{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-09-20T13:02:28.979Z","headline":"Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写","description":"Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","url":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","mainEntityOfPage":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","datePublished":"2026-09-02T00:23:26.000Z","dateModified":"2026-09-02T00:23:26.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://www.ithome.com/0/997/218.htm","https://aihot.virxact.com/items/cmtjczoyu03akrobv35uivsvg"],"canonicalUrl":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","directAnswer":{"@type":"Answer","text":"Meta推出Muse Voice Transcribe，面向实时音频处理提供边说边转写能力，并在同一流程中处理说话人区分与语音结束判断，开发者可通过Meta Model API使用该模型。","url":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","dateCreated":"2026-09-02T00:23:26.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"IT之家 source article","url":"https://www.ithome.com/0/997/218.htm","datePublished":"2026-09-02T00:23:26.000Z","provider":{"@type":"Organization","name":"IT之家","url":"https://www.ithome.com/0/997/218.htm"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmtjczoyu03akrobv35uivsvg","datePublished":"2026-09-02T00:23:26.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmtjczoyu03akrobv35uivsvg"}}],"aggregationSource":"IT之家（RSS）","originalPublisher":{"name":"IT之家","url":"https://www.ithome.com/0/997/218.htm"},"geoDeepAnswer":null,"article":{"id":"cmtjczoyu03akrobv35uivsvg","slug":"cmtjczoyu03akrobv35uivsvg","url":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","title":"Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写","title_en":"","summary":"Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","source":"IT之家（RSS）","sourceUrl":"https://www.ithome.com/0/997/218.htm","aiHotUrl":"https://aihot.virxact.com/items/cmtjczoyu03akrobv35uivsvg","publishedAt":"2026-09-02T00:23:26.000Z","category":"行业动态","score":58,"selected":false,"articleBody":["IT之家：https://www.ithome.com/ 9 月 2 日消息，科技媒体 9to5Mac 昨日（9 月 1 日）发布博文，报道称 Meta 公司推出 Muse Voice Transcribe， 这是其首个实时音频感知模型。","开发者可通过 Meta Model API 调用该能力，定价为每 1000 分钟音频 3 美元 （现汇率约合 20.2 元人民币） ，等同每小时 0.18 美元 （现汇率约合 1.2 元人民币） 。","该模型在同一流程中整合流式自动语音识别、说话人分离与端点检测，用户说话时，系统可同步输出文字，无需等待完整录音结束后再单独处理。","IT之家注：“流式”就是边说边转写，模型不必等整段音频说完才出结果，而是会一小段一小段持续输出文字，所以延迟更低，适合实时听写、会议记录、通话字幕。","Muse Voice Transcribe 可在包含 20 余名说话者的录音中分离不同发言者，还能识别说话是否结束，从而自动确定一段输入的边界。","模型训练覆盖 70 余种语言，其中 25 种语言在发布时完成验证。该模型支持长度超过 1 小时的音频，也支持句内或句间的自然语言切换。开发者还可通过语言、关键词和上下文偏置，提高特定语言、术语或场景下的识别效果。","在技术方面，Meta 为兼顾实时响应和识别准确度，引入名为自适应延迟（adaptive delay）的动态决策机制。系统不会为所有词语采用固定的速度与准确度取舍，而是针对每个词判断需要额外接收多少音频后再输出识别结果。","对于较容易识别的语音，系统可更快提交转写结果。对于发音不清、术语较多或语境复杂的词语，系统会调用更多前后文音频信息。