How to use the latest lip sync model
To use the latest lip sync model for AI-personalized videos, follow these steps:
When creating a dynamic video campaign, select the "Settings" gear in the top right corner
Select "Voice + Lip Model (Beta)"
Now, you can go ahead and create videos from this campaign, and you will be using the latest model!
How to Get Best Results for Lip Cloning
Pre-Recording Preparations
Trim Silent or Unnecessary Initial Seconds: Before submitting your video for processing, ensure the first 10 seconds are meaningful and relevant, as Sendspark’s AI prioritizes this section for lip sync calibration.
Start With a Closed Mouth: Begin recording with your mouth in a neutral, closed position before speaking. This provides the AI with a clear starting point for synchronizing movements.
Lighting is super important!
After your greeting or introductory statement, briefly pause and close your mouth again before continuing with the main content of your message. This helps set the rhythm for smoother lip sync throughout the recording.
Make sure your face is well-lit, and your lips are in focus when recording your original video.
Also, follow the standard best practices to get the best results from voice cloning (speak normally, make sure there is no background noise, leave a brief pause after saying "watermelon" before the rest of your message).
Current Limitations of Lip Cloning: Sendspark’s lip sync feature is a powerful tool. However, minor imperfections in lip movements or lip sync might appear depending on the video content, though most viewers find these subtle and unnoticeable unless closely scrutinized. The Sendspark team is continually working on updates to enhance this technology.