Canva: Refining AI video results with specific prompt engineering
CanvaDemo summary
The user demonstrates refining a Canva AI video by adding specific details like 'front view shot', 'blonde', and 'hyperrealistic' to the prompt to change the model's appearance and camera position.
Step-by-step
- Enter a detailed prompt specifying the subject, clothing material, and audience actions
- Define the camera angle by adding 'front view shot' and 'without any camera movement'
- Specify physical attributes like 'blonde' and 'wearing high heels'
- Add 'hyperrealistic' to the prompt to improve visual stability
- Remove background elements to focus on the model's generation first
Options
- Simplify the scene by removing background elements before adding them back later
Watch out for
- AI-generated videos may experience glitching at the start of the clip
Tips
- Define more specific details within the prompt to see a clear difference in the generated output
- Use 'hyperrealistic' as a keyword to help the prompt's quality
- Focus on getting the shot of the subject correct before moving forward with complex background elements
Highlights
“it is glitching at the start a bit, but I don't think that this is a very stable shot”
All demos from “Canva AI Video Generator Tutorial 2026 (Beginners)”
- 2:052:18Generate AI video of puppies using Canva Magic MediaThe user demonstrates how to navigate to the Magic Media app in Canva, select the video tab, and enter a descriptive prompt to generate a 4-second clip of puppies playing.Canva· Text to Video
- 5:060:56Generate AI video with generic promptsThe creator shows the result of using a generic prompt for a fashion model in Canva Magic Media to illustrate how the AI interprets basic instructions without specific camera angles.Canva· Text to Video
- 6:251:07Refining AI video results with specific prompt engineeringCurrentThe user demonstrates refining a Canva AI video by adding specific details like 'front view shot', 'blonde', and 'hyperrealistic' to the prompt to change the model's appearance and camera position.Canva· Text to Video
- Watch “Canva AI Video Generator Tutorial 2026 (Beginners)” →
Text to Video
10:520:27Free AI video creation with AI Video MakerThe creator shows the advanced settings in aivideomaker.ai, selecting the free model and 480p quality to generate a short clip of a tiger.Dom the AI Tutor | Tech Tutor Zones
7:521:14Generate video clips with Google V3The tutorial shows how to generate a cinematic overhead shot of a glacier from a text prompt using Canva's 'Create a video clip' tool.Brenda Cadman, Canva Verified Expert
22:370:32AI video creation in CanvaA brief walkthrough of using Canva's AI video tools to generate a clip of a cat working, including background typing sounds.Beth Ziesenis
0:501:59Generate AI video from text in CanvaThe user demonstrates how to access Magic Media in the Canva editor and use a text prompt to generate a hyperrealistic video of a robotic hand touching a holographic interface.Jason Gandy- 3:080:51Generate landscape drone footage with Magic MediaA demonstration of generating a nature scene using a text prompt for a misty pine forest drone shot with a camera pan using Canva's AI video tool.Jason Gandy
- 4:110:43Generate realistic human scenes with Canva AIThe user shows how to generate a complex cinematic close-up of a person drinking coffee in a shop using text-to-video generation.Jason Gandy
- 2:052:18Generate AI video of puppies using Canva Magic MediaThe user demonstrates how to navigate to the Magic Media app in Canva, select the video tab, and enter a descriptive prompt to generate a 4-second clip of puppies playing.The Pennywise Owl
- 5:060:56Generate AI video with generic promptsThe creator shows the result of using a generic prompt for a fashion model in Canva Magic Media to illustrate how the AI interprets basic instructions without specific camera angles.The Pennywise Owl
- 6:251:07Refining AI video results with specific prompt engineeringCurrentThe user demonstrates refining a Canva AI video by adding specific details like 'front view shot', 'blonde', and 'hyperrealistic' to the prompt to change the model's appearance and camera position.The Pennywise Owl