SPZ

What is SPZ?

SPZ is a compressed file format for 3D Gaussian Splats, designed by Niantic. It provides efficient storage of Gaussian Splat data with configurable spherical harmonics degrees and coordinate system support.

See docs/SPZ.md for more information.

CLI

$ path/to/spz info assets/racoonfamily.spz
GaussianSplat={num_points=932560, sh_degree=3, antialiased=true, median_ellipsoid_volume=0.0000000046213082, bbox=[x=-281.779541 to 258.382568, y=-240.000000 to 240.000000, z=-240.000000 to 240.000000]}

Rust

Usage

spz = { version = "0.0.6", default-features = false, features = [] }

use spz::prelude::*;

Examples

cargo run --example load_spz

Quick Start

// SPDX-License-Identifier: Apache-2.0 OR MIT

use std::path::{Path, PathBuf};

use anyhow::Result;
use spz::{GaussianSplat, UnpackOptions};

fn main() -> Result<()> {
	let mut sample_spz = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
	sample_spz.push("assets/racoonfamily.spz");

	let _gs = spz::GaussianSplat::builder()
		.filepath(sample_spz)
		.packed(true)?
		.unpack_options(
			UnpackOptions::builder()
				.to_coord_system(spz::CoordinateSystem::default())
				.build(),
		)
		.load()?;

	Ok(())
}

#[allow(unused)]
async fn load_spz_async<P>(spz_file: P) -> Result<GaussianSplat>
where
	P: AsRef<Path>,
{
	let gs = spz::GaussianSplat::builder()
		.filepath(spz_file)
		.packed(true)?
		.unpack_options(
			UnpackOptions::builder()
				.to_coord_system(spz::CoordinateSystem::default())
				.build(),
		)
		.load_async()
		.await?;

	Ok(gs)
}

API

Overview

#[derive(Clone, Debug, Default, PartialEq, Serialize, Deserialize)]
pub struct GaussianSplat {
	pub num_points: i32,
	pub spherical_harmonics_degree: i32,
	pub antialiased: bool,
	pub positions: Vec<f32>,
	pub scales: Vec<f32>,
	pub rotations: Vec<f32>,
	pub alphas: Vec<f32>,
	pub colors: Vec<f32>,
	pub spherical_harmonics: Vec<f32>,
}

Benches

Pre-Requisites

• Install gnuplot for html reports.
• Install nextest runner.

Run

just bench

• The html report of the benchmark can be found under target/criterion/report/index.html.
• View Benchmark and Profiling data on CodSpeed, (from CI runs).

Test Code Coverage

Development

Pre-Requisites

• Install the mold linker: https://github.com/rui314/mold
• Install nextest

Python

TODO.

Documentation

Further documentation is available under ./docs.

License

Licensed under either of

• Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
  • SPDX-License-Identifier: Apache-2.0
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT) at your option.
  • SPDX-License-Identifier: MIT

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.