ) - - const trigger = screen.getByRole('combobox', { name: 'Dictation Language' }) - expect(trigger.getAttribute('disabled')).not.toBeNull() -}) diff --git a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.tsx b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.tsx index 22553593c17..8d0072f6688 100644 --- a/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.tsx +++ b/src/renderer/src/components/settings/voice/VoiceDictationLanguageSetting.tsx @@ -55,12 +55,14 @@ export function VoiceDictationLanguageSetting({ const intlLocale = useIntlLocale() const modelIgnoresLanguage = selectedModel !== undefined && !selectedModel.dictationLanguages const language = voiceSettings.dictationLanguage || DICTATION_LANGUAGE_AUTO - const choices = selectedModel?.dictationLanguages ?? DICTATION_LANGUAGES - // Keep a choice the model lacks selectable rather than showing a blank trigger. - const codes = - language === DICTATION_LANGUAGE_AUTO || choices.includes(language) - ? choices - : [language, ...choices] + // Every language, even ones this model lacks: the choice also rates the other models. + const codes = [ + ...new Set([ + ...(language === DICTATION_LANGUAGE_AUTO ? [] : [language]), + ...DICTATION_LANGUAGES, + ...(selectedModel?.dictationLanguages ?? []) + ]) + ] const collator = uiCollator(NAME_COLLATOR, intlLocale) const options = codes .map((code) => ({ code, name: languageDisplayName(code, intlLocale) })) @@ -99,7 +101,6 @@ export function VoiceDictationLanguageSetting({ 'auto.components.settings.SettingsSearchableSelect.searchLanguages', 'Search languages…' )} - disabled={!voiceSettings.enabled || modelIgnoresLanguage} onValueChange={(value) => onUpdateVoiceSettings({ dictationLanguage: value })} /> } diff --git a/src/renderer/src/components/settings/voice/VoiceDictationSettingsSection.tsx b/src/renderer/src/components/settings/voice/VoiceDictationSettingsSection.tsx deleted file mode 100644 index 9edcdec2532..00000000000 --- a/src/renderer/src/components/settings/voice/VoiceDictationSettingsSection.tsx +++ /dev/null @@ -1,101 +0,0 @@ -import type { VoiceSettings } from '../../../../../shared/speech-types' -import { - dictationShortcutMode, - dictationShortcutModeSettings -} from '../../../../../shared/dictation-shortcut-mode' -import { Label } from '../../ui/label' -import { Separator } from '../../ui/separator' -import { Switch } from '../../ui/switch' -import { useShortcutLabel } from '@/hooks/useShortcutLabel' -import { useTranslate } from '@/i18n/use-translate' -import { VoiceMicrophoneSetting } from './VoiceMicrophoneSetting' - -type VoiceDictationSettingsSectionProps = { - voiceSettings: VoiceSettings - permissionPending: boolean - onToggleVoiceDictation: () => void - onUpdateVoiceSettings: (updates: Partial) => void -} - -export function VoiceDictationSettingsSection({ - voiceSettings, - permissionPending, - onToggleVoiceDictation, - onUpdateVoiceSettings -}: VoiceDictationSettingsSectionProps): React.JSX.Element { - const t = useTranslate() - const shortcutLabel = useShortcutLabel('voice.dictation') - const selectedMode = dictationShortcutMode(voiceSettings) - const modeLabels = { - auto: t('auto.components.settings.VoicePane.dictationModeAuto', 'Auto'), - hold: t('auto.components.settings.VoicePane.174da92062', 'Hold'), - toggle: t('auto.components.settings.VoicePane.118b3c2dee', 'Toggle') - } - - return ( - <> -
-
- -

- {t( - 'auto.components.settings.VoicePane.dictationShortcutDescription', - 'Press {{shortcut}} to dictate text into any focused pane.', - { shortcut: shortcutLabel } - )} -

-
- void onToggleVoiceDictation()} - className="disabled:cursor-wait disabled:opacity-70" - /> -
- - - -
-
- -

- {t( - 'auto.components.settings.VoicePane.dictationModesWithAutoDescription', - 'Auto: hold {{shortcut}} to dictate until you let go, or tap it to start and stop. Hold: dictate while {{shortcut}} is held. Toggle: press {{shortcut}} once to start, again to stop.', - { shortcut: shortcutLabel } - )} -

-
-
- {(['auto', 'hold', 'toggle'] as const).map((mode) => ( - - ))} -
-
- - - - - - - - ) -} diff --git a/src/renderer/src/components/settings/voice/VoiceInputSettings.tsx b/src/renderer/src/components/settings/voice/VoiceInputSettings.tsx new file mode 100644 index 00000000000..11645289e91 --- /dev/null +++ b/src/renderer/src/components/settings/voice/VoiceInputSettings.tsx @@ -0,0 +1,96 @@ +import type { VoiceSettings } from '../../../../../shared/speech-types' +import { + dictationShortcutMode, + dictationShortcutModeSettings, + type DictationShortcutMode +} from '../../../../../shared/dictation-shortcut-mode' +import { useAppStore } from '@/store' +import { ShortcutKeyCombo } from '@/components/ShortcutKeyCombo' +import { useShortcutKeyDetails, useShortcutLabel } from '@/hooks/useShortcutLabel' +import { useTranslate } from '@/i18n/use-translate' +import { Button } from '../../ui/button' +import { + SettingsRow, + SettingsSegmentedControl, + SettingsSubsectionHeader +} from '../form-controls/SettingsFormControls' +import { VoiceMicrophoneSetting } from './VoiceMicrophoneSetting' + +export const VOICE_MICROPHONE_ROW_ID = 'voice-microphone' + +type VoiceInputSettingsProps = { + voiceSettings: VoiceSettings + onUpdateVoiceSettings: (updates: Partial) => void +} + +export function VoiceInputSettings({ + voiceSettings, + onUpdateVoiceSettings +}: VoiceInputSettingsProps): React.JSX.Element { + const t = useTranslate() + const openSettingsTarget = useAppStore((s) => s.openSettingsTarget) + const shortcutLabel = useShortcutLabel('voice.dictation') + const shortcutKeys = useShortcutKeyDetails('voice.dictation') + + return ( +
+ +
+ + + +
+ } + /> + + value={dictationShortcutMode(voiceSettings)} + onChange={(mode) => onUpdateVoiceSettings(dictationShortcutModeSettings(mode))} + ariaLabel={t('auto.components.settings.VoicePane.ba4a900d1d', 'Dictation Mode')} + options={[ + { + value: 'auto', + label: t('auto.components.settings.VoicePane.dictationModeAuto', 'Auto') + }, + { + value: 'hold', + label: t('auto.components.settings.VoicePane.174da92062', 'Hold') + }, + { + value: 'toggle', + label: t('auto.components.settings.VoicePane.118b3c2dee', 'Toggle') + } + ]} + /> + } + /> +
+ +
+
+ + ) +} diff --git a/src/renderer/src/components/settings/voice/VoiceMicrophoneSetting.tsx b/src/renderer/src/components/settings/voice/VoiceMicrophoneSetting.tsx index 38fc213baf7..4a7fb26886c 100644 --- a/src/renderer/src/components/settings/voice/VoiceMicrophoneSetting.tsx +++ b/src/renderer/src/components/settings/voice/VoiceMicrophoneSetting.tsx @@ -1,4 +1,4 @@ -import { useCallback, useEffect, useMemo, useRef, useState } from 'react' +import { useCallback, useMemo, useState } from 'react' import { toast } from 'sonner' import type { VoiceSettings } from '../../../../../shared/speech-types' import { Button } from '../../ui/button' @@ -6,15 +6,14 @@ import { Label } from '../../ui/label' import { Select, SelectContent, SelectItem, SelectTrigger, SelectValue } from '../../ui/select' import { buildVoiceMicrophoneSelectOptions, - listVoiceMicrophoneDevices, - microphoneDeviceIdFromSelectValue, - type VoiceMicrophoneDevice + microphoneDeviceIdFromSelectValue } from '@/components/dictation/microphone-devices' import { useMountedRef } from '@/hooks/useMountedRef' import { translate } from '@/i18n/i18n' import { useTranslate } from '@/i18n/use-translate' import { useUiLanguageRevision } from '@/i18n/use-ui-language-revision' import { extractIpcErrorMessage } from '@/lib/ipc-error' +import { useVoiceMicrophoneDevices } from './use-voice-microphone-devices' type VoiceMicrophoneSettingProps = { voiceSettings: VoiceSettings @@ -62,66 +61,17 @@ function microphoneAccessErrorMessage(error: unknown): string { ) } -function sameDeviceList( - a: readonly VoiceMicrophoneDevice[], - b: readonly VoiceMicrophoneDevice[] -): boolean { - return ( - a.length === b.length && - a.every((device, index) => { - const other = b[index] - return device.deviceId === other?.deviceId && device.label === other.label - }) - ) -} - export function VoiceMicrophoneSetting({ voiceSettings, onUpdateVoiceSettings }: VoiceMicrophoneSettingProps): React.JSX.Element { const t = useTranslate() const languageRevision = useUiLanguageRevision() - const [devices, setDevices] = useState([]) - const [devicesKnown, setDevicesKnown] = useState(false) + // Rescan when dictation enables because permission and real device labels may arrive with that toggle. + const { devices, devicesKnown, refreshDevices } = useVoiceMicrophoneDevices(voiceSettings.enabled) const [accessPending, setAccessPending] = useState(false) const [accessError, setAccessError] = useState(null) const mountedRef = useMountedRef() - // Bluetooth emits repeated devicechange events; reject stale enumeration results. - const refreshGenerationRef = useRef(0) - - const refreshDevices = useCallback(async (): Promise => { - const generation = refreshGenerationRef.current + 1 - refreshGenerationRef.current = generation - if (typeof navigator === 'undefined' || !navigator.mediaDevices?.enumerateDevices) { - return - } - let next: VoiceMicrophoneDevice[] = [] - try { - next = listVoiceMicrophoneDevices(await navigator.mediaDevices.enumerateDevices()) - } catch { - next = [] - } - if (!mountedRef.current || refreshGenerationRef.current !== generation) { - return - } - setDevicesKnown(next.length > 0) - setDevices((current) => (sameDeviceList(current, next) ? current : next)) - }, [mountedRef]) - - // Rescan when dictation enables because permission and real device labels may arrive with that toggle. - useEffect(() => { - void refreshDevices() - if (typeof navigator === 'undefined' || !navigator.mediaDevices?.addEventListener) { - return - } - const handleDeviceChange = (): void => { - void refreshDevices() - } - navigator.mediaDevices.addEventListener('devicechange', handleDeviceChange) - return () => { - navigator.mediaDevices.removeEventListener('devicechange', handleDeviceChange) - } - }, [refreshDevices, voiceSettings.enabled]) // A generic stream grants discovery even when the saved device is stale. const openStreamAndRefreshDevices = useCallback(async (): Promise => { diff --git a/src/renderer/src/components/settings/voice/VoicePane.test.tsx b/src/renderer/src/components/settings/voice/VoicePane.test.tsx index 19e1fc6f5f7..0694de92ee3 100644 --- a/src/renderer/src/components/settings/voice/VoicePane.test.tsx +++ b/src/renderer/src/components/settings/voice/VoicePane.test.tsx @@ -19,7 +19,8 @@ const { useAppStoreMock, useShortcutLabelMock } = vi.hoisted(() => ({ vi.mock('@/store/index', () => ({ useAppStore: useAppStoreMock })) vi.mock('@/hooks/useShortcutLabel', () => ({ - useShortcutLabel: useShortcutLabelMock + useShortcutLabel: useShortcutLabelMock, + useShortcutKeyDetails: () => ({ keys: [], doubleTap: false }) })) vi.mock('sonner', () => ({ @@ -108,6 +109,12 @@ describe('VoicePane', () => { root.render() }) + const cloudTab = Array.from( + container.querySelectorAll('[role="radio"]') + ).find((button) => button.textContent === 'Cloud') + await act(async () => { + cloudTab?.dispatchEvent(new MouseEvent('click', { bubbles: true })) + }) const disconnect = container.querySelector( 'button[aria-label="Disconnect OpenAI API key"]' ) diff --git a/src/renderer/src/components/settings/voice/VoicePane.tsx b/src/renderer/src/components/settings/voice/VoicePane.tsx index 9b008cdd441..b29ea9ff4a7 100644 --- a/src/renderer/src/components/settings/voice/VoicePane.tsx +++ b/src/renderer/src/components/settings/voice/VoicePane.tsx @@ -12,15 +12,18 @@ import { OpenAiCompatibleProviderDialog } from './OpenAiCompatibleProviderDialog import { OpenAiTranscriptionSettingsRow } from './OpenAiTranscriptionSettingsRow' import { handleVoiceDictationToggle } from './voice-dictation-toggle' import { VoiceDictationLanguageSetting } from './VoiceDictationLanguageSetting' -import { VoiceDictationSettingsSection } from './VoiceDictationSettingsSection' +import { VOICE_MICROPHONE_ROW_ID, VoiceInputSettings } from './VoiceInputSettings' +import { VoiceSetupCard } from './VoiceSetupCard' +import { cancelSpeechModelDownload, startSpeechModelDownload } from './speech-model-download' import { VoiceModelUnloadSetting } from './VoiceModelUnloadSetting' import { VoiceSpeechModelSection } from './VoiceSpeechModelSection' -import { SettingsSwitchRow } from '../form-controls/SettingsFormControls' +import { SettingsSubsectionHeader, SettingsSwitchRow } from '../form-controls/SettingsFormControls' import { VoiceSpeechGpuSetting } from './VoiceSpeechGpuSetting' -import { matchesSettingsSearch } from '../search/settings-search' -import { getOpenaiTranscriptionSearchEntry } from './voice-pane-search' import { useTranslate } from '@/i18n/use-translate' import { readIpcErrorMessage } from '@/lib/ipc-error' +import { useLanguageReadiness } from '@/i18n/language-readiness' +import { motionSafeScrollBehavior } from '@/hooks/usePrefersReducedMotion' +import { speechModelRatingLanguage } from '../../../../../shared/speech-model-quality' type VoicePaneProps = { settings: GlobalSettings @@ -35,8 +38,10 @@ export function VoicePane({ settings, updateSettings }: VoicePaneProps): React.J const modelStates = useAppStore((s) => s.modelStates) const refreshModelStates = useAppStore((s) => s.refreshModelStates) const markFeatureTipsSeen = useAppStore((s) => s.markFeatureTipsSeen) - const settingsSearchQuery = useAppStore((s) => s.settingsSearchQuery ?? '') - const [catalog, setCatalog] = useState([]) + const systemLanguageTag = useLanguageReadiness((s) => s.detection?.detectedTag) + // null until main answers, so the card and list never show a state the catalog would contradict. + const [loadedCatalog, setCatalog] = useState(null) + const catalog = loadedCatalog ?? [] const [permissionPending, setPermissionPending] = useState(false) const [openAiDialogOpen, setOpenAiDialogOpen] = useState(false) const [openAiApiKeyDraft, setOpenAiApiKeyDraft] = useState('') @@ -143,18 +148,11 @@ export function VoicePane({ settings, updateSettings }: VoicePaneProps): React.J } const selectedModel = catalog.find((m) => m.id === voiceSettings.sttModel) - const showOpenAiSettingsRow = - voiceSettings.openAiApiKeyConfigured || - selectedModel?.provider === 'openai' || - (settingsSearchQuery.trim() !== '' && - matchesSettingsSearch(settingsSearchQuery, getOpenaiTranscriptionSearchEntry())) const compatibleModel = catalog.find((m) => m.provider === 'openai-compatible') const compatibleConfigured = compatibleModel !== undefined && modelStates.some((s) => s.id === compatibleModel.id && s.status === 'ready') - const showCompatibleSettingsRow = - compatibleConfigured || selectedModel?.provider === 'openai-compatible' const openOpenAiDialog = (modelId: string | null = null): void => { setPendingCloudModelId(modelId) @@ -167,11 +165,11 @@ export function VoicePane({ settings, updateSettings }: VoicePaneProps): React.J setCompatibleDialogOpen(true) } - const openCloudSetup = (manifest: SpeechModelManifest): void => { - if (manifest.provider === 'openai-compatible') { - openCompatibleDialog(manifest.id) + const openCloudSetup = (modelId: string): void => { + if (catalog.find((m) => m.id === modelId)?.provider === 'openai-compatible') { + openCompatibleDialog(modelId) } else { - openOpenAiDialog(manifest.id) + openOpenAiDialog(modelId) } } @@ -252,95 +250,127 @@ export function VoicePane({ settings, updateSettings }: VoicePaneProps): React.J } } + const selectedState = modelStates.find((s) => s.id === voiceSettings.sttModel) + const ratingLanguage = + speechModelRatingLanguage(voiceSettings.dictationLanguage, systemLanguageTag ?? undefined) ?? + 'en' + return ( -
- void toggleVoiceDictation()} - onUpdateVoiceSettings={updateVoiceSettings} - /> - - - - - - - - - - - updateVoiceSettings({ removeFillerWords: !voiceSettings.removeFillerWords }) - } - /> - - - - - - {selectedModel?.provider === 'local' && ( +
+ {loadedCatalog ? ( <> - + void toggleVoiceDictation()} + onDownloadModel={(modelId) => startSpeechModelDownload(modelId, refreshModelStates)} + onCancelDownload={(modelId) => cancelSpeechModelDownload(modelId, refreshModelStates)} + onShowMicrophone={() => + document + .getElementById(VOICE_MICROPHONE_ROW_ID) + ?.scrollIntoView({ behavior: motionSafeScrollBehavior(), block: 'center' }) + } + /> + + + {voiceSettings.openAiApiKeyConfigured ? ( + <> + + openOpenAiDialog(null)} + onClear={() => void clearOpenAiApiKey()} + /> + + ) : null} + {compatibleConfigured ? ( + openCompatibleDialog(null)} + onClear={() => void removeCompatibleProvider()} + /> + ) : null} + + } + /> + + ) : null} + + + +
+ + +
+ + + +
+ + + updateVoiceSettings({ removeFillerWords: !voiceSettings.removeFillerWords }) + } + /> +
+ + + + + + + +
+ +
- - )} - - {showOpenAiSettingsRow && ( - <> - - - openOpenAiDialog(null)} - onClear={() => void clearOpenAiApiKey()} - /> - - )} - - {showCompatibleSettingsRow && ( - <> - - openCompatibleDialog(null)} - onClear={() => void removeCompatibleProvider()} - /> - - )} +
+
{}} + onDownloadModel={() => {}} + onCancelDownload={() => {}} + onShowMicrophone={() => {}} + /> + ) +} + +export const newUser = defineFixture({ + width: 760, + render: () => render(false, undefined, undefined) +}) + +export const on = defineFixture({ + width: 760, + render: () => render(true, parakeet, { id: 'v3', status: 'ready' }) +}) + +export const off = defineFixture({ + width: 760, + render: () => render(false, parakeet, { id: 'v3', status: 'ready' }) +}) + +export const needsUpdate = defineFixture({ + width: 760, + render: () => render(true, parakeet, { id: 'v3', status: 'not-downloaded', needsUpdate: true }) +}) + +export const downloading = defineFixture({ + width: 760, + render: () => render(true, parakeet, { id: 'v3', status: 'downloading', progress: 0.42 }) +}) diff --git a/src/renderer/src/components/settings/voice/VoiceSetupCard.tsx b/src/renderer/src/components/settings/voice/VoiceSetupCard.tsx new file mode 100644 index 00000000000..32a11b6e2a1 --- /dev/null +++ b/src/renderer/src/components/settings/voice/VoiceSetupCard.tsx @@ -0,0 +1,177 @@ +import { CircleAlert, Check, Download, Mic } from 'lucide-react' +import type { + SpeechModelManifest, + SpeechModelState, + VoiceSettings +} from '../../../../../shared/speech-types' +import { resolveMicrophoneDevice } from '@/components/dictation/microphone-devices' +import { useShortcutLabel } from '@/hooks/useShortcutLabel' +import { useTranslate } from '@/i18n/use-translate' +import { cn } from '@/lib/utils' +import { Button } from '../../ui/button' +import { Progress } from '../../ui/progress' +import { Switch } from '../../ui/switch' +import { useVoiceMicrophoneDevices } from './use-voice-microphone-devices' + +type VoiceSetupCardProps = { + voiceSettings: VoiceSettings + selectedModel: SpeechModelManifest | undefined + selectedState: SpeechModelState | undefined + permissionPending: boolean + onToggleVoiceDictation: () => void + onDownloadModel: (modelId: string) => void + onCancelDownload: (modelId: string) => void + onShowMicrophone: () => void +} + +/** The pane's first card: dictation's on/off switch and the one problem to fix, if any. */ +export function VoiceSetupCard({ + voiceSettings, + selectedModel, + selectedState, + permissionPending, + onToggleVoiceDictation, + onDownloadModel, + onCancelDownload, + onShowMicrophone +}: VoiceSetupCardProps): React.JSX.Element { + const t = useTranslate() + const shortcut = useShortcutLabel('voice.dictation') + const { devices, devicesKnown } = useVoiceMicrophoneDevices(voiceSettings.enabled) + const microphone = resolveMicrophoneDevice({ + devices: devicesKnown ? devices : null, + preferredDeviceId: voiceSettings.microphoneDeviceId, + preferredDeviceLabel: voiceSettings.microphoneDeviceLabel + }) + const downloading = + selectedState?.status === 'downloading' || selectedState?.status === 'extracting' + const ready = selectedState?.status === 'ready' + const totalMb = Math.round((selectedModel?.sizeBytes ?? 0) / 1_000_000) + const names = { model: selectedModel?.label ?? '', size: totalMb, shortcut } + + let tone: 'neutral' | 'ok' | 'warn' = 'neutral' + let icon = + let title: string + let description: string + let body: React.ReactNode = null + + if (selectedModel && downloading) { + icon = + const percent = Math.round((selectedState?.progress ?? 0) * 100) + title = t('speechModelUpdate.downloadingTitle', 'Downloading {{model}}', names) + description = t( + 'components.settings.voice.setup.downloadingDescription', + 'You can leave this page. Dictation works once the download finishes.' + ) + body = ( +
+ + + {t('speechModelUpdate.progress', '{{done}} of {{total}} MB', { + done: Math.round((selectedState?.progress ?? 0) * totalMb), + total: totalMb + })} + + +
+ ) + } else if (selectedModel && selectedState?.needsUpdate) { + icon = + title = t('speechModelUpdate.updateTitle', 'Update your speech model') + description = t( + 'speechModelUpdate.updateDescription', + 'Dictation now runs on a faster engine. {{model}} needs a one-time download ({{size}} MB) before you can dictate again.', + names + ) + body = ( +
+ +
+ ) + } else if (!selectedModel || !ready) { + title = t('components.settings.voice.setup.title', 'Set up dictation') + description = t( + 'components.settings.voice.setup.pickModel', + 'Talk instead of typing in any pane. Download a speech model below to start; models on this computer keep your audio here.' + ) + } else if (voiceSettings.enabled && microphone.kind === 'missing') { + tone = 'warn' + icon = + title = t( + 'components.settings.voice.setup.microphoneMissingTitle', + '{{microphone}} isn’t connected', + { + microphone: + voiceSettings.microphoneDeviceLabel || + t('components.settings.voice.setup.savedMicrophone', 'Your microphone') + } + ) + description = t( + 'components.settings.voice.setup.microphoneMissingDescription', + 'Dictation uses the system default microphone until it’s back.' + ) + body = ( +
+ +
+ ) + } else if (voiceSettings.enabled) { + tone = 'ok' + icon = + title = t('components.settings.voice.setup.onTitle', 'Dictation is on') + description = + selectedModel.provider === 'local' + ? t( + 'components.settings.voice.setup.onLocal', + 'Press {{shortcut}} in any pane. {{model}} runs on this computer.', + names + ) + : t( + 'components.settings.voice.setup.onCloud', + 'Press {{shortcut}} in any pane. {{model}} runs in the cloud.', + names + ) + } else { + title = t('components.settings.voice.setup.offTitle', 'Dictation is off') + description = t( + 'components.settings.voice.setup.offDescription', + 'Turn it on to dictate with {{shortcut}} in any pane.', + names + ) + } + + return ( +
+
+ {icon} +
+
+
+
{title}
+

