Development Notes for GRT Raytracer

This file represents a moving target — features of the current stable version, CVS version, plans for the next release, and possible future features.

[Home]

Contents

• GRT 0.1 Features
• GRT 0.2 Roadmap
• Feature Futures
• Project Goals

GRT 0.1 Features

• Affine transformations
• Ambient, diffuse interaction, refraction & reflection
• Patterned filtering transparency
• Patterned non-filtering transparency
• Patterned pigment
• Patterned reflection
• Patterns: solid, checker, noise, gradient, turbulence, perlin-noise
• Primitives: plane, sphere, box
• Render to: PPM-file, window (SDL)
• noise3, turbulence, perlin-noise

GRT 0.2 Roadmap

Release date: August / September 2003

• Bounding volumes
• CSG
• Debian package
• PNG support
• Octree
• Primitives: cylinder, torus, quadrics
• SBCL compatible
• Shader based architecture
• Simulated caustics
• Variable reflection
• Shaders: plastic, metal, glass, laquer

Feature Futures

• Adaptive supersampling
• Alternative highlight scheme (microfacets)
• Animation support: frames, time, temporal antialiasing (motion blur)
• Backwards raytracing (photons)
• Distributed raytracing (focal blur & blurred reflections)
• Dumping & loading of internal data structures
• Efficiency measures: light buffer, vista buffer
• Image formats: JPEG, TIFF
• Imports from foreign file formats (DXF, POV, etc)
• Interrupted / resumed rendering
• Lights: area, parallel, spot, fading
• Map-based patterns
• Non-affine transformations
• OpenGL preview
• Patterns: marble, wood, etc.
• Perturbated normals
• Primitives: cone, metaballs, mesh, SOR, lathe, text, patches, heightfield
• Single-sample soft shadows
• Splines
• Tesselation
• Volume rendering

Project Goals

Our short term goals are easily defined: they are listed as the roadmap for the next release.

An important mid-term goal is to make GRT a good classical raytracer: reasonably efficient, with a good selection of primitives, advanced texturing capabilities, animation support, and probably integration of the scripting environment with some GUI tools.

In the long run we hope that GRT will blossom into a powerful image synthesis system, not "just" a raytracer: it is our wish to make GRT capable of supporting a variety of rendering paradigms.

Simon Adameit: What I want is a superior rendering engine like POV-Ray with cool shaders like RenderMan, and perhaps a GUI which combines the strengths of editing scenes with a text editor and modelling with the mouse - and all that written in lisp ;-)

Nikodemus Siivola: Personally, I consider GRT a blinding success if it just becomes a halfway decent raytracer. Everything else must come after that.