这种机制旨在兼顾。","Meta 称，截至 2026 年 9 月 1 日，Muse Voice Transcribe 位列 Artificial Analysis 流式语音转文本排行榜第 1 名。"],"articleImages":[{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/9/d4d2161a-3d95-4415-80a2-e6862f384d0f.png","alt":"","afterParagraph":1,"url":"/media/articles/cmtjczoyu03akrobv35uivsvg/ee554e84c3ddce3a.png"},{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/9/0e79a9bb-dd1c-4cbc-8bdb-51c587d8cc52.png","alt":"","afterParagraph":8,"url":"/media/articles/cmtjczoyu03akrobv35uivsvg/42354ad04d2f0a9d.png"},{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/9/c58580aa-c5e0-4f20-a47a-1355032532e7.png","alt":"","afterParagraph":8,"url":"/media/articles/cmtjczoyu03akrobv35uivsvg/54ebce27c0f8291f.png"},{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/9/61d8a706-7d1c-4ded-b79e-823709e021bc.png","alt":"","afterParagraph":8,"url":"/media/articles/cmtjczoyu03akrobv35uivsvg/f00e175fe710237c.png"},{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/9/0fa90faf-8e79-4e0a-9d86-5c4b54e404b3.png","alt":"","afterParagraph":8,"url":"/media/articles/cmtjczoyu03akrobv35uivsvg/e051fd016dd234d6.png"}],"mediaStatus":"ok","articleBodyZh":["Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写 这条更新来自 ithome.com，发布时间为 2026-09-02，Aioga 保留原文入口以便核验。","摘要：Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","背景：IT之家援引9to5Mac称，Muse Voice Transcribe支持20余名说话者的录音分离，支持超过1小时音频，训练覆盖70余种语言，其中25种语言在发布时完成验证，定价为每1000分钟3美元。","Aioga 观察：Aioga判断：该模型把实时转写、多人发言区分和输入边界判断组合在一起，显示出连续音频处理的整合方向。材料未提供实际延迟、准确率或各语言表现数据，自适应延迟机制的效果仍需验证。","影响与后续：可能影响：开发者需要结合调用价格、已验证语言范围、录音人数和术语场景评估适配性。支持持续输出和多人分离不代表所有音频都能稳定识别，实际效果仍可能受语音清晰度与上下文影响。 后续观察：应关注Meta Model API的实际开放范围，以及不同语言、多人录音、自然语言切换和专业术语场景中的识别表现，并核验自适应延迟对实时性与准确度的影响。"],"translationStatus":"","bodyOrigin":"source-page","editorial":{"summary":"Meta推出Muse Voice Transcribe，面向实时音频处理提供边说边转写能力，并在同一流程中处理说话人区分与语音结束判断，开发者可通过Meta Model API使用该模型。","background":"IT之家援引9to5Mac称，Muse Voice Transcribe支持20余名说话者的录音分离，支持超过1小时音频，训练覆盖70余种语言，其中25种语言在发布时完成验证，定价为每1000分钟3美元。","viewpoint":"Aioga判断：该模型把实时转写、多人发言区分和输入边界判断组合在一起，显示出连续音频处理的整合方向。材料未提供实际延迟、准确率或各语言表现数据，自适应延迟机制的效果仍需验证。","implications":"可能影响：开发者需要结合调用价格、已验证语言范围、录音人数和术语场景评估适配性。支持持续输出和多人分离不代表所有音频都能稳定识别，实际效果仍可能受语音清晰度与上下文影响。","nextStep":"后续观察：应关注Meta Model API的实际开放范围，以及不同语言、多人录音、自然语言切换和专业术语场景中的识别表现，并核验自适应延迟对实时性与准确度的影响。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-09-02T01:07:26.972Z","sourceHash":"36a4fade3f5bf072","review":{"approved":true,"groundedness":96,"clarity":94,"duplicationRisk":12,"blockingIssues":[],"notes":["“显示出连续音频处理的整合方向”属于基于材料的分析判断，已明确标注为 Aioga 判断，不构成事实冒充观点。","“实际效果仍可能受语音清晰度与上下文影响”与来源中关于发音不清、术语较多及语境复杂时需要更多上下文的信息一致。","可将“说话人区分”统一为“说话人分离”，以贴近来源术语；这属于可选措辞优化，不影响审核结论。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":1,"checks":["schema","length","source-attribution","editorial-labels","inference-boundary","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","IT之家（RSS）"],"translations":{"zh-CN":{"title":"Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写","summary":"Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","category":"行业动态","source":"IT之家","aggregationSource":"IT之家（RSS）","pageTitle":"Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写 - Aioga AI资讯","description":"Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","url":"https://www.aioga.com/news/cmtjczoyu03akrobv35uivsvg/","articleBody":["Meta 发布实时音频感知模型 Muse Voice Transcribe，支持 20 余人分轨转写 这条更新来自 ithome.com，发布时间为 2026-09-02，Aioga 保留原文入口以便核验。","摘要：Meta 推出首个实时音频感知模型 Muse Voice Transcribe，整合流式语音识别、说话人分离与端点检测，可通过 Meta Model API 调用，定价每 1000 分钟音频 3 美元。","背景：IT之家援引9to5Mac称，Muse Voice Transcribe支持20余名说话者的录音分离，支持超过1小时音频，训练覆盖70余种语言，其中25种语言在发布时完成验证，定价为每1000分钟3美元。","Aioga 观察：Aioga判断：该模型把实时转写、多人发言区分和输入边界判断组合在一起，显示出连续音频处理的整合方向。