{description}

+
+ {body} +
+ +
+ ) +} diff --git a/src/renderer/src/components/settings/voice/VoiceSpeechModelQuality.tsx b/src/renderer/src/components/settings/voice/VoiceSpeechModelQuality.tsx new file mode 100644 index 00000000000..48ea0d150eb --- /dev/null +++ b/src/renderer/src/components/settings/voice/VoiceSpeechModelQuality.tsx @@ -0,0 +1,103 @@ +import type { SpeechModelQuality } from '../../../../../shared/speech-model-quality' +import { useIntlLocale, useTranslate } from '@/i18n/use-translate' +import { uiDateTimeFormat, uiNumberFormat } from '@/i18n/ui-intl-formatters' +import { cn } from '@/lib/utils' +import { Tooltip, TooltipContent, TooltipTrigger } from '../../ui/tooltip' + +// Intl formatter caches key on option identity, so these must be module constants. +const PERCENT_FORMAT: Intl.NumberFormatOptions = { maximumFractionDigits: 1 } +const DATE_FORMAT: Intl.DateTimeFormatOptions = { dateStyle: 'medium', timeZone: 'UTC' } +const SEGMENTS = [1, 2, 3, 4, 5] as const + +export function VoiceSpeechModelQuality({ + quality +}: { + quality: SpeechModelQuality +}): React.JSX.Element { + const t = useTranslate() + const intlLocale = useIntlLocale() + if (quality.kind === 'unsupported') { + return ( + + {t('components.settings.voice.quality.unsupported', 'Doesn’t support your language')} + + ) + } + if (quality.kind === 'unmeasured') { + return ( + + {t('components.settings.voice.quality.unmeasured', 'Not measured')} + + ) + } + const { tier, errorRate } = quality + const word = [ + t('components.settings.voice.quality.poor', 'Poor'), + t('components.settings.voice.quality.fair', 'Fair'), + t('components.settings.voice.quality.good', 'Good'), + t('components.settings.voice.quality.veryGood', 'Very good'), + t('components.settings.voice.quality.excellent', 'Excellent') + ][tier - 1] + const percent = uiNumberFormat(PERCENT_FORMAT, intlLocale).format(errorRate.percent) + const values = { + percent, + clips: errorRate.clips, + date: uiDateTimeFormat(DATE_FORMAT, intlLocale).format(new Date(errorRate.measuredOn)) + } + const rate = + errorRate.metric === 'cer' + ? t( + 'components.settings.voice.quality.characterErrorRate', + '{{percent}}% character error rate', + values + ) + : t('components.settings.voice.quality.wordErrorRate', '{{percent}}% word error rate', values) + const test = + errorRate.test === 'librispeech' + ? t( + 'components.settings.voice.quality.librispeech', + 'Test: LibriSpeech read speech, {{clips}} clips, measured {{date}}.', + values + ) + : t( + 'components.settings.voice.quality.fleurs', + 'Test: FLEURS read speech, {{clips}} clips, measured {{date}}.', + values + ) + const bench = t( + 'components.settings.voice.quality.bench', + 'Orca’s benchmark on a Mac with Apple silicon.' + ) + + return ( + + + + + + +

