Enabling the browser camera when asked is recommended! Although it’s possible to mint without a camera, the minting experience is designed to be used with TensorFlow hand-tracking. The outputs are generated using the minter’s address and the seed parameter, which means that what you see is what you get!
The minting interface is not designed for mobiles.
The minting UI is divided into 3 sections: [ Setup | Params | Capture ]
[ Params ]
If you haven't recorded a customized path yet, switching to a different seed will automatically generate an algorithmic path, in case you prefer the look of that or can’t use a camera. Seed also changes the styling. You can modify the loop duration before OR after recording, which can be useful for capturing paths in slower motion or smoothing out the path.
Before minting, you should define a point in time (Preview Snap %) for capturing the thumbnail preview. [ c ] can be used to set it instead of the slider!
[ Setup ]
- The debug layer is enabled automatically when capturing a path
- You can change to any detected camera source
- Hand-track smoothing can be helpful to avoid flickering in the detection
[ Capture ]
Start recording by using the button or [ r ] - start/stop recording