材料未提供实际延迟、准确率或各语言表现数据，自适应延迟机制的效果仍需验证。","影响与后续：可能影响：开发者需要结合调用价格、已验证语言范围、录音人数和术语场景评估适配性。支持持续输出和多人分离不代表所有音频都能稳定识别，实际效果仍可能受语音清晰度与上下文影响。 后续观察：应关注Meta Model API的实际开放范围，以及不同语言、多人录音、自然语言切换和专业术语场景中的识别表现，并核验自适应延迟对实时性与准确度的影响。"]},"en":{"title":"Meta released the real-time audio perception model Muse Voice Transcribe, supporting track transcription for over 20 people","summary":"Meta launched its first real-time audio perception model, Muse Voice Transcribe, integrating streaming speech recognition, speaker separation, and endpoint detection, which can be called via the Meta Model API and priced at $3 per 1,000 minutes of audio.","category":"Industry","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta released the real-time audio perception model Muse Voice Transcribe, supporting track transcription for over 20 people - Aioga AI News","description":"Meta launched its first real-time audio perception model, Muse Voice Transcribe, integrating streaming speech recognition, speaker separation, and endpoint detection, which can be...","url":"https://www.aioga.com/en/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:22.169Z"},"ja":{"title":"Metaはリアルタイム音声知覚モデル「Muse Voice Transcribe」をリリースし、20人以上のトラック書き起こしをサポートします","summary":"Metaは初のリアルタイム音声知覚モデルであるMuse Voice Transcribeをリリースし、ストリーミング音声認識、スピーカー分離、エンドポイント検出を統合しました。これらはMeta Model APIを通じて呼び出せ、1,000分の音声あたり3ドルの価格で販売されています。","category":"業界動向","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Metaはリアルタイム音声知覚モデル「Muse Voice Transcribe」をリリースし、20人以上のトラック書き起こしをサポートします - Aioga AIニュース","description":"Metaは初のリアルタイム音声知覚モデルであるMuse Voice Transcribeをリリースし、ストリーミング音声認識、スピーカー分離、エンドポイント検出を統合しました。これらはMeta Model APIを通じて呼び出せ、1,000分の音声あたり3ドルの価格で販売されています。","url":"https://www.aioga.com/ja/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:23.620Z"},"ko":{"title":"Meta는 실시간 오디오 인지형 모델인 Muse Voice Transcribe를 출시했으며, 20명 이상의 사용자가 트랙 전사를 지원합니다","summary":"메타는 스트리밍 음성 인식, 화자 분리, 엔드포인트 감지를 통합한 최초의 실시간 오디오 인지형 모델인 뮤즈 보이스 트랜스크라이브를 출시했으며, 메타 모델 API를 통해 호출할 수 있고 1,000분당 3달러에 판매됩니다.","category":"업계 동향","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta는 실시간 오디오 인지형 모델인 Muse Voice Transcribe를 출시했으며, 20명 이상의 사용자가 트랙 전사를 지원합니다 - Aioga AI 뉴스","description":"메타는 스트리밍 음성 인식, 화자 분리, 엔드포인트 감지를 통합한 최초의 실시간 오디오 인지형 모델인 뮤즈 보이스 트랜스크라이브를 출시했으며, 메타 모델 API를 통해 호출할 수 있고 1,000분당 3달러에 판매됩니다.","url":"https://www.aioga.com/ko/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:33.139Z"},"es":{"title":"Meta lanzó el modelo de percepción de audio en tiempo real Muse Voice Transcribe, que soporta la transcripción de pistas para más de 20 personas","summary":"Meta lanzó su primer modelo de percepción de audio en tiempo real, Muse Voice Transcribe, que integra reconocimiento de voz en streaming, separación de altavoces y detección de endpoints, que puede llamarse a través de la API Meta Model y tiene un precio de 3 dólares por cada 1.000 minutos de audio.","category":"Industria","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta lanzó el modelo de percepción de audio en tiempo real Muse Voice Transcribe, que soporta la transcripción de pistas para más