{rate}

+

+ {test} {bench} +

+
+
+ ) +} diff --git a/src/renderer/src/components/settings/voice/VoiceSpeechModelRow.tsx b/src/renderer/src/components/settings/voice/VoiceSpeechModelRow.tsx new file mode 100644 index 00000000000..f27575438d7 --- /dev/null +++ b/src/renderer/src/components/settings/voice/VoiceSpeechModelRow.tsx @@ -0,0 +1,191 @@ +import { Check, Download, Loader2, Trash2, X } from 'lucide-react' +import { useTranslate } from '@/i18n/use-translate' +import { cn } from '@/lib/utils' +import { Button } from '../../ui/button' +import { Progress } from '../../ui/progress' +import type { VoiceSpeechModelRowData } from './voice-speech-model-rows' +import { VoiceSpeechModelQuality } from './VoiceSpeechModelQuality' + +export type VoiceSpeechModelRowActions = { + onUse: (modelId: string) => void + onDownload: (modelId: string) => void + onCancelDownload: (modelId: string) => void + onDelete: (modelId: string) => void + onSetUpCloud: (modelId: string) => void +} + +type VoiceSpeechModelRowProps = VoiceSpeechModelRowActions & { + row: VoiceSpeechModelRowData + active: boolean + deletePending: boolean +} + +export function VoiceSpeechModelRow({ + row: { manifest, state, quality }, + active, + deletePending, + onUse, + onDownload, + onCancelDownload, + onDelete, + onSetUpCloud +}: VoiceSpeechModelRowProps): React.JSX.Element { + const t = useTranslate() + const isCloud = manifest.provider !== 'local' + const retired = quality === null + const ready = state?.status === 'ready' + const downloading = state?.status === 'downloading' || state?.status === 'extracting' + const needsUpdate = state?.needsUpdate === true && !downloading + const totalMb = Math.round((manifest.sizeBytes ?? 0) / 1_000_000) + const percent = Math.round((state?.progress ?? 0) * 100) + + const deleteButton = ( + + ) + + let subline: React.ReactNode = null + if (retired) { + subline = t( + 'components.settings.voice.models.retired', + 'No longer offered. Delete it to free up space.' + ) + } else if (state?.status === 'extracting') { + subline = t('auto.components.settings.VoicePane.61a16c8141', 'Extracting...') + } else if (downloading) { + subline = t('speechModelUpdate.progress', '{{done}} of {{total}} MB', { + done: Math.round((state?.progress ?? 0) * totalMb), + total: totalMb + }) + } else if (needsUpdate) { + subline = ( + + {t( + 'auto.components.settings.VoicePane.updateToFasterVersion', + 'Update to the faster version ({{value0}} MB)', + { value0: totalMb } + )} + + ) + } else if (manifest.provider === 'openai') { + subline = t('components.settings.voice.models.sendsToOpenAi', 'Sends your audio to OpenAI.') + } else if (manifest.provider === 'openai-compatible') { + subline = t( + 'components.settings.voice.models.sendsToCompatible', + 'Sends your audio to the service you set up.' + ) + } + + let action: React.ReactNode + if (retired) { + action = ( + + ) + } else if (downloading) { + action = ( + <> + + + {percent}% + + + + ) + } else if (needsUpdate) { + action = ( + <> + + {deleteButton} + + ) + } else if (ready) { + action = ( + <> + {active ? ( + + + {t('components.settings.voice.models.inUse', 'In use')} + + ) : ( + + )} + {isCloud ? null : deleteButton} + + ) + } else if (isCloud) { + action = ( + + ) + } else { + action = ( + + ) + } + + return ( +
+
+
{manifest.label}
+ {subline ?
{subline}
: null} +
+
+ {quality ? : null} +
+
+ {isCloud + ? t('components.settings.voice.models.cloud', 'Cloud') + : retired + ? null + : t('auto.components.settings.VoicePane.91980ce124', '{{value0}} MB', { + value0: totalMb + })} +
+
{action}
+
+ ) +} diff --git a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.fixture.tsx b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.fixture.tsx index 6ebdd0c6c0b..42c20fd70fc 100644 --- a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.fixture.tsx +++ b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.fixture.tsx @@ -1,60 +1,112 @@ import { defineFixture } from '../../../component-fixtures/component-fixture' import { getDefaultVoiceSettings } from '../../../../../shared/constants' -import type { SpeechModelManifest } from '../../../../../shared/speech-types' +import type { + SpeechModelErrorRate, + SpeechModelManifest, + SpeechModelState +} from '../../../../../shared/speech-types' import { VoiceSpeechModelSection } from './VoiceSpeechModelSection' -function model(id: string, label: string, sizeBytes: number, description: string) { +const rates = (entries: Record): SpeechModelErrorRate[] => + Object.entries(entries).map(([language, percent]) => ({ + language, + percent, + metric: ['ja', 'ko', 'zh'].includes(language) ? 'cer' : 'wer', + test: language === 'en' ? 'librispeech' : 'fleurs', + clips: language === 'en' ? 73 : 5, + measuredOn: '2026-11-09' + })) + +function local( + id: string, + label: string, + sizeBytes: number, + languages: string[], + errorRates: SpeechModelErrorRate[], + replacedBy?: string +): SpeechModelManifest { return { id, label, - description, type: 'transducer', provider: 'local', - language: 'en', - languages: ['en'], + language: 'multilingual', + languages, + errorRates, sizeBytes, sampleRate: 16000, - streaming: false - } satisfies SpeechModelManifest + streaming: false, + replacedBy + } } -const catalog = [ - model('v3', 'Parakeet TDT v3', 739_508_576, 'Highest accuracy for 25 European languages.'), - model( - 'v2', - 'Parakeet TDT v2', - 729_574_912, - 'English only. Faster than v3 with similar accuracy.' - ), - model( - 'nemotron', - 'Nemotron 3.5', - 751_094_240, - '28 languages including Korean, Chinese and Japanese.' - ) +const catalog: SpeechModelManifest[] = [ + local('v3', 'Parakeet TDT v3', 739_508_576, ['en', 'de'], rates({ en: 3.7, de: 2.2 })), + local('v2', 'Parakeet TDT v2', 729_574_912, ['en'], rates({ en: 2.8 })), + local('zko', 'Zipformer Streaming KO', 0, ['ko'], [], 'sense'), + local('turbo', 'Whisper Large v3 Turbo', 886_381_760, ['en', 'ko'], rates({ en: 4.5, ko: 2.6 })), + local('sense', 'SenseVoice', 252_684_608, ['en', 'ko'], rates({ en: 6.9, ko: 10 })), + local('nemotron', 'Nemotron 3.5', 751_094_240, ['en', 'ko'], rates({ en: 5.3, ko: 11.8 })), + local('tiny', 'Whisper Tiny', 45_981_088, ['en', 'ko'], rates({ en: 13, ko: 24.9 })), + { + id: 'gpt', + label: 'GPT Transcribe', + type: 'openai', + provider: 'openai', + language: 'multilingual', + languages: ['en', 'ko'], + sampleRate: 16000, + streaming: false + } ] -export const updateAvailable = defineFixture({ - width: 640, - render: () => ( +function render( + sttModel: string, + modelStates: SpeechModelState[], + ratingLanguage: string +): React.JSX.Element { + return ( {}} onOpenCloudSetup={() => {}} onRefreshModelStates={() => {}} + cloudSettings={null} /> - ), - interact: (frame) => { - frame - .querySelector('button') - ?.dispatchEvent( - new PointerEvent('pointerdown', { bubbles: true, button: 0, pointerType: 'mouse' }) - ) - } + ) +} + +export const english = defineFixture({ + width: 760, + render: () => + render( + 'v3', + [ + { id: 'v3', status: 'ready' }, + { id: 'tiny', status: 'ready' }, + { id: 'sense', status: 'downloading', progress: 0.42 } + ], + 'en' + ) +}) + +export const korean = defineFixture({ + width: 760, + render: () => render('', [], 'ko') +}) + +export const updateAndRetired = defineFixture({ + width: 760, + render: () => + render( + 'v2', + [ + { id: 'v2', status: 'not-downloaded', needsUpdate: true }, + { id: 'zko', status: 'ready' } + ], + 'en' + ) }) diff --git a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.test.tsx b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.test.tsx index 6a5a0433801..b019dc293be 100644 --- a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.test.tsx +++ b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.test.tsx @@ -2,14 +2,12 @@ // @vitest-environment happy-dom import { act } from 'react' -import type { ReactNode } from 'react' import { createRoot, type Root } from 'react-dom/client' import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' import type { SpeechModelManifest, SpeechModelState } from '../../../../../shared/speech-types' import { getDefaultVoiceSettings } from '../../../../../shared/constants' const toastErrorMock = vi.hoisted(() => vi.fn()) -const menuDismissMock = vi.hoisted(() => vi.fn()) vi.mock('sonner', () => ({ toast: { @@ -18,50 +16,16 @@ vi.mock('sonner', () => ({ })) vi.mock('@/i18n/i18n', () => ({ + getIntlLocale: () => 'en', translate: (_key: string, fallback: string, values?: Record) => values ? fallback.replace('{{value0}}', values.value0) : fallback })) -vi.mock('../../ui/dropdown-menu', () => ({ - DropdownMenu: ({ children }: { children: ReactNode }) =>
{children}
, - DropdownMenuTrigger: ({ children }: { children: ReactNode }) => <>{children}, - DropdownMenuContent: ({ children }: { children: ReactNode }) =>
{children}
, - DropdownMenuItem: ({ - children, - disabled, - onSelect, - className - }: { - children: ReactNode - disabled?: boolean - onSelect?: (event: Event) => void - className?: string - }) => ( -
{ - if (!disabled) { - const selectEvent = new Event('select', { cancelable: true }) - onSelect?.(selectEvent) - if (!selectEvent.defaultPrevented) { - menuDismissMock() - } - } - }} - > - {children} -
- ) -})) - import { VoiceSpeechModelSection } from './VoiceSpeechModelSection' const localModel: SpeechModelManifest = { id: 'model-a', label: 'Local Model', - description: 'Runs offline', provider: 'local', language: 'en', languages: ['en'], @@ -107,8 +71,10 @@ function renderSection(args: { catalog={catalog} modelStates={modelStates} onUpdateVoiceSettings={vi.fn()} + ratingLanguage="en" onOpenCloudSetup={vi.fn()} onRefreshModelStates={args.refreshModelStates ?? vi.fn()} + cloudSettings={null} /> ) }) @@ -119,7 +85,6 @@ function renderSection(args: { describe('VoiceSpeechModelSection', () => { beforeEach(() => { toastErrorMock.mockReset() - menuDismissMock.mockReset() }) afterEach(() => { diff --git a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.tsx b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.tsx index aabb070722e..2d6d9b461c8 100644 --- a/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.tsx +++ b/src/renderer/src/components/settings/voice/VoiceSpeechModelSection.tsx @@ -1,247 +1,138 @@ import { useState } from 'react' import { toast } from 'sonner' import type { - VoiceSettings, SpeechModelManifest, - SpeechModelState + SpeechModelState, + VoiceSettings } from '../../../../../shared/speech-types' -import { Button } from '../../ui/button' -import { Label } from '../../ui/label' +import { useTranslate, useIntlLocale } from '@/i18n/use-translate' +import { languageDisplayName } from '@/i18n/ui-intl-formatters' import { - DropdownMenu, - DropdownMenuContent, - DropdownMenuItem, - DropdownMenuTrigger -} from '../../ui/dropdown-menu' -import { Cloud, Download, Trash2, Loader2, ChevronDown, Check } from 'lucide-react' -import { useTranslate } from '@/i18n/use-translate' -import { speechModelDescription } from './speech-model-description' -import { isOfferedSpeechModel } from '../../../../../shared/speech-model-retirement' - -function describeSpeechModelDownloadError(error: unknown): string { - const message = error instanceof Error ? error.message : String(error) - // Strip Electron’s invoke wrapper so download errors show the underlying cause. - return message.replace(/^Error invoking remote method '[^']+': (?:Error: )?/, '') -} + SettingsSegmentedControl, + SettingsSubsectionHeader +} from '../form-controls/SettingsFormControls' +import { + speechModelTab, + voiceSpeechModelRows, + type VoiceSpeechModelTab +} from './voice-speech-model-rows' +import { VoiceSpeechModelRow } from './VoiceSpeechModelRow' +import { cancelSpeechModelDownload, startSpeechModelDownload } from './speech-model-download' type VoiceSpeechModelSectionProps = { voiceSettings: VoiceSettings catalog: SpeechModelManifest[] modelStates: SpeechModelState[] + /** The language quality is rated for. */ + ratingLanguage: string onUpdateVoiceSettings: (updates: Partial) => void - /** A cloud model that isn't set up yet: open its provider's setup. */ - onOpenCloudSetup: (manifest: SpeechModelManifest) => void + onOpenCloudSetup: (modelId: string) => void onRefreshModelStates: () => void + /** Provider rows (API keys, base URL) shown under the cloud models. */ + cloudSettings: React.ReactNode } export function VoiceSpeechModelSection({ voiceSettings, catalog, modelStates, + ratingLanguage, onUpdateVoiceSettings, onOpenCloudSetup, - onRefreshModelStates + onRefreshModelStates, + cloudSettings }: VoiceSpeechModelSectionProps): React.JSX.Element { const t = useTranslate() - const [pendingDeleteModelIds, setPendingDeleteModelIds] = useState>(() => new Set()) - const getModelState = (id: string): SpeechModelState | undefined => - modelStates.find((s) => s.id === id) - + const intlLocale = useIntlLocale() const selectedModel = catalog.find((m) => m.id === voiceSettings.sttModel) - const selectedModelState = voiceSettings.sttModel - ? getModelState(voiceSettings.sttModel) - : undefined - const selectedIsReady = selectedModelState?.status === 'ready' + const [chosenTab, setTab] = useState(null) + const tab = chosenTab ?? (selectedModel ? speechModelTab(selectedModel) : 'local') + const [pendingDeleteModelIds, setPendingDeleteModelIds] = useState>( + () => new Set() + ) + const rows = voiceSpeechModelRows(catalog, modelStates, tab, ratingLanguage) + + const setDeletePending = (modelId: string, pending: boolean): void => + setPendingDeleteModelIds((prev) => { + const next = new Set(prev) + if (pending) { + next.add(modelId) + } else { + next.delete(modelId) + } + return next + }) + + const deleteModel = (modelId: string): void => { + if (pendingDeleteModelIds.has(modelId)) { + return + } + setDeletePending(modelId, true) + void window.api.speech + .deleteModel(modelId) + .then(onRefreshModelStates) + .catch(() => + toast.error(t('auto.components.settings.VoicePane.68de13f72c', 'Failed to delete model.')) + ) + .finally(() => setDeletePending(modelId, false)) + } return ( -
-
- -

- {selectedModel && selectedIsReady - ? `${selectedModel.label} — ${speechModelDescription(selectedModel)}` - : t( - 'auto.components.settings.VoicePane.e24f7d43d2', - 'Select a speech model. Local models run offline; cloud models require an API key.' - )} -

-
- - - - - - {catalog.map((manifest) => { - const mState = getModelState(manifest.id) - const isReady = mState?.status === 'ready' - const isCloud = manifest.provider !== 'local' - const isRetired = !isOfferedSpeechModel(manifest) - // Retired local models stay listed only while downloaded, so their files can be deleted. - if (isRetired && (isCloud || !isReady)) { - return null - } - const isDownloading = - mState?.status === 'downloading' || mState?.status === 'extracting' - const isActive = voiceSettings.sttModel === manifest.id - // Only the old sherpa-onnx files are on disk; downloading replaces them. - const needsUpdate = mState?.needsUpdate === true && !isDownloading - const deletePending = pendingDeleteModelIds.has(manifest.id) - const sizeMb = manifest.sizeBytes ? Math.round(manifest.sizeBytes / 1_000_000) : null - - return ( - { - if (isRetired) { - event.preventDefault() - } else if (isReady) { - onUpdateVoiceSettings({ sttModel: manifest.id }) - } else if (isCloud) { - onOpenCloudSetup(manifest) - } else if (!isDownloading) { - // Keep the menu open so download progress remains visible. - event.preventDefault() - void window.api.speech.downloadModel(manifest.id).catch((error: unknown) => - toast.error( - t( - 'auto.components.settings.VoicePane.cfde55c7b0', - 'Failed to download model.' - ), - // Include the raw cause because it is the diagnosable failure users can report. - { description: describeSpeechModelDownloadError(error) } - ) - ) - } - }} - className={`group flex items-center gap-2.5 py-2.5 ${ - !isCloud && !isReady && !isDownloading ? 'opacity-50' : '' - }`} - > - - {isActive && isReady ? ( - - ) : isDownloading ? ( - - ) : isCloud ? ( - - ) : null} - -
-
- {manifest.label} - {!isCloud && ( - - {manifest.streaming - ? t('auto.components.settings.VoicePane.d504ab05f0', 'streaming') - : t('auto.components.settings.VoicePane.8f4d2a51d7', 'offline')} - - )} - {manifest.recommended && ( - - {t('auto.components.settings.VoicePane.1ba81c0ff0', 'recommended')} - - )} - - {isDownloading && mState?.progress !== undefined - ? mState.status === 'extracting' - ? t('auto.components.settings.VoicePane.61a16c8141', 'Extracting...') - : `${Math.round(mState.progress * 100)}%` - : isCloud - ? null - : t('auto.components.settings.VoicePane.91980ce124', '{{value0}} MB', { - value0: sizeMb - })} - -
-

- {speechModelDescription(manifest)} -

- {needsUpdate && ( -

- {t( - 'auto.components.settings.VoicePane.updateToFasterVersion', - 'Update to the faster version ({{value0}} MB)', - { value0: sizeMb } - )} -

- )} -
- {!isCloud && (isReady || needsUpdate) ? ( - - ) : !isCloud && !isReady && !isDownloading ? ( - - - - ) : null} -
- ) - })} -
-
+ value={tab} + onChange={setTab} + ariaLabel={t('components.settings.voice.models.where', 'Where models run')} + options={[ + { + value: 'local', + label: t('components.settings.voice.models.local', 'On this computer') + }, + { value: 'cloud', label: t('components.settings.voice.models.cloud', 'Cloud') } + ]} + /> + } + /> + {tab === 'cloud' ? ( +

+ {t( + 'components.settings.voice.models.cloudDescription', + 'Cloud models send your audio to the provider. Models on this computer keep it here.' + )} +

+ ) : null} +
+ {rows.map((row) => ( +
+ onUpdateVoiceSettings({ sttModel: modelId })} + onDownload={(modelId) => startSpeechModelDownload(modelId, onRefreshModelStates)} + onCancelDownload={(modelId) => + cancelSpeechModelDownload(modelId, onRefreshModelStates) + } + onDelete={deleteModel} + onSetUpCloud={onOpenCloudSetup} + /> +
+ ))} +
+ {tab === 'cloud' ? cloudSettings : null}
) } diff --git a/src/renderer/src/components/settings/voice/speech-model-description.ts b/src/renderer/src/components/settings/voice/speech-model-description.ts deleted file mode 100644 index 1f1f0abffb4..00000000000 --- a/src/renderer/src/components/settings/voice/speech-model-description.ts +++ /dev/null @@ -1,69 +0,0 @@ -import type { SpeechModelManifest } from '../../../../../shared/speech-types' -import { translate } from '@/i18n/i18n' - -// Models transcribe-cpp can't run; their users moved to SenseVoice or Parakeet. -const retiredForEngine = (): string => - translate( - 'settings.voice.modelDescriptions.retiredForEngine', - 'No longer offered: the speech engine can no longer run it. Delete it to free up space.' - ) - -const DESCRIPTIONS: Record string> = { - 'parakeet-tdt-0.6b-v3-int8': () => - translate( - 'settings.voice.modelDescriptions.parakeetV3', - 'Highest accuracy for 25 European languages. Punctuation, capitalization, and word-level timestamps.' - ), - 'parakeet-tdt-0.6b-v2-int8': () => - translate( - 'settings.voice.modelDescriptions.parakeetV2', - 'English only. Faster than v3 with similar accuracy. Punctuation and capitalization.' - ), - 'zipformer-bilingual-zh-en': retiredForEngine, - 'paraformer-bilingual-zh-en': retiredForEngine, - 'zipformer-streaming-en-20m': () => - translate( - 'settings.voice.modelDescriptions.zipformerStreamingEnRetired', - 'No longer offered: it drops the start of sentences. Delete it to free up space.' - ), - 'zipformer-streaming-zh-14m': retiredForEngine, - 'zipformer-streaming-korean': retiredForEngine, - 'parakeet-tdt-ctc-0.6b-ja-int8': retiredForEngine, - 'whisper-tiny': () => - translate( - 'settings.voice.modelDescriptions.whisperTiny', - '90+ languages. Lower accuracy than Parakeet but broadest language coverage.' - ), - 'whisper-large-v3-turbo': () => - translate( - 'settings.voice.modelDescriptions.whisperLargeV3Turbo', - 'Best accuracy for Japanese, Korean, and Chinese, and 90+ languages. Slower than Parakeet.' - ), - 'sense-voice-zh-en-ja-ko-yue': () => - translate( - 'settings.voice.modelDescriptions.senseVoice', - 'Chinese, English, Japanese, Korean, and Cantonese with automatic language detection.' - ), - 'openai-compatible': () => - translate( - 'settings.voice.modelDescriptions.openAiCompatible', - 'Any service with an OpenAI-style transcription API, such as Groq, Mistral or a local server.' - ), - 'nemotron-3.5-asr-streaming-0.6b': () => - translate( - 'settings.voice.modelDescriptions.nemotron', - 'English, Spanish, German and many more languages, with automatic language detection.' - ), - 'openai-gpt-transcribe': () => - translate( - 'settings.voice.modelDescriptions.gptTranscribe', - 'Cloud transcription with high accuracy. Requires an OpenAI API key.' - ) -} - -/** Main's catalog arrives in English over IPC; unknown ids from a newer main keep it. */ -export function speechModelDescription(manifest: SpeechModelManifest): string { - return Object.hasOwn(DESCRIPTIONS, manifest.id) - ? DESCRIPTIONS[manifest.id]() - : manifest.description -} diff --git a/src/renderer/src/components/settings/voice/speech-model-download.ts b/src/renderer/src/components/settings/voice/speech-model-download.ts new file mode 100644 index 00000000000..db4b882650e --- /dev/null +++ b/src/renderer/src/components/settings/voice/speech-model-download.ts @@ -0,0 +1,26 @@ +import { toast } from 'sonner' +import { translate } from '@/i18n/i18n' + +function describeSpeechModelDownloadError(error: unknown): string { + const message = error instanceof Error ? error.message : String(error) + // Strip Electron’s invoke wrapper so download errors show the underlying cause. + return message.replace(/^Error invoking remote method '[^']+': (?:Error: )?/, '') +} + +/** Starts a model download; a failure becomes a toast, and model states refresh either way. */ +export function startSpeechModelDownload(modelId: string, refreshModelStates: () => void): void { + void window.api.speech + .downloadModel(modelId) + .catch((error: unknown) => + toast.error( + translate('auto.components.settings.VoicePane.cfde55c7b0', 'Failed to download model.'), + // Include the raw cause because it is the diagnosable failure users can report. + { description: describeSpeechModelDownloadError(error) } + ) + ) + .finally(refreshModelStates) +} + +export function cancelSpeechModelDownload(modelId: string, refreshModelStates: () => void): void { + void window.api.speech.cancelDownload(modelId).finally(refreshModelStates) +} diff --git a/src/renderer/src/components/settings/voice/use-voice-microphone-devices.ts b/src/renderer/src/components/settings/voice/use-voice-microphone-devices.ts new file mode 100644 index 00000000000..de2f0bf55af --- /dev/null +++ b/src/renderer/src/components/settings/voice/use-voice-microphone-devices.ts @@ -0,0 +1,67 @@ +import { useCallback, useEffect, useRef, useState } from 'react' +import { + listVoiceMicrophoneDevices, + type VoiceMicrophoneDevice +} from '@/components/dictation/microphone-devices' +import { useMountedRef } from '@/hooks/useMountedRef' + +function sameDeviceList( + a: readonly VoiceMicrophoneDevice[], + b: readonly VoiceMicrophoneDevice[] +): boolean { + return ( + a.length === b.length && + a.every((device, index) => { + const other = b[index] + return device.deviceId === other?.deviceId && device.label === other.label + }) + ) +} + +/** The live microphone list, rescanned on devicechange and whenever `rescanKey` changes. */ +export function useVoiceMicrophoneDevices(rescanKey: unknown): { + devices: VoiceMicrophoneDevice[] + devicesKnown: boolean + refreshDevices: () => Promise +} { + const [devices, setDevices] = useState([]) + const [devicesKnown, setDevicesKnown] = useState(false) + const mountedRef = useMountedRef() + // Bluetooth emits repeated devicechange events; reject stale enumeration results. + const refreshGenerationRef = useRef(0) + + const refreshDevices = useCallback(async (): Promise => { + const generation = refreshGenerationRef.current + 1 + refreshGenerationRef.current = generation + if (typeof navigator === 'undefined' || !navigator.mediaDevices?.enumerateDevices) { + return + } + let next: VoiceMicrophoneDevice[] = [] + try { + next = listVoiceMicrophoneDevices(await navigator.mediaDevices.enumerateDevices()) + } catch { + next = [] + } + if (!mountedRef.current || refreshGenerationRef.current !== generation) { + return + } + setDevicesKnown(next.length > 0) + setDevices((current) => (sameDeviceList(current, next) ? current : next)) + }, [mountedRef]) + + useEffect(() => { + void refreshDevices() + if (typeof navigator === 'undefined' || !navigator.mediaDevices?.addEventListener) { + return + } + const handleDeviceChange = (): void => { + void refreshDevices() + } + navigator.mediaDevices.addEventListener('devicechange', handleDeviceChange) + return () => { + navigator.mediaDevices.removeEventListener('devicechange', handleDeviceChange) + } + }, [refreshDevices, rescanKey]) + + return { devices, devicesKnown, refreshDevices } +} diff --git a/src/renderer/src/components/settings/voice/voice-speech-model-rows.test.ts b/src/renderer/src/components/settings/voice/voice-speech-model-rows.test.ts new file mode 100644 index 00000000000..0521ff111ed --- /dev/null +++ b/src/renderer/src/components/settings/voice/voice-speech-model-rows.test.ts @@ -0,0 +1,42 @@ +import { describe, expect, it } from 'vitest' +import type { SpeechModelManifest } from '../../../../../shared/speech-types' +import { voiceSpeechModelRows } from './voice-speech-model-rows' + +const model = ( + id: string, + languages: string[], + extra: Partial = {} +): SpeechModelManifest => ({ + id, + label: id, + type: 'transducer', + provider: 'local', + language: 'multilingual', + languages, + sampleRate: 16000, + streaming: false, + ...extra +}) + +describe('voiceSpeechModelRows', () => { + it('puts models without the language last and lists retired models only while on disk', () => { + const catalog = [ + model('english-only', ['en']), + model('retired-on-disk', ['ko'], { replacedBy: 'korean' }), + model('korean', ['en', 'ko']), + model('retired-gone', ['ko'], { replacedBy: 'korean' }), + model('cloud', ['ko'], { provider: 'openai' }) + ] + const rows = voiceSpeechModelRows( + catalog, + [{ id: 'retired-on-disk', status: 'ready' }], + 'local', + 'ko' + ) + expect(rows.map((row) => row.manifest.id)).toEqual([ + 'korean', + 'english-only', + 'retired-on-disk' + ]) + }) +}) diff --git a/src/renderer/src/components/settings/voice/voice-speech-model-rows.ts b/src/renderer/src/components/settings/voice/voice-speech-model-rows.ts new file mode 100644 index 00000000000..5e58965d70f --- /dev/null +++ b/src/renderer/src/components/settings/voice/voice-speech-model-rows.ts @@ -0,0 +1,47 @@ +import type { SpeechModelManifest, SpeechModelState } from '../../../../../shared/speech-types' +import { isOfferedSpeechModel } from '../../../../../shared/speech-model-retirement' +import { + speechModelQuality, + type SpeechModelQuality +} from '../../../../../shared/speech-model-quality' + +export type VoiceSpeechModelTab = 'local' | 'cloud' + +export type VoiceSpeechModelRowData = { + manifest: SpeechModelManifest + state: SpeechModelState | undefined + /** null for a retired model, which is only listed so its files can be deleted. */ + quality: SpeechModelQuality | null +} + +export function speechModelTab(manifest: SpeechModelManifest): VoiceSpeechModelTab { + return manifest.provider === 'local' ? 'local' : 'cloud' +} + +/** A tab's rows: models for the language first, then ones that can't handle it, then retired downloads. */ +export function voiceSpeechModelRows( + catalog: readonly SpeechModelManifest[], + modelStates: readonly SpeechModelState[], + tab: VoiceSpeechModelTab, + language: string +): VoiceSpeechModelRowData[] { + const supported: VoiceSpeechModelRowData[] = [] + const unsupported: VoiceSpeechModelRowData[] = [] + const retired: VoiceSpeechModelRowData[] = [] + for (const manifest of catalog) { + if (speechModelTab(manifest) !== tab) { + continue + } + const state = modelStates.find((s) => s.id === manifest.id) + if (!isOfferedSpeechModel(manifest)) { + // Retired models stay listed only while their files are on disk. + if (manifest.provider === 'local' && (state?.status === 'ready' || state?.needsUpdate)) { + retired.push({ manifest, state, quality: null }) + } + continue + } + const quality = speechModelQuality(manifest, language) + ;(quality.kind === 'unsupported' ? unsupported : supported).push({ manifest, state, quality }) + } + return [...supported, ...unsupported, ...retired] +} diff --git a/src/renderer/src/i18n/locales/de.json b/src/renderer/src/i18n/locales/de.json index e8b5fce4702..7281314d8d6 100644 --- a/src/renderer/src/i18n/locales/de.json +++ b/src/renderer/src/i18n/locales/de.json @@ -972,19 +972,6 @@ "missingHost": "Die Proxy-URL muss einen Host enthalten." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Höchste Genauigkeit für 25 europäische Sprachen. Zeichensetzung, Groß- und Kleinschreibung und Zeitstempel auf Wortebene.", - "parakeetV2": "Nur Englisch. Schneller als v3 bei ähnlicher Genauigkeit. Zeichensetzung und Groß- und Kleinschreibung.", - "zipformerStreamingEnRetired": "Wird nicht mehr angeboten: Es verschluckt den Anfang von Sätzen. Löschen Sie es, um Speicherplatz freizugeben.", - "whisperTiny": "Über 90 Sprachen. Geringere Genauigkeit als Parakeet, aber die größte Sprachabdeckung.", - "senseVoice": "Chinesisch, Englisch, Japanisch, Koreanisch und Kantonesisch mit automatischer Erkennung der gesprochenen Sprache.", - "gptTranscribe": "Cloud-Transkription mit hoher Genauigkeit. Benötigt einen OpenAI-API-Schlüssel.", - "whisperLargeV3Turbo": "Beste Genauigkeit für Japanisch, Koreanisch und Chinesisch sowie über 90 Sprachen. Langsamer als Parakeet.", - "retiredForEngine": "Nicht mehr verfügbar: Die Spracherkennung kann dieses Modell nicht mehr ausführen. Löschen Sie es, um Speicherplatz freizugeben.", - "nemotron": "Englisch, Spanisch, Deutsch und viele weitere Sprachen, mit automatischer Spracherkennung." - } - }, "diagnostics": { "errors": { "creatingDisabled": "das Erstellen von Dateien zur Prüfung ist deaktiviert", @@ -8946,7 +8933,6 @@ "7807c11c4d": "Allgemein", "6855b0f77d": "Die grundlegenden Abläufe abschließen, die Orca für die parallele Arbeit mit Agenten nützlich machen.", "6d119427ef": "Checkliste für den Einstieg", - "eb1176a14e": "Lokales Sprachdiktat mit Modellen auf dem Gerät.", "5063bb47a5": "Spracheingabe", "7118953f14": "Agenten die Steuerung beliebiger Apps auf dem Computer ermöglichen.", "c9841721cb": "Computersteuerung", @@ -9461,7 +9447,6 @@ "VoicePane": { "6fa734ed95": "{{value0}} löschen", "68de13f72c": "Modell konnte nicht gelöscht werden.", - "1ba81c0ff0": "empfohlen", "cfde55c7b0": "Modell konnte nicht heruntergeladen werden.", "43fd4f454b": "Spracherkennungsmodell", "ba4a900d1d": "Diktatmodus", @@ -9472,10 +9457,6 @@ "506df81ba6": "OpenAI-API-Schlüssel gespeichert", "91980ce124": "{{value0}} MB", "61a16c8141": "Wird entpackt…", - "8f4d2a51d7": "offline", - "d504ab05f0": "Streaming", - "fbe5990716": "Modell auswählen", - "e24f7d43d2": "Ein Spracherkennungsmodell auswählen. Lokale Modelle funktionieren offline; Cloud-Modelle benötigen einen API-Schlüssel.", "174da92062": "Gedrückt halten", "118b3c2dee": "Umschalten", "ad5d036ecc": "Mikrofonberechtigung konnte nicht angefordert werden. Das Diktat wurde nicht aktiviert.", @@ -9483,7 +9464,6 @@ "1eac933202": "Die macOS-Einstellungen „Privacy & Security“ wurden geöffnet. Nach der Zugriffsfreigabe das Diktat erneut aktivieren.", "cd9fe37556": "Mikrofonberechtigung erteilt", "openAiKeyName": "OpenAI-Transkriptionsschlüssel", - "dictationShortcutDescription": "Mit {{shortcut}} Text in den jeweils fokussierten Bereich diktieren.", "dictationModeAuto": "Automatisch", "dictationModesWithAutoDescription": "Automatisch: Halten Sie {{shortcut}} gedrückt, um bis zum Loslassen zu diktieren, oder tippen Sie kurz zum Starten und Stoppen. Gedrückt halten: Diktieren, solange {{shortcut}} gedrückt bleibt. Umschalten: {{shortcut}} einmal zum Starten und erneut zum Stoppen drücken.", "speechGpuLabel": "GPU verwenden", @@ -20700,6 +20680,67 @@ "fillerWords": { "title": "Füllwörter entfernen", "description": "Entfernt Laute wie „äh“ und „ähm“ aus diktiertem Text, passend zur Diktiersprache." + }, + "sectionDescription": "Sprechen statt tippen, in jedem Bereich. Die Sprache wird auf diesem Computer erkannt, außer du wählst ein Cloud-Modell.", + "input": { + "title": "Eingabe", + "shortcut": "Tastenkürzel", + "shortcutDescription": "Startet das Diktieren im fokussierten Bereich.", + "changeShortcut": "Ändern" + }, + "language": { + "title": "Sprache" + }, + "cleanup": { + "title": "Bereinigung" + }, + "advanced": { + "title": "Erweitert" + }, + "setup": { + "downloadingDescription": "Du kannst diese Seite verlassen. Diktieren funktioniert, sobald der Download fertig ist.", + "title": "Diktieren einrichten", + "pickModel": "Sprechen statt tippen, in jedem Bereich. Lade unten ein Sprachmodell herunter; Modelle auf diesem Computer behalten deine Audiodaten hier.", + "microphoneMissingTitle": "{{microphone}} ist nicht verbunden", + "microphoneMissingDescription": "Diktieren nutzt das Standardmikrofon des Systems, bis es wieder da ist.", + "changeMicrophone": "Mikrofon ändern", + "onTitle": "Diktieren ist an", + "onLocal": "Drücke {{shortcut}} in einem beliebigen Bereich. {{model}} läuft auf diesem Computer.", + "onCloud": "Drücke {{shortcut}} in einem beliebigen Bereich. {{model}} läuft in der Cloud.", + "offTitle": "Diktieren ist aus", + "offDescription": "Schalte es ein, um mit {{shortcut}} in jedem Bereich zu diktieren.", + "savedMicrophone": "Dein Mikrofon" + }, + "quality": { + "unsupported": "Unterstützt deine Sprache nicht", + "unmeasured": "Nicht gemessen", + "poor": "Schwach", + "fair": "Mäßig", + "good": "Gut", + "veryGood": "Sehr gut", + "excellent": "Hervorragend", + "characterErrorRate": "{{percent}} % Zeichenfehlerrate", + "wordErrorRate": "{{percent}} % Wortfehlerrate", + "librispeech": "Test: LibriSpeech (vorgelesene Sprache), {{clips}} Clips, gemessen am {{date}}.", + "fleurs": "Test: FLEURS (vorgelesene Sprache), {{clips}} Clips, gemessen am {{date}}.", + "bench": "Orcas Benchmark auf einem Mac mit Apple Silicon.", + "label": "Qualität: {{word}}" + }, + "models": { + "retired": "Wird nicht mehr angeboten. Lösche es, um Speicherplatz freizugeben.", + "sendsToOpenAi": "Sendet deine Audiodaten an OpenAI.", + "delete": "Löschen", + "cancelDownload": "Download abbrechen", + "inUse": "In Verwendung", + "use": "Verwenden", + "setUpCloud": "Einrichten", + "download": "Herunterladen", + "cloud": "Cloud", + "description": "Die Qualität ist für {{language}} bewertet. Fahre über eine Bewertung, um den Test dahinter zu sehen.", + "where": "Wo Modelle laufen", + "local": "Auf diesem Computer", + "cloudDescription": "Cloud-Modelle senden deine Audiodaten an den Anbieter. Modelle auf diesem Computer behalten sie hier.", + "sendsToCompatible": "Sendet deine Audiodaten an den Dienst, den du eingerichtet hast." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/en.json b/src/renderer/src/i18n/locales/en.json index 03b11ac3bb5..d28e80b8a05 100644 --- a/src/renderer/src/i18n/locales/en.json +++ b/src/renderer/src/i18n/locales/en.json @@ -1144,20 +1144,6 @@ "missingHost": "Proxy URL must include a host." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Highest accuracy for 25 European languages. Punctuation, capitalization, and word-level timestamps.", - "parakeetV2": "English only. Faster than v3 with similar accuracy. Punctuation and capitalization.", - "zipformerStreamingEnRetired": "No longer offered: it drops the start of sentences. Delete it to free up space.", - "whisperTiny": "90+ languages. Lower accuracy than Parakeet but broadest language coverage.", - "senseVoice": "Chinese, English, Japanese, Korean, and Cantonese with automatic language detection.", - "gptTranscribe": "Cloud transcription with high accuracy. Requires an OpenAI API key.", - "openAiCompatible": "Any service with an OpenAI-style transcription API, such as Groq, Mistral or a local server.", - "whisperLargeV3Turbo": "Best accuracy for Japanese, Korean, and Chinese, and 90+ languages. Slower than Parakeet.", - "retiredForEngine": "No longer offered: the speech engine can no longer run it. Delete it to free up space.", - "nemotron": "English, Spanish, German and many more languages, with automatic language detection." - } - }, "diagnostics": { "errors": { "creatingDisabled": "creating review files is disabled", @@ -9264,7 +9250,6 @@ "7807c11c4d": "General", "6855b0f77d": "Finish the core workflows that make Orca useful for parallel agent work.", "6d119427ef": "Onboarding checklist", - "eb1176a14e": "Local speech-to-text dictation with on-device models.", "5063bb47a5": "Voice", "7118953f14": "Enable agents to control any app on your computer.", "c9841721cb": "Computer Use", @@ -9783,7 +9768,6 @@ "VoicePane": { "6fa734ed95": "Delete {{value0}}", "68de13f72c": "Failed to delete model.", - "1ba81c0ff0": "recommended", "cfde55c7b0": "Failed to download model.", "43fd4f454b": "Speech Model", "ba4a900d1d": "Dictation Mode", @@ -9794,10 +9778,6 @@ "506df81ba6": "OpenAI API key saved", "91980ce124": "{{value0}} MB", "61a16c8141": "Extracting...", - "8f4d2a51d7": "offline", - "d504ab05f0": "streaming", - "fbe5990716": "Select Model", - "e24f7d43d2": "Select a speech model. Local models run offline; cloud models require an API key.", "174da92062": "Hold", "118b3c2dee": "Toggle", "ad5d036ecc": "Could not request microphone permission. Voice dictation was not enabled.", @@ -9805,7 +9785,6 @@ "1eac933202": "Opened macOS Privacy & Security. Enable dictation again after granting access.", "cd9fe37556": "Microphone permission granted", "openAiKeyName": "Your OpenAI transcription key", - "dictationShortcutDescription": "Press {{shortcut}} to dictate text into any focused pane.", "dictationModeAuto": "Auto", "dictationModesWithAutoDescription": "Auto: hold {{shortcut}} to dictate until you let go, or tap it to start and stop. Hold: dictate while {{shortcut}} is held. Toggle: press {{shortcut}} once to start, again to stop.", "speechGpuLabel": "Use GPU", @@ -21401,6 +21380,67 @@ "edit": "Edit", "removeLabel": "Remove OpenAI-compatible provider", "setUp": "Set up" + }, + "sectionDescription": "Talk instead of typing in any pane. Speech is recognized on this computer unless you pick a cloud model.", + "input": { + "title": "Input", + "shortcut": "Shortcut", + "shortcutDescription": "Starts dictation in the focused pane.", + "changeShortcut": "Change" + }, + "language": { + "title": "Language" + }, + "cleanup": { + "title": "Cleanup" + }, + "advanced": { + "title": "Advanced" + }, + "setup": { + "downloadingDescription": "You can leave this page. Dictation works once the download finishes.", + "title": "Set up dictation", + "pickModel": "Talk instead of typing in any pane. Download a speech model below to start; models on this computer keep your audio here.", + "microphoneMissingTitle": "{{microphone}} isn’t connected", + "microphoneMissingDescription": "Dictation uses the system default microphone until it’s back.", + "changeMicrophone": "Change Microphone", + "onTitle": "Dictation is on", + "onLocal": "Press {{shortcut}} in any pane. {{model}} runs on this computer.", + "onCloud": "Press {{shortcut}} in any pane. {{model}} runs in the cloud.", + "offTitle": "Dictation is off", + "offDescription": "Turn it on to dictate with {{shortcut}} in any pane.", + "savedMicrophone": "Your microphone" + }, + "quality": { + "unsupported": "Doesn’t support your language", + "unmeasured": "Not measured", + "poor": "Poor", + "fair": "Fair", + "good": "Good", + "veryGood": "Very good", + "excellent": "Excellent", + "characterErrorRate": "{{percent}}% character error rate", + "wordErrorRate": "{{percent}}% word error rate", + "librispeech": "Test: LibriSpeech read speech, {{clips}} clips, measured {{date}}.", + "fleurs": "Test: FLEURS read speech, {{clips}} clips, measured {{date}}.", + "bench": "Orca’s benchmark on a Mac with Apple silicon.", + "label": "Quality: {{word}}" + }, + "models": { + "retired": "No longer offered. Delete it to free up space.", + "sendsToOpenAi": "Sends your audio to OpenAI.", + "sendsToCompatible": "Sends your audio to the service you set up.", + "delete": "Delete", + "cancelDownload": "Cancel download", + "inUse": "In use", + "use": "Use", + "setUpCloud": "Set up", + "download": "Download", + "cloud": "Cloud", + "description": "Quality is rated for {{language}}. Hover a rating to see the test behind it.", + "where": "Where models run", + "local": "On this computer", + "cloudDescription": "Cloud models send your audio to the provider. Models on this computer keep it here." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/es.json b/src/renderer/src/i18n/locales/es.json index 006202c63ab..cc1294d922c 100644 --- a/src/renderer/src/i18n/locales/es.json +++ b/src/renderer/src/i18n/locales/es.json @@ -503,19 +503,6 @@ "missingHost": "La URL de proxy debe incluir un host." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Máxima precisión para 25 idiomas europeos. Puntuación, mayúsculas y minúsculas y marcas de tiempo por palabra.", - "parakeetV2": "Solo inglés. Más rápido que v3 con una precisión similar. Puntuación y mayúsculas y minúsculas.", - "zipformerStreamingEnRetired": "Ya no se ofrece: omite el comienzo de las frases. Elimínalo para liberar espacio.", - "whisperTiny": "Más de 90 idiomas. Menor precisión que Parakeet, pero con la mayor cobertura de idiomas.", - "senseVoice": "Chino, inglés, japonés, coreano y cantonés con detección automática de idioma.", - "gptTranscribe": "Transcripción en la nube con alta precisión. Requiere una clave API de OpenAI.", - "whisperLargeV3Turbo": "La mejor precisión en japonés, coreano y chino, y más de 90 idiomas. Más lento que Parakeet.", - "retiredForEngine": "Ya no se ofrece: el motor de voz ya no puede ejecutarlo. Elimínelo para liberar espacio.", - "nemotron": "Inglés, español, alemán y muchos idiomas más, con detección automática del idioma." - } - }, "diagnostics": { "errors": { "creatingDisabled": "la creación de archivos de revisión está desactivada", @@ -8529,7 +8516,6 @@ "7807c11c4d": "General", "6855b0f77d": "Termina los flujos de trabajo principales que hacen que Orca sea útil para el trabajo de agentes en paralelo.", "6d119427ef": "Lista de verificación de incorporación", - "eb1176a14e": "Dictado local de voz a texto con modelos en el dispositivo.", "5063bb47a5": "Voz", "7118953f14": "Permite que los agentes controlen cualquier app en tu computadora.", "c9841721cb": "Computer Use", @@ -9039,7 +9025,6 @@ }, "VoicePane": { "68de13f72c": "No se pudo eliminar el modelo.", - "1ba81c0ff0": "recomendado", "cfde55c7b0": "No se pudo descargar el modelo.", "43fd4f454b": "Modelo de voz", "ba4a900d1d": "Modo de dictado", @@ -9050,10 +9035,6 @@ "506df81ba6": "Clave API de OpenAI guardada", "91980ce124": "{{value0}} MB", "61a16c8141": "Extrayendo...", - "8f4d2a51d7": "sin conexión", - "d504ab05f0": "streaming", - "fbe5990716": "Seleccionar modelo", - "e24f7d43d2": "Selecciona un modelo de voz. Los modelos locales funcionan sin conexión; los modelos en la nube requieren una clave API.", "174da92062": "Mantener", "118b3c2dee": "Alternar", "ad5d036ecc": "No se pudo solicitar permiso para el micrófono. El dictado de voz no se habilitó.", @@ -9062,7 +9043,6 @@ "cd9fe37556": "Permiso de micrófono concedido", "6fa734ed95": "Eliminar {{value0}}", "openAiKeyName": "Su clave de transcripción OpenAI", - "dictationShortcutDescription": "Pulsa {{shortcut}} para dictar texto en cualquier panel con foco.", "dictationModeAuto": "Automático", "dictationModesWithAutoDescription": "Automático: mantén {{shortcut}} para dictar hasta que lo sueltes, o tócalo para empezar y parar. Mantener: dicta mientras {{shortcut}} se mantiene presionado. Alternar: pulsa {{shortcut}} una vez para empezar y otra para parar.", "speechGpuLabel": "Usar GPU", @@ -20488,6 +20468,67 @@ "fillerWords": { "title": "Quitar muletillas", "description": "Quita sonidos como «ehm» y «em» del texto dictado, según el idioma de dictado." + }, + "sectionDescription": "Habla en lugar de escribir en cualquier panel. La voz se reconoce en este equipo, salvo que elijas un modelo en la nube.", + "input": { + "title": "Entrada", + "shortcut": "Atajo", + "shortcutDescription": "Inicia el dictado en el panel activo.", + "changeShortcut": "Cambiar" + }, + "language": { + "title": "Idioma" + }, + "cleanup": { + "title": "Limpieza" + }, + "advanced": { + "title": "Avanzado" + }, + "setup": { + "downloadingDescription": "Puedes salir de esta página. El dictado funcionará cuando termine la descarga.", + "title": "Configurar el dictado", + "pickModel": "Habla en lugar de escribir en cualquier panel. Descarga abajo un modelo de voz para empezar; los modelos de este equipo mantienen tu audio aquí.", + "microphoneMissingTitle": "{{microphone}} no está conectado", + "microphoneMissingDescription": "El dictado usa el micrófono predeterminado del sistema hasta que vuelva.", + "changeMicrophone": "Cambiar micrófono", + "onTitle": "El dictado está activado", + "onLocal": "Pulsa {{shortcut}} en cualquier panel. {{model}} se ejecuta en este equipo.", + "onCloud": "Pulsa {{shortcut}} en cualquier panel. {{model}} se ejecuta en la nube.", + "offTitle": "El dictado está desactivado", + "offDescription": "Actívalo para dictar con {{shortcut}} en cualquier panel.", + "savedMicrophone": "Tu micrófono" + }, + "quality": { + "unsupported": "No admite tu idioma", + "unmeasured": "Sin medir", + "poor": "Mala", + "fair": "Regular", + "good": "Buena", + "veryGood": "Muy buena", + "excellent": "Excelente", + "characterErrorRate": "{{percent}} % de error por carácter", + "wordErrorRate": "{{percent}} % de error por palabra", + "librispeech": "Prueba: LibriSpeech (habla leída), {{clips}} clips, medida el {{date}}.", + "fleurs": "Prueba: FLEURS (habla leída), {{clips}} clips, medida el {{date}}.", + "bench": "Benchmark de Orca en un Mac con Apple silicon.", + "label": "Calidad: {{word}}" + }, + "models": { + "retired": "Ya no se ofrece. Bórralo para liberar espacio.", + "sendsToOpenAi": "Envía tu audio a OpenAI.", + "delete": "Borrar", + "cancelDownload": "Cancelar descarga", + "inUse": "En uso", + "use": "Usar", + "setUpCloud": "Configurar", + "download": "Descargar", + "cloud": "Nube", + "description": "La calidad se valora para {{language}}. Pasa el cursor sobre una valoración para ver la prueba.", + "where": "Dónde se ejecutan los modelos", + "local": "En este equipo", + "cloudDescription": "Los modelos en la nube envían tu audio al proveedor. Los de este equipo lo mantienen aquí.", + "sendsToCompatible": "Envía tu audio al servicio que configuraste." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/fr.json b/src/renderer/src/i18n/locales/fr.json index c9887b7ca7f..c0198f16351 100644 --- a/src/renderer/src/i18n/locales/fr.json +++ b/src/renderer/src/i18n/locales/fr.json @@ -742,19 +742,6 @@ "missingHost": "L’URL du proxy doit inclure un hôte." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Précision maximale pour 25 langues européennes. Ponctuation, majuscules et minuscules, et horodatage par mot.", - "parakeetV2": "Anglais uniquement. Plus rapide que v3 avec une précision similaire. Ponctuation, majuscules et minuscules.", - "zipformerStreamingEnRetired": "N’est plus proposé : il omet le début des phrases. Supprimez-le pour libérer de l’espace.", - "whisperTiny": "Plus de 90 langues. Moins précis que Parakeet, mais offre la plus large couverture linguistique.", - "senseVoice": "Chinois, anglais, japonais, coréen et cantonais avec détection automatique de la langue.", - "gptTranscribe": "Transcription dans le cloud avec une grande précision. Nécessite une clé API OpenAI.", - "whisperLargeV3Turbo": "Meilleure précision pour le japonais, le coréen et le chinois, et plus de 90 langues. Plus lent que Parakeet.", - "retiredForEngine": "N’est plus proposé : le moteur vocal ne peut plus l’exécuter. Supprimez-le pour libérer de l’espace.", - "nemotron": "Anglais, espagnol, allemand et bien d’autres langues, avec détection automatique de la langue." - } - }, "diagnostics": { "errors": { "creatingDisabled": "la création de fichiers de révision est désactivée", @@ -8814,7 +8801,6 @@ "7807c11c4d": "Général", "6855b0f77d": "Terminez les workflows essentiels qui rendent Orca utile au travail parallèle avec des agents.", "6d119427ef": "Checklist d'intégration", - "eb1176a14e": "Dictée vocale locale avec modèles embarqués sur l'appareil.", "5063bb47a5": "Voix", "7118953f14": "Permettez aux agents de contrôler n'importe quelle app de votre ordinateur.", "c9841721cb": "Computer Use", @@ -9329,7 +9315,6 @@ "VoicePane": { "6fa734ed95": "Supprimer {{value0}}", "68de13f72c": "Échec de la suppression du modèle.", - "1ba81c0ff0": "recommandé", "cfde55c7b0": "Échec du téléchargement du modèle.", "43fd4f454b": "Modèle vocal", "ba4a900d1d": "Mode de dictée", @@ -9340,10 +9325,6 @@ "506df81ba6": "Clé API OpenAI enregistrée", "91980ce124": "{{value0}} Mo", "61a16c8141": "Extraction...", - "8f4d2a51d7": "hors ligne", - "d504ab05f0": "streaming", - "fbe5990716": "Sélectionner un modèle", - "e24f7d43d2": "Sélectionnez un modèle vocal. Les modèles locaux fonctionnent hors ligne ; les modèles cloud nécessitent une clé API.", "174da92062": "Maintien", "118b3c2dee": "Bascule", "ad5d036ecc": "Impossible de demander l'autorisation d'accès au micro. La dictée vocale n'a pas été activée.", @@ -9351,7 +9332,6 @@ "1eac933202": "Panneau Confidentialité et sécurité de macOS ouvert. Réactivez la dictée après avoir accordé l'accès.", "cd9fe37556": "Autorisation d'accès au micro accordée", "openAiKeyName": "Votre clé de transcription OpenAI", - "dictationShortcutDescription": "Appuyez sur {{shortcut}} pour dicter du texte dans n'importe quel volet actif.", "dictationModeAuto": "Auto", "dictationModesWithAutoDescription": "Auto : maintenez {{shortcut}} pour dicter jusqu’à ce que vous relâchiez, ou appuyez brièvement pour démarrer et arrêter. Maintien : dictez tant que {{shortcut}} est maintenu. Bascule : appuyez une fois sur {{shortcut}} pour démarrer, puis à nouveau pour arrêter.", "speechGpuLabel": "Utiliser le GPU", @@ -20805,6 +20785,67 @@ "fillerWords": { "title": "Supprimer les hésitations", "description": "Supprime les sons comme « euh » et « hum » du texte dicté, selon la langue de dictée." + }, + "sectionDescription": "Parlez au lieu de taper, dans n’importe quel volet. La parole est reconnue sur cet ordinateur, sauf si vous choisissez un modèle cloud.", + "input": { + "title": "Entrée", + "shortcut": "Raccourci", + "shortcutDescription": "Lance la dictée dans le volet actif.", + "changeShortcut": "Modifier" + }, + "language": { + "title": "Langue" + }, + "cleanup": { + "title": "Nettoyage" + }, + "advanced": { + "title": "Avancé" + }, + "setup": { + "downloadingDescription": "Vous pouvez quitter cette page. La dictée fonctionnera une fois le téléchargement terminé.", + "title": "Configurer la dictée", + "pickModel": "Parlez au lieu de taper, dans n’importe quel volet. Téléchargez un modèle vocal ci-dessous pour commencer ; les modèles de cet ordinateur gardent votre audio ici.", + "microphoneMissingTitle": "{{microphone}} n’est pas connecté", + "microphoneMissingDescription": "La dictée utilise le micro par défaut du système jusqu’à son retour.", + "changeMicrophone": "Changer de micro", + "onTitle": "La dictée est activée", + "onLocal": "Appuyez sur {{shortcut}} dans n’importe quel volet. {{model}} tourne sur cet ordinateur.", + "onCloud": "Appuyez sur {{shortcut}} dans n’importe quel volet. {{model}} tourne dans le cloud.", + "offTitle": "La dictée est désactivée", + "offDescription": "Activez-la pour dicter avec {{shortcut}} dans n’importe quel volet.", + "savedMicrophone": "Votre micro" + }, + "quality": { + "unsupported": "Ne prend pas en charge votre langue", + "unmeasured": "Non mesuré", + "poor": "Faible", + "fair": "Passable", + "good": "Bonne", + "veryGood": "Très bonne", + "excellent": "Excellente", + "characterErrorRate": "{{percent}} % d’erreurs par caractère", + "wordErrorRate": "{{percent}} % d’erreurs par mot", + "librispeech": "Test : LibriSpeech (parole lue), {{clips}} extraits, mesuré le {{date}}.", + "fleurs": "Test : FLEURS (parole lue), {{clips}} extraits, mesuré le {{date}}.", + "bench": "Benchmark d’Orca sur un Mac avec puce Apple.", + "label": "Qualité : {{word}}" + }, + "models": { + "retired": "N’est plus proposé. Supprimez-le pour libérer de l’espace.", + "sendsToOpenAi": "Envoie votre audio à OpenAI.", + "delete": "Supprimer", + "cancelDownload": "Annuler le téléchargement", + "inUse": "Utilisé", + "use": "Utiliser", + "setUpCloud": "Configurer", + "download": "Télécharger", + "cloud": "Cloud", + "description": "La qualité est évaluée pour {{language}}. Survolez une note pour voir le test.", + "where": "Où les modèles tournent", + "local": "Sur cet ordinateur", + "cloudDescription": "Les modèles cloud envoient votre audio au fournisseur. Ceux de cet ordinateur le gardent ici.", + "sendsToCompatible": "Envoie votre audio au service que vous avez configuré." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/ja.json b/src/renderer/src/i18n/locales/ja.json index 613f7cac950..21f5f68a36a 100644 --- a/src/renderer/src/i18n/locales/ja.json +++ b/src/renderer/src/i18n/locales/ja.json @@ -497,19 +497,6 @@ "missingHost": "プロキシ URL にはホストが必要です。" } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "ヨーロッパの 25 言語で最高の精度。句読点、大文字・小文字、単語単位のタイムスタンプに対応。", - "parakeetV2": "英語のみ。v3 と同程度の精度で、より高速。句読点と大文字・小文字に対応。", - "zipformerStreamingEnRetired": "提供を終了しました。文の冒頭が抜け落ちます。削除して空き容量を確保してください。", - "whisperTiny": "90 以上の言語に対応。Parakeet より精度は低いものの、対応言語が最も豊富。", - "senseVoice": "中国語、英語、日本語、韓国語、広東語に対応。言語を自動検出。", - "gptTranscribe": "高精度なクラウド文字起こし。OpenAI API キーが必要です。", - "whisperLargeV3Turbo": "日本語・韓国語・中国語で最も高い精度。90以上の言語に対応。Parakeetより低速です。", - "retiredForEngine": "提供終了: 音声エンジンでこのモデルを実行できなくなりました。削除して空き容量を確保してください。", - "nemotron": "英語、スペイン語、ドイツ語など多数の言語に対応。言語を自動判別します。" - } - }, "diagnostics": { "errors": { "creatingDisabled": "確認用ファイルの作成は無効になっています", @@ -8273,7 +8260,6 @@ "7807c11c4d": "一般", "6855b0f77d": "Orca を並列 Agent 作業に活用するためのコアワークフローを完了します。", "6d119427ef": "オンボーディングチェックリスト", - "eb1176a14e": "オンデバイスモデルを使用したローカルの音声からテキストへのディクテーション。", "5063bb47a5": "音声", "7118953f14": "Agent がコンピュータ上の任意のアプリを制御できるようにします。", "c9841721cb": "コンピュータ操作", @@ -8783,7 +8769,6 @@ }, "VoicePane": { "68de13f72c": "モデルの削除に失敗しました。", - "1ba81c0ff0": "推奨", "cfde55c7b0": "モデルのダウンロードに失敗しました。", "43fd4f454b": "音声モデル", "ba4a900d1d": "ディクテーションモード", @@ -8794,10 +8779,6 @@ "506df81ba6": "OpenAI API キーが保存されました", "91980ce124": "{{value0}} MB", "61a16c8141": "抽出中…", - "8f4d2a51d7": "オフライン", - "d504ab05f0": "ストリーミング", - "fbe5990716": "モデルの選択", - "e24f7d43d2": "音声モデルを選択します。ローカルモデルはオフラインで実行されます。クラウドモデルには API キーが必要です。", "174da92062": "ホールド", "118b3c2dee": "トグル", "ad5d036ecc": "マイクの許可を要求できませんでした。音声ディクテーションが有効になっていませんでした。", @@ -8806,7 +8787,6 @@ "cd9fe37556": "マイクの許可が与えられました", "6fa734ed95": "{{value0}} を削除", "openAiKeyName": "OpenAI 転写キー", - "dictationShortcutDescription": "{{shortcut}} を押すと、フォーカスされているペインにテキストを音声入力できます。", "dictationModeAuto": "自動", "dictationModesWithAutoDescription": "自動:{{shortcut}} を押し続けると離すまで音声入力し、短く押すと開始と停止を切り替えます。ホールド:{{shortcut}} を押している間、音声入力します。トグル:{{shortcut}} を1回押すと開始し、もう一度押すと停止します。", "speechGpuLabel": "GPU を使用", @@ -19981,6 +19961,67 @@ "fillerWords": { "title": "フィラーを削除", "description": "音声入力の言語に合わせて、「えーと」「うーん」などのつなぎ言葉を入力テキストから取り除きます。" + }, + "sectionDescription": "どのペインでも、入力の代わりに話せます。クラウドモデルを選ばない限り、音声はこのコンピューター上で認識されます。", + "input": { + "title": "入力", + "shortcut": "ショートカット", + "shortcutDescription": "フォーカス中のペインで音声入力を開始します。", + "changeShortcut": "変更" + }, + "language": { + "title": "言語" + }, + "cleanup": { + "title": "クリーンアップ" + }, + "advanced": { + "title": "詳細設定" + }, + "setup": { + "downloadingDescription": "このページを離れても大丈夫です。ダウンロードが終わると音声入力が使えます。", + "title": "音声入力を設定", + "pickModel": "どのペインでも、入力の代わりに話せます。下の音声モデルをダウンロードして始めましょう。このコンピューター上のモデルなら音声は外に出ません。", + "microphoneMissingTitle": "{{microphone}} が接続されていません", + "microphoneMissingDescription": "接続されるまで、システムの既定のマイクを使います。", + "changeMicrophone": "マイクを変更", + "onTitle": "音声入力はオンです", + "onLocal": "どのペインでも {{shortcut}} を押します。{{model}} はこのコンピューター上で動作します。", + "onCloud": "どのペインでも {{shortcut}} を押します。{{model}} はクラウドで動作します。", + "offTitle": "音声入力はオフです", + "offDescription": "オンにすると、どのペインでも {{shortcut}} で音声入力できます。", + "savedMicrophone": "選択したマイク" + }, + "quality": { + "unsupported": "お使いの言語に非対応", + "unmeasured": "未測定", + "poor": "低い", + "fair": "まずまず", + "good": "良い", + "veryGood": "とても良い", + "excellent": "非常に優秀", + "characterErrorRate": "文字誤り率 {{percent}}%", + "wordErrorRate": "単語誤り率 {{percent}}%", + "librispeech": "テスト:LibriSpeech(朗読音声)、{{clips}} クリップ、{{date}} に測定。", + "fleurs": "テスト:FLEURS(朗読音声)、{{clips}} クリップ、{{date}} に測定。", + "bench": "Apple シリコン搭載 Mac での Orca のベンチマーク。", + "label": "品質:{{word}}" + }, + "models": { + "retired": "提供終了。削除すると空き容量が増えます。", + "sendsToOpenAi": "音声を OpenAI に送信します。", + "delete": "削除", + "cancelDownload": "ダウンロードをキャンセル", + "inUse": "使用中", + "use": "使用", + "setUpCloud": "設定", + "download": "ダウンロード", + "cloud": "クラウド", + "description": "品質は{{language}}での評価です。評価にカーソルを合わせるとテストの詳細が表示されます。", + "where": "モデルの実行場所", + "local": "このコンピューター", + "cloudDescription": "クラウドモデルは音声をプロバイダーに送信します。このコンピューター上のモデルなら音声は外に出ません。", + "sendsToCompatible": "音声を設定したサービスに送信します。" } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/ko.json b/src/renderer/src/i18n/locales/ko.json index 0f99b624163..66a37aee302 100644 --- a/src/renderer/src/i18n/locales/ko.json +++ b/src/renderer/src/i18n/locales/ko.json @@ -497,19 +497,6 @@ "missingHost": "프록시 URL에는 호스트가 포함되어야 합니다." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "유럽 25개 언어에서 가장 높은 정확도. 문장 부호, 대소문자, 단어별 타임스탬프 지원.", - "parakeetV2": "영어 전용. v3와 정확도는 비슷하며 속도는 더 빠릅니다. 문장 부호와 대소문자 지원.", - "zipformerStreamingEnRetired": "더 이상 제공되지 않습니다. 문장 앞부분이 누락됩니다. 삭제하여 공간을 확보하세요.", - "whisperTiny": "90개 이상 언어 지원. Parakeet보다 정확도는 낮지만 지원 언어가 가장 다양합니다.", - "senseVoice": "중국어, 영어, 일본어, 한국어, 광둥어를 지원하며 언어를 자동으로 감지합니다.", - "gptTranscribe": "정확도가 높은 클라우드 전사. OpenAI API 키가 필요합니다.", - "whisperLargeV3Turbo": "일본어, 한국어, 중국어에서 가장 정확하며 90개 이상의 언어를 지원합니다. Parakeet보다 느립니다.", - "retiredForEngine": "더 이상 제공되지 않음: 음성 엔진에서 더 이상 실행할 수 없습니다. 삭제하여 공간을 확보하세요.", - "nemotron": "영어, 스페인어, 독일어 등 다양한 언어를 지원하며 언어를 자동으로 감지합니다." - } - }, "diagnostics": { "errors": { "creatingDisabled": "검토 파일 생성이 비활성화되어 있습니다", @@ -8236,7 +8223,6 @@ "7807c11c4d": "일반", "6855b0f77d": "Orca를 병렬 에이전트 작업에 유용하게 만드는 핵심 워크플로를 완료합니다.", "6d119427ef": "온보딩 체크리스트", - "eb1176a14e": "온디바이스 모델을 사용한 로컬 음성-텍스트 받아쓰기.", "5063bb47a5": "음성", "7118953f14": "agents가 컴퓨터의 모든 앱을 제어할 수 있도록 하세요.", "c9841721cb": "컴퓨터 사용", @@ -8746,7 +8732,6 @@ }, "VoicePane": { "68de13f72c": "모델을 삭제하지 못했습니다.", - "1ba81c0ff0": "추천", "cfde55c7b0": "모델을 다운로드하지 못했습니다.", "43fd4f454b": "음성 모델", "ba4a900d1d": "받아쓰기 모드", @@ -8757,10 +8742,6 @@ "506df81ba6": "OpenAI API 키가 저장되었습니다.", "91980ce124": "{{value0}}MB", "61a16c8141": "추출 중...", - "8f4d2a51d7": "오프라인", - "d504ab05f0": "스트리밍", - "fbe5990716": "모델 선택", - "e24f7d43d2": "음성 모델을 선택하세요. 로컬 모델은 오프라인으로 실행됩니다. 클라우드 모델에는 API 키가 필요합니다.", "174da92062": "길게 누르기", "118b3c2dee": "토글", "ad5d036ecc": "마이크 권한을 요청할 수 없습니다. 음성 받아쓰기가 활성화되지 않았습니다.", @@ -8769,7 +8750,6 @@ "cd9fe37556": "마이크 권한이 부여되었습니다.", "6fa734ed95": "{{value0}} 삭제", "openAiKeyName": "OpenAI 트랜스크립션 키", - "dictationShortcutDescription": "{{shortcut}} 단축키를 눌러 포커스된 패널에 텍스트를 받아쓰세요.", "dictationModeAuto": "자동", "dictationModesWithAutoDescription": "자동: {{shortcut}} 단축키를 길게 누르면 손을 뗄 때까지 받아쓰기하고, 짧게 누르면 시작하고 중지합니다. 길게 누르기: {{shortcut}} 단축키를 누르고 있는 동안 받아쓰기합니다. 