vimaec / s3d

A simple and efficient file format for storing and transmitting 3D scene data (geometry, scene graph, assets, etc.)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

S3D

The S3D file format, is a simple alternative to FBX, Collada (DAE), USD, and glTF formats for storing and transmitting 3D scene graphs and dependent assets. The S3D data layout is the BFAST data container format.

Unlike other formats S3D is agnostic about how tools choose to encode specific assets such as meshes, skin weights, or materials. It is a container for the scene graph that make it easy to efficiently get and set transform data and parent-child relationships. By not specifying these things S3D is less prone to change, much easier to use, and far more flexible.

Specification

The S3D file format is a BFAST file consisting of 3 to 7 data-buffers with the following contents specific interpretation:

  1. JSON manifest - Contains meta-information about the file and the assets encoded using UTF-8.
  2. Binary scene graph - A BFAST container of binary data describing the scene graph.
    1. Transforms - An array of matrices representing global transforms encoded as 4x4 single or double precision floating point vaues
    2. Geometry indices - An array of indices indicating which geometric asset is associated
    3. Material indices - An index indicating which material is associated with this node
    4. Parent indices - An array of indices indicating the parent node
  3. Geometric assets- A BFAST container of geometric assets.
  4. Material assets (optional) - A BFAST container of material descriptors.
  5. Texture assets (optional) - A BFAST container containing texture assets.
  6. Object metadata (optional) - UTF-8 encoded JSON properties.
  7. Additional assets (optional) - A BFAST container of additional data files associated with the scene.

The specific encoding of geometric, material, or texture assets is not specified by the format.

Why not OBJ?

OBJ is inefficient to read and write, is opinionated, and does not enable encoding of instancing of geometry.

Why not Collada (DAE)?

Collada is a very complex and bloated file format that is inefficient to read and write.

Why not glTF?

The glTF is a complex and opinionated format that dictates how different assets (e.g. meshes, morph targets, materials, etc.) are stored.

Writing a compliant importer, exporter, or validator for glTF is a massive undertaking. For example, see

Why not FBX?

Like glTF FBX is also a complex and opinionated format, but additionally is closed making it unusable for web applications.

Why not USD?

The USD file format does not have a specification. You have to use the USD library or recreate its reading and writing. There

About

A simple and efficient file format for storing and transmitting 3D scene data (geometry, scene graph, assets, etc.)

License:MIT License