Anyone can easily create browser-accessible photorealistic 3D scenes without code
In the latest update, we made a significant improvement to the user experience and changed the layout and design of some key ui components.
In the previous version, the camera system struggled to orbit around objects smoothly, and moving the camera could have been more convenient. The new camera system has optimized the algorithm, allowing for adjustable camera movement speed and switching between “Orbit” and “Fly” modes.
In the previous version, Splat Edit had a panel. In the latest version, we’ve integrated the Splat Edit functionality directly into the canvas, allowing you to activate Splat Edit mode through the toolbar.
The new Splat Edit mode retains most of the functionality of the old version. After selecting splat files in the scene using tools like React and Brush, you can delete or remove them.
In the previous version of the client, when training with images, all files had to be placed in a folder named “images.” Training would start by dragging one image from this folder into the client interface, and the training process would use all the images in that folder.
The image folder can now be named anything. To start training, users can drag the entire folder into the client interface or select multiple images from the folder and drag them into the interface to begin training.
SHBand, the value of this parameter, can change the performance of the splat file. The larger the SHBand, the better it can represent high-frequency colors.