토글: {{shortcut}} 단축키를 한 번 누르면 받아쓰기를 시작하고, 다시 누르면 중지합니다.", "speechGpuLabel": "GPU 사용", @@ -19981,6 +19961,67 @@ "fillerWords": { "title": "군말 제거", "description": "받아쓰기 언어에 맞춰 받아쓴 텍스트에서 “음”, “흠” 같은 소리를 지웁니다." + }, + "sectionDescription": "어느 창에서든 입력 대신 말하세요. 클라우드 모델을 고르지 않으면 음성은 이 컴퓨터에서 인식됩니다.", + "input": { + "title": "입력", + "shortcut": "단축키", + "shortcutDescription": "포커스된 창에서 받아쓰기를 시작합니다.", + "changeShortcut": "변경" + }, + "language": { + "title": "언어" + }, + "cleanup": { + "title": "정리" + }, + "advanced": { + "title": "고급" + }, + "setup": { + "downloadingDescription": "이 페이지를 나가도 됩니다. 다운로드가 끝나면 받아쓰기를 쓸 수 있습니다.", + "title": "받아쓰기 설정", + "pickModel": "어느 창에서든 입력 대신 말하세요. 아래에서 음성 모델을 다운로드해 시작하세요. 이 컴퓨터의 모델은 오디오를 밖으로 보내지 않습니다.", + "microphoneMissingTitle": "{{microphone}}이(가) 연결되어 있지 않습니다", + "microphoneMissingDescription": "다시 연결될 때까지 시스템 기본 마이크를 사용합니다.", + "changeMicrophone": "마이크 변경", + "onTitle": "받아쓰기가 켜져 있습니다", + "onLocal": "어느 창에서든 {{shortcut}}을(를) 누르세요. {{model}}은(는) 이 컴퓨터에서 실행됩니다.", + "onCloud": "어느 창에서든 {{shortcut}}을(를) 누르세요. {{model}}은(는) 클라우드에서 실행됩니다.", + "offTitle": "받아쓰기가 꺼져 있습니다", + "offDescription": "켜면 어느 창에서든 {{shortcut}}으로 받아쓸 수 있습니다.", + "savedMicrophone": "선택한 마이크" + }, + "quality": { + "unsupported": "사용하는 언어를 지원하지 않음", + "unmeasured": "측정 안 됨", + "poor": "낮음", + "fair": "보통", + "good": "좋음", + "veryGood": "매우 좋음", + "excellent": "탁월함", + "characterErrorRate": "문자 오류율 {{percent}}%", + "wordErrorRate": "단어 오류율 {{percent}}%", + "librispeech": "테스트: LibriSpeech(낭독 음성), 클립 {{clips}}개, {{date}} 측정.", + "fleurs": "테스트: FLEURS(낭독 음성), 클립 {{clips}}개, {{date}} 측정.", + "bench": "Apple 실리콘 Mac에서 측정한 Orca 벤치마크.", + "label": "품질: {{word}}" + }, + "models": { + "retired": "더 이상 제공되지 않습니다. 삭제하면 공간을 확보할 수 있습니다.", + "sendsToOpenAi": "오디오를 OpenAI로 보냅니다.", + "delete": "삭제", + "cancelDownload": "다운로드 취소", + "inUse": "사용 중", + "use": "사용", + "setUpCloud": "설정", + "download": "다운로드", + "cloud": "클라우드", + "description": "품질은 {{language}} 기준입니다. 평가에 마우스를 올리면 테스트 내용을 볼 수 있습니다.", + "where": "모델 실행 위치", + "local": "이 컴퓨터", + "cloudDescription": "클라우드 모델은 오디오를 제공업체로 보냅니다. 이 컴퓨터의 모델은 오디오를 밖으로 보내지 않습니다.", + "sendsToCompatible": "오디오를 설정한 서비스로 보냅니다." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/pt.json b/src/renderer/src/i18n/locales/pt.json index d7b928d3ad3..d756f00569c 100644 --- a/src/renderer/src/i18n/locales/pt.json +++ b/src/renderer/src/i18n/locales/pt.json @@ -1717,19 +1717,6 @@ "missingHost": "A URL de proxy deve incluir um host." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Máxima precisão para 25 idiomas europeus. Pontuação, maiúsculas e minúsculas e marcas de tempo por palavra.", - "parakeetV2": "Somente inglês. Mais rápido que v3 com precisão semelhante. Pontuação e maiúsculas e minúsculas.", - "zipformerStreamingEnRetired": "Não é mais oferecido: ele perde o início das frases. Exclua-o para liberar espaço.", - "whisperTiny": "Mais de 90 idiomas. Menor precisão que Parakeet, mas com a maior cobertura de idiomas.", - "senseVoice": "Chinês, inglês, japonês, coreano e cantonês com detecção automática de idioma.", - "gptTranscribe": "Transcrição na nuvem com alta precisão. Exige uma chave de API da OpenAI.", - "whisperLargeV3Turbo": "A melhor precisão em japonês, coreano e chinês, e mais de 90 idiomas. Mais lento que o Parakeet.", - "retiredForEngine": "Não é mais oferecido: o mecanismo de fala não consegue mais executá-lo. Exclua-o para liberar espaço.", - "nemotron": "Inglês, espanhol, alemão e muitos outros idiomas, com detecção automática de idioma." - } - }, "diagnostics": { "errors": { "creatingDisabled": "a criação de arquivos de revisão está desativada", @@ -5147,7 +5134,6 @@ "7807c11c4d": "Geral", "6855b0f77d": "Conclua os fluxos de trabalho essenciais que tornam o Orca útil para trabalho paralelo com agentes.", "6d119427ef": "Checklist de introdução", - "eb1176a14e": "Ditado de voz para texto local com modelos no dispositivo.", "5063bb47a5": "Voz", "7118953f14": "Permita que agentes controlem qualquer aplicativo no seu computador.", "c9841721cb": "Uso do Computador", @@ -6480,7 +6466,6 @@ "VoicePane": { "6fa734ed95": "Excluir {{value0}}", "68de13f72c": "Falha ao excluir o modelo.", - "1ba81c0ff0": "recomendado", "cfde55c7b0": "Falha ao baixar o modelo.", "43fd4f454b": "Modelo de Voz", "ba4a900d1d": "Modo de Ditado", @@ -6491,10 +6476,6 @@ "506df81ba6": "Chave de API da OpenAI salva", "91980ce124": "{{value0}} MB", "61a16c8141": "Extraindo...", - "8f4d2a51d7": "off-line", - "d504ab05f0": "streaming", - "fbe5990716": "Selecionar Modelo", - "e24f7d43d2": "Selecione um modelo de voz. Modelos locais funcionam off-line; modelos na nuvem exigem uma chave de API.", "174da92062": "Manter Pressionado", "118b3c2dee": "Alternar", "ad5d036ecc": "Não foi possível solicitar a permissão do microfone. O ditado por voz não foi ativado.", @@ -6502,7 +6483,6 @@ "1eac933202": "Privacidade e Segurança do macOS aberta. Ative o ditado novamente após conceder o acesso.", "cd9fe37556": "Permissão do microfone concedida", "openAiKeyName": "Sua chave de transcrição da OpenAI", - "dictationShortcutDescription": "Pressione {{shortcut}} para ditar texto em qualquer painel com foco.", "dictationModeAuto": "Automático", "dictationModesWithAutoDescription": "Automático: mantenha {{shortcut}} pressionado para ditar até soltar, ou toque para iniciar e parar. Manter pressionado: dite enquanto mantém {{shortcut}} pressionado. Alternar: pressione {{shortcut}} uma vez para iniciar e novamente para parar.", "speechGpuLabel": "Usar GPU", @@ -21670,6 +21650,67 @@ "fillerWords": { "title": "Remover hesitações", "description": "Remove sons como “hã” e “hum” do texto ditado, conforme o idioma do ditado." + }, + "sectionDescription": "Fale em vez de digitar em qualquer painel. A fala é reconhecida neste computador, a menos que você escolha um modelo na nuvem.", + "input": { + "title": "Entrada", + "shortcut": "Atalho", + "shortcutDescription": "Inicia o ditado no painel em foco.", + "changeShortcut": "Alterar" + }, + "language": { + "title": "Idioma" + }, + "cleanup": { + "title": "Limpeza" + }, + "advanced": { + "title": "Avançado" + }, + "setup": { + "downloadingDescription": "Você pode sair desta página. O ditado funciona quando o download terminar.", + "title": "Configurar o ditado", + "pickModel": "Fale em vez de digitar em qualquer painel. Baixe um modelo de voz abaixo para começar; modelos neste computador mantêm seu áudio aqui.", + "microphoneMissingTitle": "{{microphone}} não está conectado", + "microphoneMissingDescription": "O ditado usa o microfone padrão do sistema até ele voltar.", + "changeMicrophone": "Trocar microfone", + "onTitle": "O ditado está ativado", + "onLocal": "Pressione {{shortcut}} em qualquer painel. {{model}} roda neste computador.", + "onCloud": "Pressione {{shortcut}} em qualquer painel. {{model}} roda na nuvem.", + "offTitle": "O ditado está desativado", + "offDescription": "Ative para ditar com {{shortcut}} em qualquer painel.", + "savedMicrophone": "Seu microfone" + }, + "quality": { + "unsupported": "Não suporta seu idioma", + "unmeasured": "Não medido", + "poor": "Ruim", + "fair": "Razoável", + "good": "Boa", + "veryGood": "Muito boa", + "excellent": "Excelente", + "characterErrorRate": "{{percent}}% de erro por caractere", + "wordErrorRate": "{{percent}}% de erro por palavra", + "librispeech": "Teste: LibriSpeech (fala lida), {{clips}} clipes, medido em {{date}}.", + "fleurs": "Teste: FLEURS (fala lida), {{clips}} clipes, medido em {{date}}.", + "bench": "Benchmark do Orca em um Mac com Apple silicon.", + "label": "Qualidade: {{word}}" + }, + "models": { + "retired": "Não é mais oferecido. Exclua para liberar espaço.", + "sendsToOpenAi": "Envia seu áudio para a OpenAI.", + "delete": "Excluir", + "cancelDownload": "Cancelar download", + "inUse": "Em uso", + "use": "Usar", + "setUpCloud": "Configurar", + "download": "Baixar", + "cloud": "Nuvem", + "description": "A qualidade é avaliada para {{language}}. Passe o cursor sobre uma avaliação para ver o teste.", + "where": "Onde os modelos rodam", + "local": "Neste computador", + "cloudDescription": "Modelos na nuvem enviam seu áudio ao provedor. Os deste computador o mantêm aqui.", + "sendsToCompatible": "Envia seu áudio para o serviço que você configurou." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/ru.json b/src/renderer/src/i18n/locales/ru.json index 466dfb0eaf1..f54ffcfda1d 100644 --- a/src/renderer/src/i18n/locales/ru.json +++ b/src/renderer/src/i18n/locales/ru.json @@ -1014,19 +1014,6 @@ "missingHost": "URL прокси должен содержать адрес сервера." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "Наивысшая точность для 25 европейских языков. Пунктуация, регистр букв и временные метки для каждого слова.", - "parakeetV2": "Только английский. Быстрее v3 при схожей точности. Пунктуация и регистр букв.", - "zipformerStreamingEnRetired": "Больше не предлагается: теряет начало фраз. Удалите модель, чтобы освободить место.", - "whisperTiny": "Более 90 языков. Точность ниже, чем у Parakeet, но самый широкий охват языков.", - "senseVoice": "Китайский, английский, японский, корейский и кантонский с автоматическим определением языка.", - "gptTranscribe": "Облачная расшифровка с высокой точностью. Требуется API-ключ OpenAI.", - "whisperLargeV3Turbo": "Лучшая точность для японского, корейского и китайского, а также более 90 языков. Медленнее, чем Parakeet.", - "retiredForEngine": "Больше не предлагается: движок распознавания речи больше не может её запускать. Удалите её, чтобы освободить место.", - "nemotron": "Английский, испанский, немецкий и многие другие языки, с автоматическим определением языка." - } - }, "diagnostics": { "errors": { "creatingDisabled": "создание файлов для проверки отключено", @@ -9298,7 +9285,6 @@ "7807c11c4d": "Общие", "6855b0f77d": "Завершите настройку основных процессов Orca для параллельной работы агентов.", "6d119427ef": "Знакомство с Orca", - "eb1176a14e": "Локальная голосовая диктовка с моделями на устройстве.", "5063bb47a5": "Голос", "7118953f14": "Разрешите агентам управлять любыми приложениями на компьютере.", "c9841721cb": "Управление компьютером", @@ -9813,7 +9799,6 @@ "VoicePane": { "6fa734ed95": "Удалить {{value0}}", "68de13f72c": "Не удалось удалить модель.", - "1ba81c0ff0": "рекомендуется", "cfde55c7b0": "Не удалось скачать модель.", "43fd4f454b": "Речевая модель", "ba4a900d1d": "Режим диктовки", @@ -9824,10 +9809,6 @@ "506df81ba6": "API-ключ OpenAI сохранён", "91980ce124": "{{value0}} МБ", "61a16c8141": "Распаковка…", - "8f4d2a51d7": "без сети", - "d504ab05f0": "потоковая обработка", - "fbe5990716": "Выберите модель", - "e24f7d43d2": "Выберите речевую модель. Локальные модели работают без сети; облачным нужен API-ключ.", "174da92062": "Удержание", "118b3c2dee": "Переключение", "ad5d036ecc": "Не удалось запросить доступ к микрофону. Голосовая диктовка не включена.", @@ -9835,7 +9816,6 @@ "1eac933202": "Открыт раздел macOS «Конфиденциальность и безопасность». После выдачи доступа включите диктовку снова.", "cd9fe37556": "Доступ к микрофону разрешён", "openAiKeyName": "Ваш ключ расшифровки OpenAI", - "dictationShortcutDescription": "Нажмите {{shortcut}}, чтобы диктовать текст в любую активную панель.", "dictationModeAuto": "Авто", "dictationModesWithAutoDescription": "Авто: удерживайте {{shortcut}}, чтобы диктовать до отпускания, или нажмите коротко, чтобы начать и остановить. Удержание: диктуйте, пока удерживаете {{shortcut}}. Переключение: нажмите {{shortcut}} для начала и ещё раз для остановки.", "speechGpuLabel": "Использовать GPU", @@ -21387,6 +21367,67 @@ "fillerWords": { "title": "Удалять слова-паразиты", "description": "Убирает звуки вроде «э» и «эм» из надиктованного текста с учётом языка диктовки." + }, + "sectionDescription": "Говорите вместо того, чтобы печатать, в любой панели. Речь распознаётся на этом компьютере, если вы не выберете облачную модель.", + "input": { + "title": "Ввод", + "shortcut": "Сочетание клавиш", + "shortcutDescription": "Запускает диктовку в активной панели.", + "changeShortcut": "Изменить" + }, + "language": { + "title": "Язык" + }, + "cleanup": { + "title": "Очистка" + }, + "advanced": { + "title": "Дополнительно" + }, + "setup": { + "downloadingDescription": "Можно уйти с этой страницы. Диктовка заработает, когда загрузка закончится.", + "title": "Настроить диктовку", + "pickModel": "Говорите вместо того, чтобы печатать, в любой панели. Скачайте модель речи ниже, чтобы начать; модели на этом компьютере не отправляют звук наружу.", + "microphoneMissingTitle": "{{microphone}} не подключён", + "microphoneMissingDescription": "Пока он не вернётся, диктовка использует системный микрофон по умолчанию.", + "changeMicrophone": "Сменить микрофон", + "onTitle": "Диктовка включена", + "onLocal": "Нажмите {{shortcut}} в любой панели. {{model}} работает на этом компьютере.", + "onCloud": "Нажмите {{shortcut}} в любой панели. {{model}} работает в облаке.", + "offTitle": "Диктовка выключена", + "offDescription": "Включите, чтобы диктовать с помощью {{shortcut}} в любой панели.", + "savedMicrophone": "Ваш микрофон" + }, + "quality": { + "unsupported": "Не поддерживает ваш язык", + "unmeasured": "Не измерено", + "poor": "Плохо", + "fair": "Средне", + "good": "Хорошо", + "veryGood": "Очень хорошо", + "excellent": "Отлично", + "characterErrorRate": "{{percent}}% ошибок в символах", + "wordErrorRate": "{{percent}}% ошибок в словах", + "librispeech": "Тест: LibriSpeech (чтение вслух), клипов: {{clips}}, измерено {{date}}.", + "fleurs": "Тест: FLEURS (чтение вслух), клипов: {{clips}}, измерено {{date}}.", + "bench": "Бенчмарк Orca на Mac с Apple silicon.", + "label": "Качество: {{word}}" + }, + "models": { + "retired": "Больше не предлагается. Удалите, чтобы освободить место.", + "sendsToOpenAi": "Отправляет звук в OpenAI.", + "delete": "Удалить", + "cancelDownload": "Отменить загрузку", + "inUse": "Используется", + "use": "Использовать", + "setUpCloud": "Настроить", + "download": "Скачать", + "cloud": "Облако", + "description": "Качество оценено для языка: {{language}}. Наведите на оценку, чтобы увидеть тест.", + "where": "Где работают модели", + "local": "На этом компьютере", + "cloudDescription": "Облачные модели отправляют звук провайдеру. Модели на этом компьютере оставляют его здесь.", + "sendsToCompatible": "Отправляет звук в настроенный вами сервис." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/tr.json b/src/renderer/src/i18n/locales/tr.json index 3b7a679fe24..91f2984beb7 100644 --- a/src/renderer/src/i18n/locales/tr.json +++ b/src/renderer/src/i18n/locales/tr.json @@ -948,19 +948,6 @@ "missingHost": "Proxy URL'si bir ana makine içermelidir." } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "25 Avrupa dili için en yüksek doğruluk. Noktalama, büyük/küçük harf kullanımı ve sözcük düzeyinde zaman damgaları.", - "parakeetV2": "Yalnızca İngilizce. Benzer doğrulukla v3'ten daha hızlı. Noktalama ve büyük/küçük harf kullanımı.", - "zipformerStreamingEnRetired": "Artık sunulmuyor: cümlelerin başını atlıyor. Yer açmak için silin.", - "whisperTiny": "90'dan fazla dil. Parakeet'ten daha düşük doğruluk, ancak en geniş dil kapsamı.", - "senseVoice": "Otomatik dil algılamayla Çince, İngilizce, Japonca, Korece ve Kantonca.", - "gptTranscribe": "Yüksek doğrulukla bulut transkripsiyonu. OpenAI API anahtarı gerektirir.", - "whisperLargeV3Turbo": "Japonca, Korece ve Çince için en iyi doğruluk, ayrıca 90'dan fazla dil. Parakeet'ten daha yavaş.", - "retiredForEngine": "Artık sunulmuyor: konuşma motoru bu modeli artık çalıştıramıyor. Yer açmak için silin.", - "nemotron": "İngilizce, İspanyolca, Almanca ve daha birçok dil; dil otomatik algılanır." - } - }, "diagnostics": { "errors": { "creatingDisabled": "inceleme dosyası oluşturma devre dışı", @@ -8930,7 +8917,6 @@ "7807c11c4d": "Genel", "6855b0f77d": "Orca'yı paralel ajan çalışması için kullanışlı kılan temel iş akışlarını tamamlayın.", "6d119427ef": "Başlangıç kontrol listesi", - "eb1176a14e": "Cihaz üzerindeki modellerle yerel sesten metne dikte.", "5063bb47a5": "Ses", "7118953f14": "Ajanların bilgisayarınızdaki herhangi bir uygulamayı kontrol etmesini sağlayın.", "c9841721cb": "Bilgisayar Kullanımı", @@ -9445,7 +9431,6 @@ "VoicePane": { "6fa734ed95": "{{value0}} silinsin mi", "68de13f72c": "Model silinemedi.", - "1ba81c0ff0": "önerilen", "cfde55c7b0": "Model indirilemedi.", "43fd4f454b": "Konuşma Modeli", "ba4a900d1d": "Dikte Modu", @@ -9456,10 +9441,6 @@ "506df81ba6": "OpenAI API anahtarı kaydedildi", "91980ce124": "{{value0}} MB", "61a16c8141": "Çıkarılıyor...", - "8f4d2a51d7": "çevrimdışı", - "d504ab05f0": "akış", - "fbe5990716": "Model Seç", - "e24f7d43d2": "Bir konuşma modeli seçin. Yerel modeller çevrimdışı çalışır; bulut modelleri bir API anahtarı gerektirir.", "174da92062": "Basılı Tut", "118b3c2dee": "Aç/kapat", "ad5d036ecc": "Mikrofon izni istenemedi. Sesli dikte etkinleştirilmedi.", @@ -9467,7 +9448,6 @@ "1eac933202": "macOS Gizlilik ve Güvenlik açıldı. Erişimi verdikten sonra dikteyi tekrar etkinleştirin.", "cd9fe37556": "Mikrofon izni verildi", "openAiKeyName": "OpenAI transkripsiyon anahtarınız", - "dictationShortcutDescription": "Odaktaki herhangi bir bölmeye metin dikte etmek için {{shortcut}} tuşlarına basın.", "dictationModeAuto": "Otomatik", "dictationModesWithAutoDescription": "Otomatik: bırakana kadar dikte etmek için {{shortcut}} tuşlarını basılı tutun veya başlatıp durdurmak için kısaca dokunun. Basılı tut: {{shortcut}} basılıyken dikte edin. Aç/kapat: başlatmak için {{shortcut}} tuşlarına bir kez, durdurmak için tekrar basın.", "speechGpuLabel": "GPU kullan", @@ -20682,6 +20662,67 @@ "fillerWords": { "title": "Dolgu seslerini kaldır", "description": "Dikte edilen metinden “hmm” gibi duraksama seslerini dikte diline göre çıkarır." + }, + "sectionDescription": "Her bölmede yazmak yerine konuşun. Bulut modeli seçmediğiniz sürece konuşma bu bilgisayarda tanınır.", + "input": { + "title": "Giriş", + "shortcut": "Kısayol", + "shortcutDescription": "Odaktaki bölmede dikteyi başlatır.", + "changeShortcut": "Değiştir" + }, + "language": { + "title": "Dil" + }, + "cleanup": { + "title": "Temizleme" + }, + "advanced": { + "title": "Gelişmiş" + }, + "setup": { + "downloadingDescription": "Bu sayfadan ayrılabilirsiniz. İndirme bitince dikte çalışır.", + "title": "Dikteyi kur", + "pickModel": "Her bölmede yazmak yerine konuşun. Başlamak için aşağıdan bir konuşma modeli indirin; bu bilgisayardaki modeller sesinizi burada tutar.", + "microphoneMissingTitle": "{{microphone}} bağlı değil", + "microphoneMissingDescription": "Geri gelene kadar dikte sistemin varsayılan mikrofonunu kullanır.", + "changeMicrophone": "Mikrofonu değiştir", + "onTitle": "Dikte açık", + "onLocal": "Herhangi bir bölmede {{shortcut}} tuşuna basın. {{model}} bu bilgisayarda çalışır.", + "onCloud": "Herhangi bir bölmede {{shortcut}} tuşuna basın. {{model}} bulutta çalışır.", + "offTitle": "Dikte kapalı", + "offDescription": "Her bölmede {{shortcut}} ile dikte etmek için açın.", + "savedMicrophone": "Mikrofonunuz" + }, + "quality": { + "unsupported": "Dilinizi desteklemiyor", + "unmeasured": "Ölçülmedi", + "poor": "Zayıf", + "fair": "Orta", + "good": "İyi", + "veryGood": "Çok iyi", + "excellent": "Mükemmel", + "characterErrorRate": "%{{percent}} karakter hata oranı", + "wordErrorRate": "%{{percent}} kelime hata oranı", + "librispeech": "Test: LibriSpeech (sesli okuma), {{clips}} klip, ölçüm tarihi {{date}}.", + "fleurs": "Test: FLEURS (sesli okuma), {{clips}} klip, ölçüm tarihi {{date}}.", + "bench": "Orca'nın Apple silikonlu bir Mac'teki kıyaslaması.", + "label": "Kalite: {{word}}" + }, + "models": { + "retired": "Artık sunulmuyor. Yer açmak için silin.", + "sendsToOpenAi": "Sesinizi OpenAI'ye gönderir.", + "delete": "Sil", + "cancelDownload": "İndirmeyi iptal et", + "inUse": "Kullanımda", + "use": "Kullan", + "setUpCloud": "Kur", + "download": "İndir", + "cloud": "Bulut", + "description": "Kalite {{language}} için değerlendirildi. Testi görmek için bir puanın üzerine gelin.", + "where": "Modellerin çalıştığı yer", + "local": "Bu bilgisayarda", + "cloudDescription": "Bulut modelleri sesinizi sağlayıcıya gönderir. Bu bilgisayardaki modeller sesi burada tutar.", + "sendsToCompatible": "Sesinizi kurduğunuz hizmete gönderir." } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/zh-Hant.json b/src/renderer/src/i18n/locales/zh-Hant.json index 51bfeb5d8b5..46897c5cfd5 100644 --- a/src/renderer/src/i18n/locales/zh-Hant.json +++ b/src/renderer/src/i18n/locales/zh-Hant.json @@ -921,19 +921,6 @@ "missingHost": "代理 URL 必須包含主機。" } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "為 25 種歐洲語言提供最高準確度。支援標點符號、大小寫及詞級時間戳記。", - "parakeetV2": "僅限英文。準確度與 v3 相近,速度更快。支援標點符號及大小寫。", - "zipformerStreamingEnRetired": "已不再提供:它會遺漏句子開頭。刪除它以釋放空間。", - "whisperTiny": "支援 90 多種語言。準確度低於 Parakeet,但語言涵蓋範圍最廣。", - "senseVoice": "支援中文、英文、日文、韓文及粵語,並自動偵測語言。", - "gptTranscribe": "準確度高的雲端轉錄。需要 OpenAI API 金鑰。", - "whisperLargeV3Turbo": "日語、韓語和中文準確度最高,支援 90 多種語言。速度比 Parakeet 慢。", - "retiredForEngine": "已不再提供:語音引擎已無法執行此模型。刪除它以釋放空間。", - "nemotron": "支援英語、西班牙語、德語等多種語言,可自動辨識語言。" - } - }, "diagnostics": { "errors": { "creatingDisabled": "已停用建立檢視檔案", @@ -8767,7 +8754,6 @@ "7807c11c4d": "一般", "6855b0f77d": "完成核心工作流程,讓 Orca 有效協助代理程式並行工作。", "6d119427ef": "入門清單", - "eb1176a14e": "使用裝置上模型進行本機語音到文字聽寫。", "5063bb47a5": "語音", "7118953f14": "使代理程式能夠控制您電腦上的任何應用程式。", "c9841721cb": "電腦控制", @@ -9282,7 +9268,6 @@ "VoicePane": { "6fa734ed95": "刪除 {{value0}}", "68de13f72c": "刪除模型失敗。", - "1ba81c0ff0": "推薦", "cfde55c7b0": "下載模型失敗。", "43fd4f454b": "語音模型", "ba4a900d1d": "聽寫模式", @@ -9293,10 +9278,6 @@ "506df81ba6": "OpenAI API 金鑰已儲存", "91980ce124": "{{value0}} MB", "61a16c8141": "正在解壓縮...", - "8f4d2a51d7": "離線", - "d504ab05f0": "串流", - "fbe5990716": "選擇模型", - "e24f7d43d2": "選擇語音模型。本機模型離線執行;雲端模型需要 API 金鑰。", "174da92062": "按住", "118b3c2dee": "切換", "ad5d036ecc": "無法請求麥克風權限。未啟用語音聽寫。", @@ -9304,7 +9285,6 @@ "1eac933202": "已開啟 macOS 的「隱私權與安全性」。授予存取權限後再次啟用聽寫。", "cd9fe37556": "已授予麥克風權限", "openAiKeyName": "您的 OpenAI 轉錄金鑰", - "dictationShortcutDescription": "按 {{shortcut}} 可將文字聽寫到目前取得焦點的任意窗格中。", "dictationModeAuto": "自動", "dictationModesWithAutoDescription": "自動:按住 {{shortcut}} 聽寫,放開即停止;輕按則開始或停止。按住:按住 {{shortcut}} 期間進行聽寫。切換:按一次 {{shortcut}} 開始,再按一次停止。", "speechGpuLabel": "使用 GPU", @@ -20384,6 +20364,67 @@ "fillerWords": { "title": "移除語氣詞", "description": "依聽寫語言,從聽寫文字中移除「嗯」「呃」之類的語氣詞。" + }, + "sectionDescription": "在任何面板中用說話取代打字。除非你選擇雲端模型,否則語音都在這台電腦上辨識。", + "input": { + "title": "輸入", + "shortcut": "快捷鍵", + "shortcutDescription": "在目前聚焦的面板中開始聽寫。", + "changeShortcut": "變更" + }, + "language": { + "title": "語言" + }, + "cleanup": { + "title": "清理" + }, + "advanced": { + "title": "進階" + }, + "setup": { + "downloadingDescription": "你可以離開此頁面。下載完成後即可使用聽寫。", + "title": "設定聽寫", + "pickModel": "在任何面板中用說話取代打字。先在下方下載一個語音模型;本機模型會把你的音訊留在這台電腦上。", + "microphoneMissingTitle": "{{microphone}} 未連接", + "microphoneMissingDescription": "在它重新連接前,聽寫會使用系統預設麥克風。", + "changeMicrophone": "更換麥克風", + "onTitle": "聽寫已開啟", + "onLocal": "在任何面板中按 {{shortcut}}。{{model}} 在這台電腦上執行。", + "onCloud": "在任何面板中按 {{shortcut}}。{{model}} 在雲端執行。", + "offTitle": "聽寫已關閉", + "offDescription": "開啟後,可在任何面板中按 {{shortcut}} 聽寫。", + "savedMicrophone": "你的麥克風" + }, + "quality": { + "unsupported": "不支援你的語言", + "unmeasured": "未測量", + "poor": "差", + "fair": "普通", + "good": "良好", + "veryGood": "很好", + "excellent": "優秀", + "characterErrorRate": "字元錯誤率 {{percent}}%", + "wordErrorRate": "詞錯誤率 {{percent}}%", + "librispeech": "測試:LibriSpeech(朗讀語音),{{clips}} 段音訊,測於 {{date}}。", + "fleurs": "測試:FLEURS(朗讀語音),{{clips}} 段音訊,測於 {{date}}。", + "bench": "Orca 在 Apple 晶片 Mac 上的基準測試。", + "label": "品質:{{word}}" + }, + "models": { + "retired": "已不再提供。刪除即可釋放空間。", + "sendsToOpenAi": "會把你的音訊傳送給 OpenAI。", + "delete": "刪除", + "cancelDownload": "取消下載", + "inUse": "使用中", + "use": "使用", + "setUpCloud": "設定", + "download": "下載", + "cloud": "雲端", + "description": "品質依{{language}}評定。將指標停在評分上可查看所用測試。", + "where": "模型執行位置", + "local": "這台電腦", + "cloudDescription": "雲端模型會把你的音訊傳送給服務商。本機模型會把音訊留在這台電腦上。", + "sendsToCompatible": "會把你的音訊傳送到你設定的服務。" } }, "repositoryHooks": { diff --git a/src/renderer/src/i18n/locales/zh.json b/src/renderer/src/i18n/locales/zh.json index 48c96ba1076..067d098593f 100644 --- a/src/renderer/src/i18n/locales/zh.json +++ b/src/renderer/src/i18n/locales/zh.json @@ -497,19 +497,6 @@ "missingHost": "代理 URL 必须包含主机。" } }, - "voice": { - "modelDescriptions": { - "parakeetV3": "为 25 种欧洲语言提供最高准确度。支持标点、大小写和词级时间戳。", - "parakeetV2": "仅限英语。准确度与 v3 相近,速度更快。支持标点和大小写。", - "zipformerStreamingEnRetired": "已不再提供:它会丢失句子开头。删除它以释放空间。", - "whisperTiny": "支持 90 多种语言。准确度低于 Parakeet,但语言覆盖最广。", - "senseVoice": "支持中文、英语、日语、韩语和粤语,并自动检测语言。", - "gptTranscribe": "准确度高的云端转录。需要 OpenAI API 密钥。", - "whisperLargeV3Turbo": "日语、韩语和中文准确度最高,支持 90 多种语言。速度比 Parakeet 慢。", - "retiredForEngine": "已不再提供:语音引擎已无法运行此模型。删除它以释放空间。", - "nemotron": "支持英语、西班牙语、德语等多种语言,可自动识别语言。" - } - }, "diagnostics": { "errors": { "creatingDisabled": "已禁用创建查看文件", @@ -8236,7 +8223,6 @@ "7807c11c4d": "通用", "6855b0f77d": "完成使 Orca 对于并行智能体工作有用的核心工作流程。", "6d119427ef": "入门清单", - "eb1176a14e": "使用设备上模型进行本地语音到文本听写。", "5063bb47a5": "语音", "7118953f14": "使智能体能够控制您计算机上的任何应用程序。", "c9841721cb": "计算机控制", @@ -8746,7 +8732,6 @@ }, "VoicePane": { "68de13f72c": "删除模型失败。", - "1ba81c0ff0": "推荐", "cfde55c7b0": "下载模型失败。", "43fd4f454b": "语音模型", "ba4a900d1d": "听写模式", @@ -8757,10 +8742,6 @@ "506df81ba6": "OpenAI API 密钥已保存", "91980ce124": "{{value0}} MB", "61a16c8141": "正在提取...", - "8f4d2a51d7": "离线", - "d504ab05f0": "流媒体", - "fbe5990716": "选择型号", - "e24f7d43d2": "选择语音模型,本地模型离线运行;云模型需要 API 密钥。", "174da92062": "按住", "118b3c2dee": "切换", "ad5d036ecc": "无法请求麦克风权限。未启用语音听写。", @@ -8769,7 +8750,6 @@ "cd9fe37556": "已授予麦克风权限", "6fa734ed95": "删除 {{value0}}", "openAiKeyName": "您的 OpenAI 转录密钥", - "dictationShortcutDescription": "按 {{shortcut}} 可将文本听写到当前聚焦的任意窗格中。", "dictationModeAuto": "自动", "dictationModesWithAutoDescription": "自动:按住 {{shortcut}} 听写,松开即停止;轻按则开始或停止。按住:按住 {{shortcut}} 期间进行听写。切换:按一次 {{shortcut}} 开始,再按一次停止。", "speechGpuLabel": "使用 GPU", @@ -19981,6 +19961,67 @@ "fillerWords": { "title": "移除语气词", "description": "按听写语言,从听写文本中去掉“嗯”“呃”之类的语气词。" + }, + "sectionDescription": "在任意面板中用说话代替打字。除非你选择云端模型,否则语音都在这台电脑上识别。", + "input": { + "title": "输入", + "shortcut": "快捷键", + "shortcutDescription": "在当前聚焦的面板中开始听写。", + "changeShortcut": "更改" + }, + "language": { + "title": "语言" + }, + "cleanup": { + "title": "清理" + }, + "advanced": { + "title": "高级" + }, + "setup": { + "downloadingDescription": "你可以离开此页面。下载完成后即可使用听写。", + "title": "设置听写", + "pickModel": "在任意面板中用说话代替打字。先在下方下载一个语音模型;本机模型会把你的音频留在这台电脑上。", + "microphoneMissingTitle": "{{microphone}} 未连接", + "microphoneMissingDescription": "在它重新连接前,听写会使用系统默认麦克风。", + "changeMicrophone": "更换麦克风", + "onTitle": "听写已开启", + "onLocal": "在任意面板中按 {{shortcut}}。{{model}} 在这台电脑上运行。", + "onCloud": "在任意面板中按 {{shortcut}}。{{model}} 在云端运行。", + "offTitle": "听写已关闭", + "offDescription": "开启后,可在任意面板中按 {{shortcut}} 听写。", + "savedMicrophone": "你的麦克风" + }, + "quality": { + "unsupported": "不支持你的语言", + "unmeasured": "未测量", + "poor": "差", + "fair": "一般", + "good": "良好", + "veryGood": "很好", + "excellent": "优秀", + "characterErrorRate": "字符错误率 {{percent}}%", + "wordErrorRate": "词错误率 {{percent}}%", + "librispeech": "测试:LibriSpeech(朗读语音),{{clips}} 段音频,测于 {{date}}。", + "fleurs": "测试:FLEURS(朗读语音),{{clips}} 段音频,测于 {{date}}。", + "bench": "Orca 在 Apple 芯片 Mac 上的基准测试。", + "label": "质量:{{word}}" + }, + "models": { + "retired": "已不再提供。删除即可释放空间。", + "sendsToOpenAi": "会把你的音频发送给 OpenAI。", + "delete": "删除", + "cancelDownload": "取消下载", + "inUse": "使用中", + "use": "使用", + "setUpCloud": "设置", + "download": "下载", + "cloud": "云端", + "description": "质量按{{language}}评定。将指针悬停在评分上可查看所用测试。", + "where": "模型运行位置", + "local": "这台电脑", + "cloudDescription": "云端模型会把你的音频发送给服务商。本机模型会把音频留在这台电脑上。", + "sendsToCompatible": "会把你的音频发送到你设置的服务。" } }, "repositoryHooks": { diff --git a/src/shared/speech-types.ts b/src/shared/speech-types.ts index 252e211284b..d30efb2dca5 100644 --- a/src/shared/speech-types.ts +++ b/src/shared/speech-types.ts @@ -29,7 +29,6 @@ export type SpeechModelErrorRate = { export type SpeechModelManifest = { id: string label: string - description: string type: SpeechModelType provider: SpeechModelProvider /** The text pipeline's hint: one language, a bilingual pair such as 'zh-en', or 'multilingual'. */