#The voiceId in the websocket connection corresponds to a particular voice. If I want the ssml parsi

1 messages · Page 1 of 1 (latest)

minor ginkgoBOT
#
AI-generated Answer for suraj_63976

The behavior when using a different voice than the one specified by the voice_id in a WebSocket connection is not explicitly detailed in the provided content. However, it is important to ensure that the voice_id you provide matches the voice you intend to use, as mismatches can lead to errors or unexpected results.

For SSML parsing, the voice's compatibility with the selected model and features (like SSML) is crucial. If the voice is not properly configured or supported for the model you're using, it may result in issues such as incorrect output or errors during generation.

To avoid such problems, ensure that:

  1. The voice_id corresponds to the correct voice you want to use.
  2. The voice is compatible with the model and features you are enabling, such as SSML parsing.

If you encounter specific issues, you can verify the voice details in the My Voices section or use the API to confirm the correct voice_id. For further assistance, you may also refer to the Streaming API documentation.

soft iron
#

give me an example of a working SSML with elevenlabs Jessica voice

minor ginkgoBOT
#
AI-generated Answer for suraj_63976

ElevenLabs supports SSML (Speech Synthesis Markup Language) for controlling various speech elements like pauses, pronunciation, and emphasis. Below is an example of a working SSML snippet for the "Jessica" voice:

<speak>
  <voice name="Jessica">
    Hello, <break time="500ms"/> welcome to ElevenLabs. 
    <emphasis level="strong">This is an example</emphasis> of SSML in action.
    You can also control <prosody rate="slow" pitch="low">speech rate and pitch</prosody>.
  </voice>
</speak>

Key Features in the Example:

  1. <voice>: Specifies the voice to use (e.g., "Jessica").
  2. <break>: Adds a pause (e.g., 500ms).
  3. <emphasis>: Highlights specific words or phrases.
  4. <prosody>: Adjusts speech rate and pitch.

Make sure to use the "Eleven English V1" or "Eleven Turbo V2" models, as they support SSML. If you encounter issues, ensure your API request includes the correct voice ID and SSML content.

soft iron
#

does the eleven_flash_v2_5 model support SSML ?