#i uploaded my own voice for PVC and successfully completed verification. When i try to use it, it
1 messages · Page 1 of 1 (latest)
After successfully verifying your Professional Voice Clone, it may take some time for the custom model to be fine-tuned and ready for use. The estimated processing time is approximately 3 hours. Please try accessing your voice clone again after this period. If the issue persists, feel free to reach out for further assistance.
how can tell from the platform that it is not ready yet or tuning is going on
To check the status of your Professional Voice Clone, go to the "My Voices" section. Find your voice entry and select 'View.' You will see the recommended models and their status. If a model is not yet fine-tuned, it will be indicated with a plus sign, meaning further training is needed.
The fine-tuning process can take between 6-24 hours, depending on the queue. You can monitor the progress by hovering over each model to check its status. Once the process is complete, you will be notified, and the icon next to the model will change, indicating that your voice is ready for use.
what happens to the tts if the recorded samples have filler words and some stuttering here and there, basically normal conversational speech but considered to be imperfect for diction
If the recorded samples for your Professional Voice Clone contain filler words and stuttering, it may lead to unwanted pauses and extraneous noises in the text-to-speech output. To improve the quality, it's recommended to train the AI voice on samples without filler pauses. This can help reduce extraneous noises and achieve a more consistent output. Adjusting your approach to the pauses and training data can lead to better results.
Your samples after cleanup have duration of 3327 seconds. Ideal duration is more than 2 hours.
The effectiveness of your voice clone largely depends on the quality and consistency of the samples rather than just the quantity. While having over 2 hours of high-quality samples is recommended for optimal results, a shorter duration may still yield a good quality clone if the samples are consistent and of high quality.
If your samples contain filler words or stuttering, it might affect the final quality of the voice clone. It's important to provide clean and clear recordings to ensure the best possible outcome.
if clone is not good, can i delete it, re-record samples and start again
Yes, you can delete your existing Professional Voice Clone and start the process again with new recordings. To do this, navigate to "Voices" in the left sidebar, go to "Personal," locate your voice, click "View," and then select "Delete."
When re-uploading, ensure you provide at least 2 hours of high-quality audio. The audio should feature only a single speaker, without background noise, excessive reverb, music, or long pauses. This will help improve the accuracy and quality of the voice clone. If you need further assistance, feel free to ask.
Is it interesting to use Pronunciation Dictionaries for basic french words that the AI misprunonce ? Or is it just for the non french World like proper noun ?