de 20 personas - Aioga Noticias de IA","description":"Meta lanzó su primer modelo de percepción de audio en tiempo real, Muse Voice Transcribe, que integra reconocimiento de voz en streaming, separación de altavoces y detección de end...","url":"https://www.aioga.com/es/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:31.520Z"},"fr":{"title":"Meta a publié le modèle de perception audio en temps réel Muse Voice Transcribe, qui prend en charge la transcription de pistes pour plus de 20 personnes","summary":"Meta a lancé son premier modèle de perception audio en temps réel, Muse Voice Transcribe, intégrant la reconnaissance vocale en streaming, la séparation des haut-parleurs et la détection des points de terminaison, qui peuvent être appelés via l’API Meta Model et sont évalués à 3 $ pour 1 000 minutes audio.","category":"Industrie","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta a publié le modèle de perception audio en temps réel Muse Voice Transcribe, qui prend en charge la transcription de pistes pour plus de 20 personnes - Aioga Actualités IA","description":"Meta a lancé son premier modèle de perception audio en temps réel, Muse Voice Transcribe, intégrant la reconnaissance vocale en streaming, la séparation des haut-parleurs et la dét...","url":"https://www.aioga.com/fr/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:42.652Z"},"de":{"title":"Meta veröffentlichte das Echtzeit-Audio-Wahrnehmungsmodell Muse Voice Transcribe, das die Übertragung von Spuren für über 20 Personen unterstützt","summary":"Meta brachte sein erstes Echtzeit-Audiowahrnehmungsmodell auf den Markt, Muse Voice Transcribe, das Streaming-Spracherkennung, Lautsprechertrennung und Endpunkterkennung integriert, die über die Meta Model API aufgerufen werden kann und 3 US-Dollar pro 1.000 Minuten Audio kostet.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta veröffentlichte das Echtzeit-Audio-Wahrnehmungsmodell Muse Voice Transcribe, das die Übertragung von Spuren für über 20 Personen unterstützt - Aioga KI-News","description":"Meta brachte sein erstes Echtzeit-Audiowahrnehmungsmodell auf den Markt, Muse Voice Transcribe, das Streaming-Spracherkennung, Lautsprechertrennung und Endpunkterkennung integriert...","url":"https://www.aioga.com/de/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:41.916Z"},"pt-BR":{"title":"A Meta lançou o modelo de percepção de áudio em tempo real Muse Voice Transcribe, que suporta transcrição de faixas para mais de 20 pessoas","summary":"A Meta lançou seu primeiro modelo de percepção de áudio em tempo real, Muse Voice Transcribe, integrando reconhecimento de fala em streaming, separação de falantes e detecção de endpoints, que pode ser chamado via API Meta Model e tem preço de $3 por 1.000 minutos de áudio.