ElevenLabs:語音轉語音

上傳源音訊檔案以應用新的語音語調或克隆不同的聲音,產生修改後的音訊輸出。

Input

A recording of someone speaking; the words, timing and delivery are kept and the voice is replaced.

  • 0:00 / 0:00

The ElevenLabs voice to speak in.

Advanced settings

Multilingual works for any language; English v2 is tuned for English.

Clean up a noisy recording first.

Leave empty for a new result each run. Enter a number to repeat one.

How closely the result matches the chosen voice.

Pushes the result closer to the voice, a little slower.

Speaking speed (0.7 to 1.2).

Lower is more expressive and varied; higher is steadier.

Exaggerates the voice's style (0 to 0.2).

Output
Example
00:00

試試範例

關於此工作流程

此 ComfyUI 工作流程利用 ElevenLabs Speech-to-Speech API 進行語音轉換。你可提供來源音訊(用 LoadAudio 或 RecordAudio),選擇目標語音(ElevenLabsVoiceSelector 選預設語音,或 ElevenLabsInstantVoiceClone 建立新語音),ElevenLabsSpeechToSpeech 節點會產生新錄音,保留原始內容與節奏,但聲音特質改為所選語音。可用 SaveAudioMP3 或 SaveAudioOpus 儲存輸出。

技術上,流程將來源音訊與目標語音 ID 一同送入 ElevenLabsSpeechToSpeech。選項一,ElevenLabsVoiceSelector 會查詢 ElevenLabs 帳號並輸出選定語音。選項二,ElevenLabsInstantVoiceClone 上傳乾淨參考音訊,回傳新語音資源,再送入 Speech-to-Speech 節點。Note 節點提供指引,部分節點預設旁路,僅需啟用所需部分。設計上方便 A/B 測試預設語音與即時複製,並用明確儲存節點簡化檔案管理。

問答