Hi everyone! I'm trying to create a consistent voice for my company's internal training videos using ElevenLabs. We have a lot of content, and using a real narrator every time isn't scalable.
I've been experimenting with their Voice Lab. Here's my basic workflow so far:
1. I recorded a clean, 5-minute audio sample of our training manager reading a script.
2. Uploaded it in the Voice Lab to create a "cloned" voice.
3. Generated a test script with the new voice.
The initial clone sounds pretty good, but sometimes it mispronounces our internal product names. I tried adding those terms to the pronunciation dictionary with IPA, but results are mixed.
Here's an example of the dictionary entry I used:
```
{
"word": "XyloCore",
"ipa": "zaɪloʊkɔːr"
}
```
Has anyone else done this for internal use? Any tips for getting more consistent pronunciation on very specific jargon? Also, is the Professional plan sufficient for this kind of bulk generation, or should I look at the next tier?
Thanks in advance for any advice