ElevenLabs Voices in Rapport - Dec 2024

https://elevenlabs.io/docs/api-reference/get-voices

https://elevenlabs.io/docs/voices/default-voices#current-default-voices

The table below describes the voice models along with their speaker boost values and associated labels.

Explanation of Terms:

Fine-Tuned Models: The models that have been fine-tuned specifically for that voice.
Non-Fine-Tuned Models: Models compatible with the voice but not fine-tuned.
Speaker Boost: Indicates whether the speaker boost is set to True or False for that voice.

This table provides a comprehensive view of each voice, showing their fine-tuned and non-fine-tuned models along with the speaker boost status.

Fine-Tuned Models: This refers to whether a specific model has been customized or adjusted (fine-tuned) to better match the voice's characteristics. Fine-tuning can improve the performance of a model for a particular voice, making it sound more natural or accurate according to the intended use case.
Speaker Boost: This is a separate setting that can be turned on or off (True or False). Speaker Boost generally refers to an enhancement that makes the voice louder, clearer, or more prominent when it is used. It is not dependent on whether the model has been fine-tuned.

So, to summarize, while both fine-tuning and speaker boost can affect the output quality of a voice, they operate independently of each other.

Voice	VoiceID	Fine tuned models	Non-fine tuned models	Use case	Age	Gender	Accent	Description
Aria	9BWtsMINqrJLrRacOk9x	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2	None	social media	middle-aged	female	American	expressive
Roger	CwhRBWXzGAHq8TQ4Fs17	eleven_multilingual_v2, eleven_turbo_v2	eleven_turbo_v2_5 (failed)	social media	middle-aged	male	American	confident
Sarah	EXAVITQu4vr4xnSDxMaL	None	eleven_turbo_v2, eleven_multilingual_v2, eleven_turbo_v2_5	news	young	female	american	soft
Laura	FGY2WhTYpPnrIDTdsKH5	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2	None	social media	young	female	American	upbeat
Charlie	IKne3meq5aSn9XLyUdCD	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5	conversational	middle aged	male	Australian	natural
George	JBFqnCBsd6RMkjVDRZzb	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	narration	middle aged	male	British	warm
Callum	N2lVS1w4EtoT3dr4eOWO	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5	characters	middle-aged	male	Transatlantic	intense
River	SAz9YHcvj6GT2YYXdXww	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2, eleven_multilingual_sts_v2	None	social media	middle-aged	non-binary	American	confident
Liam	TX3LPaxmHKxFdv7VOQHJ	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5, eleven_flash_v2_5, eleven_flash_v2	narration	young	male	American	articulate
Charlotte	XB0fDUnXU5powFXDhCwa	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5	characters	young	female	Swedish	seductive
Alice	Xb7hH8MSUJpSbSDYk0k2	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	news	middle-aged	female	British	confident
Matilda	XrExE9yKIg1WjnnlVkGX	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5	narration	middle-aged	female	American	friendly
Will	bIHbv24MWmeRgasZH58o	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2	None	social media	young	male	American	friendly
Jessica	cgSgspJ2msm6clMCkdW9	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2	None	conversational	young	female	American	expressive
Eric	cjVigY5qzO86Huf0OWal	eleven_multilingual_v2, eleven_turbo_v2_5, eleven_turbo_v2	None	conversational	middle-aged	male	American	friendly
Chris	iP95p4xoKVk53GoZ742B	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	conversational	middle-aged	male	American	casual
Brian	nPczCjzI2devNBz1zQrb	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	narration	middle-aged	male	American	deep
Daniel	onwK4e9ZLuTAKqWW03F9	eleven_turbo_v2	eleven_multilingual_v1, eleven_multilingual_v2, eleven_turbo_v2_5	news	middle-aged	male	British	authoritative
Lily	pFZP5JQG7iQjIQuC4Bku	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	narration	middle-aged	female	British	warm
Bill	pqHfZKP75CvOlQylNhV4	eleven_turbo_v2	eleven_multilingual_v2, eleven_turbo_v2_5	narration	old	male	American	trustworthy

Default pre-made voice settings

These settings can be used in the Rapport User Interface and are set in the TTS Args

JSON

{
   "model_id":"eleven_flash_v2_5",
   "voice_settings":{
      "stability":0.5,
      "similarity_boost":0.8,
      "style":0.0,
      "use_speaker_boost":true
   }
}

ElevenLabs Voice Settings in Rapport

Under Project Settings set the Text to Speech (TTS) option to be ElevenLabs.
The voice settings can then be set in the TTS Args field inon the user interface.

To do this we can enter a small piece of JSON code as shown below. This example uses the model eleven_turbo_v2_5

The TTS arguments field is a write field and will validate if an incorrect JSON format has been entered. If you copy the format shown above then everything should be fine.