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"A Meta lançou o modelo de percepção de áudio em tempo real Muse Voice Transcribe, que suporta transcrição de faixas para mais de 20 pessoas - Aioga Notícias de IA","description":"A Meta lançou seu primeiro modelo de percepção de áudio em tempo real, Muse Voice Transcribe, integrando reconhecimento de fala em streaming, separação de falantes e detecção de en...","url":"https://www.aioga.com/pt-BR/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:52.412Z"},"ru":{"title":"Meta выпустила модель восприятия звука в реальном времени Muse Voice Transcribe, поддерживающую транскрипцию дорожек для более чем 20 человек","summary":"Meta выпустила свою первую модель восприятия аудио в реальном времени — Muse Voice Transcribe, интегрирующую распознавание потоковой речи, разделение динамиков и обнаружение конечных точек, которую можно вызвать через API Meta Model и оценивать $3 за 1000 минут аудио.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta выпустила модель восприятия звука в реальном времени Muse Voice Transcribe, поддерживающую транскрипцию дорожек для более чем 20 человек - Aioga Новости ИИ","description":"Meta выпустила свою первую модель восприятия аудио в реальном времени — Muse Voice Transcribe, интегрирующую распознавание потоковой речи, разделение динамиков и обнаружение конечн...","url":"https://www.aioga.com/ru/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:04:51.176Z"},"ar":{"title":"أصدرت ميتا نموذج إدراك الصوت في الوقت الحقيقي Muse Voice Transcribe، الذي يدعم نسخ المسارات لأكثر من 20 شخصا","summary":"أطلقت ميتا أول نموذج إدراك صوتي في الوقت الحقيقي لها، Muse Voice Transcribe، الذي يدمج التعرف على الكلام المتدفق، وفصل السماعات، واكتشاف نقاط النهاية، والذي يمكن استدعاؤه عبر واجهة برمجة تطبيقات Meta Model وتسعر بسعر 3 دولارات لكل 1000 دقيقة من الصوت.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"أصدرت ميتا نموذج إدراك الصوت في الوقت الحقيقي Muse Voice Transcribe، الذي يدعم نسخ المسارات لأكثر من 20 شخصا - Aioga أخبار الذكاء الاصطناعي","description":"أطلقت ميتا أول نموذج إدراك صوتي في الوقت الحقيقي لها، Muse Voice Transcribe، الذي يدمج التعرف على الكلام المتدفق، وفصل السماعات، واكتشاف نقاط النهاية، والذي يمكن استدعاؤه عبر واجهة...","url":"https://www.aioga.com/ar/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:01.356Z"},"hi":{"title":"Meta ने रीयल-टाइम ऑडियो परसेप्शन मॉडल Muse Voice Transcribe जारी किया, जो 20 से अधिक लोगों के लिए ट्रैक ट्रांसक्रिप्शन का समर्थन करता है","summary":"मेटा ने अपना पहला रीयल-टाइम ऑडियो परसेप्शन मॉडल, म्यूज़ वॉयस ट्रांसक्राइब लॉन्च किया, जिसमें स्ट्रीमिंग स्पीच रिकग्निशन, स्पीकर सेपरेशन और एंडपॉइंट डिटेक्शन को एकीकृत किया गया है, जिसे मेटा मॉडल एपीआई के माध्यम से कॉल किया जा सकता है और इसकी कीमत $3 प्रति 1,000 मिनट के ऑडियो को है।","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta ने रीयल-टाइम ऑडियो परसेप्शन मॉडल Muse Voice Transcribe जारी किया, जो 20 से अधिक लोगों के लिए ट्रैक ट्रांसक्रिप्शन का समर्थन करता है - Aioga AI समाचार","description":"मेटा ने अपना पहला रीयल-टाइम ऑडियो परसेप्शन मॉडल, म्यूज़ वॉयस ट्रांसक्राइब लॉन्च किया, जिसमें स्ट्रीमिंग स्पीच रिकग्निशन, स्पीकर सेपरेशन और एंडपॉइंट डिटेक्शन को एकीकृत किया गया है,...","url":"https://www.aioga.com/hi/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:00.989Z"},"it":{"title":"Meta ha rilasciato il modello di percezione audio in tempo reale Muse Voice Transcribe, che supporta la trascrizione delle tracce per oltre 20 persone","summary":"Meta ha lanciato il suo primo modello di percezione audio in tempo reale, Muse Voice Transcribe, che integra riconoscimento vocale in streaming, separazione degli altoparlanti e rilevamento degli endpoint, che può essere chiamato tramite l'API Meta Model e costato 3$ ogni 1.000 minuti di audio.