When filming content for LipDub, its important to understand some of the best practices so you can be confident that your footage will produce the best possible result.
Footage can range from very simple to extremely complex. The more dynamic the content to higher the risk for artifacts. Most content lies is some where in the middle, so we always recommend to just try it out and see if it like the results! 😊
Here’s some of our Filming best practices for you to keep in mind:
- Capture at least 30seconds to 1min of footage when the actor is speaking. More the better 😄
- There are diminishing returns when finetuning the AI model so after 5min+ of data, no need to provide any more.
- Capture a wide diversity of mouth shapes
- This is typically already captured if you have that 30 second to 1min of footage of them speaking
- Consistent lighting throughout the video
- Try to avoid occlusions of the face area & extreme angles if possible
For more info on what to avoid in your footage if its possible, please check out this article: