Emotion level dubbing

更新时间:
复制 MD 格式

Emotion level dubbing controls the emotional intensity of dubbed audio in speech translation. Set an emotion level to reduce source-language intonation patterns that can make cross-language dubbing sound unnatural to native speakers of the target language.

How it works

In speech translation, the system uses voice cloning by default to replicate the original speaker's voice, emotions, and prosody for the target language dubbing. However, cross-language dubbing can carry the source language's intonation patterns into the target language, producing dubbing that sounds accented and unnatural.

Emotion level dubbing addresses this by weakening the source language's intonation and emotional characteristics. It provides two emotion control modes:

Emotion levelValueDescription
Slight emotionSlightEmotionRetains the basic intonation of the original audio. The dubbing sounds natural with slight emotional inflections.
No emotionNoEmotionUses a completely flat and stable tone for dubbing with no emotional inflections.
Note

If you do not set the EmotionLevel parameter, the system uses the default voice cloning mode for dubbing.

Emotion level dubbing is an enhancement configuration for speech translation. It is not an independent translation level. Only speech translation supports the EmotionLevel parameter. Subtitle translation does not involve dubbing and does not require this parameter. For more information about video translation features including subtitle translation, speech translation, and lip-sync translation, see the video translation overview.

Scenarios

The default voice cloning mode preserves the original speaker's intonation and emotional characteristics. During cross-language dubbing, the source language's intonation patterns are carried into the target language, causing the dubbed audio to have a noticeable foreign accent:

  • Chinese to English — Chinese intonation patterns are carried into the English dubbing, causing the English to sound unnatural with abnormal emotional inflections that do not match native English speech patterns.

  • English to Chinese — English intonation patterns are carried into the Chinese dubbing, causing the Chinese to sound like a non-native speaker with unnatural rhythm and tones.

    Emotion level dubbing does not regenerate the dubbing entirely in the target language's native intonation. Instead, it weakens the source language's intonation and emotional characteristics by removing or reducing the strong emotional coloring from the original language, making the dubbed audio sound more neutral, stable, and natural for native speakers of the target language.

The following table compares the effect of different dubbing modes for Chinese-English translation:

Dubbing modeEffectAdvantagesDisadvantagesScenarios
Default mode (voice cloning)Fully preserves the original speaker's intonation and emotionsRich emotional expression with distinctive original voice characteristicsNoticeable source language accent in cross-language dubbingContent that requires high emotional fidelity, such as film commentary and emotional short videos
Slight emotion (SlightEmotion)Weakens source language intonation while retaining slight emotional inflectionsIntonation tends to be stable, reduces source language accent while preserving some emotional coloringWeakened emotional expression, not suitable for highly emotional contentCorporate promotions, product introductions, online courses, and other content that requires stable intonation
No emotion (NoEmotion)Completely flat and stable tone with no emotional inflectionsMaximally removes source language accent, producing the most neutral and stable dubbingComplete absence of emotional expression, not suitable for content that requires emotional impactNews broadcasts, technical explanations, operation demonstrations, and other extremely neutral content
Note

NoEmotion maximally weakens emotional and intonation inflections, producing completely stable dubbing. SlightEmotion retains a small amount of original intonation, producing results between the default mode and no emotion. Select the appropriate level based on your content style.

Usage

Emotion level dubbing is available only through the API. The console does not support this configuration.

Parameter reference

When you call the SubmitVideoTranslationJob operation, set the emotion level by using the EditingConfig.SpeechTranslate.EmotionLevel field.

ParameterTypeRequiredValuesDescription
EmotionLevelStringNoSlightEmotionSlight emotion: retains the basic intonation of the original audio. The dubbing sounds natural with slight emotional inflections.
NoEmotionNo emotion: uses a completely flat and stable tone with no emotional inflections.

Examples

The following example translates a Chinese video into English with slight emotion dubbing:

{
  "InputConfig": {
    "Type": "Video",
    "Video": "1628ae20c36******8f6f7c77a6302"
  },
  "EditingConfig": {
    "SourceLanguage": "zh",
    "TargetLanguage": "en",
    "NeedSpeechTranslate": true,
    "SpeechTranslate": {
      "EmotionLevel": "SlightEmotion"
    }
  },
  "Title": "Slight emotion dubbing example",
  "OutputConfig": {
    "MediaURL": "https://****.oss-cn-shanghai.aliyuncs.com/output.mp4"
  }
}

The following example translates an English video into Chinese with no emotion dubbing:

{
  "InputConfig": {
    "Type": "Video",
    "Video": "a3f9b21d457******c8e2a1b3d4e5f6a7b"
  },
  "EditingConfig": {
    "SourceLanguage": "en",
    "TargetLanguage": "zh",
    "NeedSpeechTranslate": true,
    "SpeechTranslate": {
      "EmotionLevel": "NoEmotion"
    }
  },
  "Title": "No emotion dubbing example",
  "OutputConfig": {
    "MediaURL": "https://****.oss-cn-shanghai.aliyuncs.com/output_noemotion.mp4"
  }
}

Limits

The following limits apply to emotion level dubbing:

  • Supported languagesEmotionLevel takes effect only when the target language is Chinese (zh) or English (en). If you set a different target language, this parameter is ignored and the system uses the default dubbing mode.

Source languageTarget languageSupported
Chinese (zh)English (en)Yes
English (en)Chinese (zh)Yes
Other language combinationsNot supported
  • Supported regions — Emotion level dubbing is available in the Chinese mainland regions and Singapore.

RegionSupported
China (Shanghai)Yes
China (Beijing)Yes
China (Shenzhen)Yes
China (Hangzhou)Yes
SingaporeYes
  • Access method — Available only through the API. The console does not provide a configuration entry.

  • Parameter priority — If you set both EmotionLevel and VoiceCloneMode, the EmotionLevel setting takes priority.

  • Default behavior — If you do not pass the EmotionLevel parameter, the system uses the default voice cloning mode. All target languages supported by speech translation remain unaffected and continue to use the default dubbing mode.