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta ha rilasciato il modello di percezione audio in tempo reale Muse Voice Transcribe, che supporta la trascrizione delle tracce per oltre 20 persone - Aioga Notizie IA","description":"Meta ha lanciato il suo primo modello di percezione audio in tempo reale, Muse Voice Transcribe, che integra riconoscimento vocale in streaming, separazione degli altoparlanti e ri...","url":"https://www.aioga.com/it/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:10.956Z"},"nl":{"title":"Meta bracht het realtime audioperceptiemodel Muse Voice Transcribe uit, dat tracktranscriptie ondersteunt voor meer dan 20 mensen","summary":"Meta lanceerde zijn eerste realtime audioperceptiemodel, Muse Voice Transcribe, dat streaming spraakherkenning, luidsprekerafscheiding en eindpuntdetectie integreert, die via de Meta Model API kan worden opgeroepen en $3 per 1.000 minuten audio kost.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta bracht het realtime audioperceptiemodel Muse Voice Transcribe uit, dat tracktranscriptie ondersteunt voor meer dan 20 mensen - Aioga AI-nieuws","description":"Meta lanceerde zijn eerste realtime audioperceptiemodel, Muse Voice Transcribe, dat streaming spraakherkenning, luidsprekerafscheiding en eindpuntdetectie integreert, die via de Me...","url":"https://www.aioga.com/nl/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:10.345Z"},"tr":{"title":"Meta, 20'den fazla kişi için parça transkripsiyonunu destekleyen gerçek zamanlı ses algısı modeli Muse Voice Transcribe'ı piyasaya sürdü","summary":"Meta, ilk gerçek zamanlı ses algılama modeli olan Muse Voice Transcribe'ı başlattı; bu model, akış konuşma tanıma, hoparlör ayrımı ve uç nokta algılamasını entegre ediyor; bu model, Meta Model API üzerinden çağrılabiliyor ve 1.000 dakikalık ses başına 3 dolar fiyatına sunuluyor.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta, 20'den fazla kişi için parça transkripsiyonunu destekleyen gerçek zamanlı ses algısı modeli Muse Voice Transcribe'ı piyasaya sürdü - Aioga AI Haberleri","description":"Meta, ilk gerçek zamanlı ses algılama modeli olan Muse Voice Transcribe'ı başlattı; bu model, akış konuşma tanıma, hoparlör ayrımı ve uç nokta algılamasını entegre ediyor; bu model...","url":"https://www.aioga.com/tr/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:20.237Z"},"vi":{"title":"Meta đã phát hành mô hình nhận thức âm thanh thời gian thực Muse Voice Transcribe, hỗ trợ phiên âm bản nhạc cho hơn 20 người","summary":"Meta đã ra mắt mô hình nhận thức âm thanh thời gian thực đầu tiên, Muse Voice Transcribe, tích hợp nhận dạng giọng nói trực tuyến, tách biệt loa và phát hiện điểm cuối, có thể gọi qua API Meta Model và có giá $3 cho mỗi 1.000 phút âm thanh.