v4.6.0

• Added support for generating LOD (lod-meta.json) format

gsbox cut -i lod0.ply -l 0 -i lod1.ply -l 1 -i lod2.ply -l 2 -cs 30000 -of sog -e env.ply -o output/lod-meta.json

gsbox-amd64-linux-v4.6.0.zip
gsbox-arm64-darwin-v4.6.0.zip
gsbox-win64-v4.6.0.zip

v4.5.5

• Synchronized the new block data formats 23 and 230 from SPX v3.

v4.5.4

• Performance Optimization: kmeans

v4.5.3

• Fixed a bug introduced in v4.5.2 (loss of spherical harmonic coefficients during PLY parsing)
• Optimized the K-means algorithm to streamline palette data
• Added support for outputting Level 1 or Level 2 spherical harmonic coefficients in SOG format

v4.5.2

Performance Optimization: Morton Sort

v4.5.1

• Performance Optimization: Clustering, ply/splat Reading

v4.5.0

• Optimizing the palette sorting significantly improves performance.

v4.4.1

• Some optimizations, including ignoring clustering calculations when sh0 to improve performance, etc.

v4.4.0

• Performance optimization, including leveraging existing palettes to avoid recalculations, will be very useful for sog2spz, spx2spz, sog2spx, and spx2sog.

• Add the command-line argument -q, with a range of 1-9 (default to 5), which can automatically set the related parameters for WebP compression and spherical harmonic coefficient clustering calculations based on the level.

For SPZ, levels 1-8 affect the calculation of spherical harmonic coefficients. Generally, the higher the level, the better the quality and the longer the computation time. At level 9, the spherical harmonic coefficients are encoded in the usual way.

For SPX and SOG, the levels affect both the calculation of spherical harmonic coefficients and the WebP compression parameters. Generally, the higher the level, the better the quality and the longer the computation time.

If the relevant parameters are manually set, they will no longer be determined by the quality level. Manual adjustments can be made when necessary to achieve optimization goals.

v4.3.1

• Fix spx v3 read/write bug