Explore our changelog for detailed updates.Recommended Equipment Specifications:
System Requirements: Windows 10 or later, Nvidia GeForce RTX 1060 or higher.
For optimal performance, we recommend the Nvidia GeForce RTX 3060 with 12GB of memory.
1.fix mask alpha
2.fix prune crash
3.fix post process texture create unitialized when resized window
4.fix when depth is invalid
1.Added support for exporting WebViewer HTML
2.Added support for exporting camera pose JSON files
3.Enhanced error handling for training data load failures
4.Fixed issues with importing RealityCapture data
5.Resolved issues with automatic MCMC model selection when initial sparse point cloud data is missing
6.Optimized memory usage by automatically selecting data pack loading methods based on dataset size
7.Reduced pre-allocated Splat memory usage during Vulkan rendering
0. Added support for automatic updates starting from version 33.
1. Rewrote the training algorithm framework to improve training speed and rendering performance.
2. Added the SplatMCMC model, which no longer depends on an initial sparse point cloud.
3. Added external camera import support, including colmap, realitycapture, and metashape.
4. Added support for importing .obj and .gltf models and mesh rendering.
5. Added the reduced.ply format to further compress data storage.
6. Improved the Splat editing tool, optimizing brush operations.
7. Added SplatData histogram statistics for selection editing.
8. Fixed the crash issue during training and model import when using non-English paths.
9. Fixed the occasional garbled text issue when exporting Splat to a file.
1. Improved camera controls.
2. Introduced a new Splat Edit interface
3. Now, it supports training with images by dragging and dropping any named image folder or a specific number of selected images
4. Added SH Band UI adjustments
5. Fixed several bugs
1. Added three sparse reconstruction methods. Users can select the desired type in the UI SfmSolver, with Glomapper set as the default for the fastest performance. Sparse point clouds are displayed during the Sfm stage.
2. Introduced a Mask option to focus on training objects within the effective areas of images. The corresponding mask files will be loaded if a mask folder is available. The alpha channel will be used as the mask for PNG training images.
3. Added a Resource Panel to view project resource files.
4. Added an option in GaussianComponent to exclude it from rendering. This can be set to false during training to speed up the process.
5. Added a ViewCube Axis.
6. Introduced Splat model merging and export functionality.
7. Fixed issues with Chinese path names.
8. Fixed the display issue of camera perspectives during training.
Added Grid display
Added Keyframe windowAdded support for MOV video format
Fixed scene load/save issuesFixed crash issue after editing PLY files
Fixed cropping issue in Crops
Improved camera movement experience for smoother operation
Fixed issue preventing new project creation in window
Fixed crash issue during training with GTX 1060 GPU
Optimized UI operations
Fixed abnormal video data processing issue
1. vulkan gpu buffer auto-release
2. Fixed ply, compressply output issues
1.Optimise Splat for micro-rasterized rendering
2.Fix occasional crashes in the Resize window
3.Added sky and HDR background rendering
4.Asynchronous loading of models and display of loading status.
5.Support for importing and rendering splat models rebuilt on other platforms.
6.The Brush displays a range marker when selected.
1.Added installation packages (launcher, updater) and file associations.
2.Added camera data processing progress for project loading.
3.Asynchronous scene loading.
4.Added pack data to reduce video memory usage (preliminary).
5.Added batch loader.
6.Scene display will immediately update after loading is complete.
7.Fixed display issues with multiple Gaussian models.
8.Added densifyIteration in the UI to control the number of densification iterations.
9.Fixed the bug where the process would not exit if the project creation window was closed immediately upon program startup.