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta đã phát hành mô hình nhận thức âm thanh thời gian thực Muse Voice Transcribe, hỗ trợ phiên âm bản nhạc cho hơn 20 người - Tin tức AI Aioga","description":"Meta đã ra mắt mô hình nhận thức âm thanh thời gian thực đầu tiên, Muse Voice Transcribe, tích hợp nhận dạng giọng nói trực tuyến, tách biệt loa và phát hiện điểm cuối, có thể gọi...","url":"https://www.aioga.com/vi/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:20.635Z"},"id":{"title":"Meta merilis model persepsi audio real-time Muse Voice Transcribe, yang mendukung transkripsi trek untuk lebih dari 20 orang","summary":"Meta meluncurkan model persepsi audio real-time pertamanya, Muse Voice Transcribe, yang mengintegrasikan pengenalan suara streaming, pemisahan speaker, dan deteksi endpoint, yang dapat dipanggil melalui API Meta Model dan dihargai $3 per 1.000 menit audio.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta merilis model persepsi audio real-time Muse Voice Transcribe, yang mendukung transkripsi trek untuk lebih dari 20 orang - Berita AI Aioga","description":"Meta meluncurkan model persepsi audio real-time pertamanya, Muse Voice Transcribe, yang mengintegrasikan pengenalan suara streaming, pemisahan speaker, dan deteksi endpoint, yang d...","url":"https://www.aioga.com/id/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:30.096Z"},"th":{"title":"Meta ได้เปิดตัวโมเดลการรับรู้เสียงแบบเรียลไทม์ Muse Voice Transcribe ซึ่งรองรับการถอดเสียงแทร็กสําหรับผู้ใช้มากกว่า 20 คน","summary":"Meta เปิดตัวโมเดลการรับรู้เสียงแบบเรียลไทม์ครั้งแรก Muse Voice Transcribe ซึ่งผสานการรู้จําเสียงพูดแบบสตรีมมิ่ง การแยกเสียงพูด และการตรวจจับปลายทาง ซึ่งสามารถเรียกผ่าน Meta Model API และมีราคา $3 ต่อ 1,000 นาทีของเสียง","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta ได้เปิดตัวโมเดลการรับรู้เสียงแบบเรียลไทม์ Muse Voice Transcribe ซึ่งรองรับการถอดเสียงแทร็กสําหรับผู้ใช้มากกว่า 20 คน - ข่าว AI Aioga","description":"Meta เปิดตัวโมเดลการรับรู้เสียงแบบเรียลไทม์ครั้งแรก Muse Voice Transcribe ซึ่งผสานการรู้จําเสียงพูดแบบสตรีมมิ่ง การแยกเสียงพูด และการตรวจจับปลายทาง ซึ่งสามารถเรียกผ่าน Meta Model A...","url":"https://www.aioga.com/th/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:30.174Z"},"pl":{"title":"Meta wypuściła model percepcji dźwięku w czasie rzeczywistym Muse Voice Transcribe, wspierający transkrypcję utworów dla ponad 20 osób","summary":"Meta wprowadziła swój pierwszy model percepcji dźwięku w czasie rzeczywistym, Muse Voice Transcribe, integrujący rozpoznawanie mowy strumieniowej, separację głośników oraz wykrywanie punktów końcowych, który można wywołać za pomocą Meta Model API i kosztować 3 USD za 1000 minut nagrania audio.","category":"行业动态","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Meta wypuściła model percepcji dźwięku w czasie rzeczywistym Muse Voice Transcribe, wspierający transkrypcję utworów dla ponad 20 osób - Aioga Wiadomości AI","description":"Meta wprowadziła swój pierwszy model percepcji dźwięku w czasie rzeczywistym, Muse Voice Transcribe, integrujący rozpoznawanie mowy strumieniowej, separację głośników oraz wykrywan...","url":"https://www.aioga.com/pl/news/cmtjczoyu03akrobv35uivsvg/","contentTranslated":true,"sourceHash":"ca4be88a3379fd27","translatedAt":"2026-09-02T01:05:39.632Z"}},"evidenceTier":"verified-news","reviewStatus":"automated-ingest","indexable":true,"editorialCover":""}}