This article will help you solve common problems with creating your Overdub voice.
Common error messages
This error message occurs when the training composition contains less than 10 minutes of recorded audio in the timeline.
To resolve, please import or record additional speech audio into the composition.
You may encounter these error messages either in-app or through email if, after submitting a custom Overdub voice for training, our system couldn't match the Voice ID to at least 10 minutes of the accompanying training data. This can occur when:
- The training data has large gaps of silence between speech audio.
- The training data contains multiple speakers and less than 10 minutes of the intended speaker's voice was matched.
- The voice ID statement does not closely match the training data, most often due to differences in recording quality, background noise, etc.
Check your Voice ID
- Review your Voice ID and make sure it closely matches the training audio's tone and audio quality.
- Consider re-recording the Voice ID under the same conditions as your training audio.
- Once the Voice ID is consistent with your training audio, submit the AI voice for training.
Import or record additional speech
If the issue persists please import or record additional speech audio into the composition, or improve your recording conditions. Again, make sure that the audio quality of the Voice ID and training audio are as close as possible to avoid issues.
This error message indicates that the voice ID was either missing or incomplete. To resolve:
- Re-submit the training
- Select the appropriate recording device
- Make sure to record the complete voice ID statement
If you've received a "something went wrong" error message, here are some things you can try.
- Make sure there are no visual pins in your script - like text, captions, etc.
- Try recreating your project from scratch in a new voice - you can re-import the media or copy/paste from the other project.
If none of that works, please contact support and we'll help.
This error happens when you try to record your Voice ID using a multi-channel microphone. Currently, we only support recording the Voice ID consent statement in mono format.
To fix this problem, please record the Voice ID using a different microphone that supports either mono or stereo recording.
Create from voice is currently broken
You can only create overdub voices from scratch for now.
Unable to Record in Training Project
This should be fixed in an upcoming update, but if you're blocked due to issues with the Recorder in your training project, here you can try this workaround (recording is frozen, not starting, blocked by an error:)
1. Open a new Storyboard project, and record your audio (including training data, and the consent statement).
2. Highlight the full transcript and right-click to Copy.
3. Paste your transcript into the training project and hit Submit Training Data.
4. You will be prompted to record a new consent statement. Click record, then immediately hit stop.
5. You can finally click submit to finalize your training data.
If you don't like the sound of your Overdub voice or are having trouble using it
Check out our troubleshooting guide for using your Overdub voice.