AI Retakes
Need to fix something you said without re-recording your whole video? Whether you made a small mistake, want to reword a sentence, or just need to tweak a detail, AI Retakes makes it easy.
What are AI Retakes?
AI Retakes let you replace a section of your voiceover in a CANVID project by either:
- Typing what you want to say (the app will generate it using your voice)
- Speaking the new audio directly with your microphone
It automatically replaces the audio, updates your video to lip-sync your webcam footage (if used), and keeps everything else intact without affecting your timeline.
When to Use It
- You said the wrong word, number, or date
- You want to improve your delivery
- You need to fix a mistake but the rest of the video is great
- You changed your mind about how something should be phrased
How to Access AI Retakes
Generate a transcript
Before you can us AI Retakes, the app needs a transcript of your video
Select the part to replace
In the Captions sidebar, highlight the words you want to update. (maximum of 15 seconds only)
Select Replace Audio
A context menu should appear. Choose Replace Audio to get started.
Select processing mode
- Cloud: Use CANVID's own cloud infrastructure
- Local: Use your own computer's system resources (Intel AI PCs or desktops with Intel Arc graphics)
Option 1: Typing
Use this option if you want to use AI Voice cloning to create new audio with text-to speect. Click the box to the left that says Type what you want to say.
Adjust selection
In this section, you can fine-tune the audio selection for your chosen words or sentence by adjusting the handles. Use the play button to make sure the audio flows naturally and isn’t cut off abruptly for the best results.
Click Next Step: Replace Audio with Text to continue.
Replace audio with text
Re-type the full sentence, making sure to include the changes you want to apply.
Click Finish when done.
Start processing
The app will begin processing your new sentence.
Done
Review the transcript to confirm the changes you’ve made.
Option 2: Speaking
Use this option to record the new audio directly with your microphone. Click the box to the right that says Speak what you want to say to get started.
Adjust selection
Similar to the first option, you can also adjust the handles in this section to fine-tune the audio if necessary. Use the play button to make sure the audio flows naturally and isn’t cut off abruptly for the best results.
Record audio
- Check if you're using an active microphone using the device selection below the Re-record button.
- When ready, Click Re-record to start recording your new sentence. Click Stop when done.
- Your newly recorded audio will automatically playback so you can listen. When you're happy, click Next Step: Review Changes to continue.
Review changes
You will see a transcript of your new voice recording. Check if everything is correct, then click Finish.
Start processing
The app will begin processing your new sentence.
Done
Review the transcript to confirrm the changes you've made.
Tips for Best Results
- It’s best to select and replace an entire sentence or at least a substantial segment/chunk, not just one or two isolated words
- Check that the segment you’re replacing doesn’t phonetically merge with the words before or after it. In other words, make sure to select a chunk that audio wise can stands alone, so it can blend nice with the audio before and after
- Make sure the selection is 15 seconds or less
- Try to match the tone and pace of your original voice (when not using Text-to-speech)
- Keep your background quiet if you are using a microphone