- # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # I # # # #
- ** Example Voice Design v3:**
ElevenLabs has launched a new generation voice generation tool, Voice Design v3, which allows users to create characterized voice that fits the character set through texttips.
-
The new version of the model has a higher sound quality and greater expression.
-
Support 70+ Language and hundreds of local accents suitable for production grade.
-
Major updates include: (b) More detailed control over the personality and rhythm of speech roles (e.g. tone, speed, tone change, age and sex);
-
Support for Eleven v3 and new audio labels;
-
Enhancement of the oral recognition process;
-
Higher guaranteed audio output;
Voice Design v3 supports more than 70 languages and local accents, suitable for various creative and commercial uses. Officially, many live cases, including Japanese warriors, witches, French women and funny aliens, show the flexibility and performance of the system. The functionality is now open to all users. Online experience: https://elevenlabs.io/voice-design
# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # I # # # #
- Voice generation based on description You just have to enter a descriptive text that allows the system to produce the corresponding voice immediately. For example:
-
** Example: “A calm, tenacious, dumb old cowboy with a strong Southern American accent.”
-
** Generate voice**: Looks like it’s just me and you left…”
- Highly modifiable parameters
The user can customise the following speech features:
🧠 3. 模型支持的角色风格 Voice Design supports the creation of a variety of voice styles, whether written or imagined:
-
Real style: “A 20-year-old Indian woman with a high voice, a soft voice, a slow and gentle voice.”
-
“Middle-aged English males, sound magnetic, formal and clear.”
Role style:
-
“A rude, angry old pirate, often shouts.”
-
“An alien with a sharp voice, a croak and a funny mouth.”
- Use of Guides and Reminders
To help users better design their voice, Steven Labs provides a special tip writing guide, which includes dozens of style templates, building techniques, best practices, and allows newcomers to start writing easily.
⚙️ 5. 语音设计 API（开发者功能） Voice Design provides an API interface (currently Alpha phase, limited to the Console):
-
** Generate interface**: Enter prompt and text to return a sound preview.
-
Save interface: Save the resulting voice in the user voice library.
For developers such as product integration, automated creation, personalized content generation, etc.
** Example Voice Design v3:**
** Sound reminder: A calm, tough and gruff old cowboy with an deaf, grively, loud American accent. ** Sound reminder: A calm and loud maker with a dick Japanese accent. Soft, whiffery, low tone with a calm and gentle paching. ** Sound reminder:** She has a croaky, harsh, shrill, high-pitch voice that crackles. ** Sound reminder:** A low, while strong French voice with a sick French audience. Cool, attracted and secret, with a hit of mystery. ** Soundings, voices. I must say your voice models are out of this world! ** Voice reminder:** The friendly mythical God, Zeus, with a huge deep power voice. Charming, proud, strong and dramatic. ** Sound reminder: **A very old, cranky and croaky African-American grandma. 80 years old. Very horace, grumpy, shrill and frested. A very old, grumpy and dumb African-American grandmother, 80 years old.