This project was done as part of the Computer Graphics module by Prof. Dr. Mario Botsch at the Bielefeld University in the summer term of 2011. It is an image manipulation program which implements its algorithms in shader programs running on the GPU.
We aimed to make this image manipulation program flexible and extensible.
Therefore, we decided to save the source code of the shaders in their own files
which will be loaded and compiled at runtime. This makes it easy to add new
algorithms: just write a shader and you can load it from the program.
However, it would be nice if one could control the various parameters of the algorithms with GUI controls. But a normal shader program does not provide the information what the purpose and valid range of a parameter is. This is the point were the Annotated Shading Language (ASL) comes into play.
The Annotated Shading Language allows to add information like description and parameter range of a parameter to the program. Because all ASL elements are written within comments, the shader program can still be compiled without using an ASL shader compiler.
Because ASL was developed only as part of a rather small project, there is still room for many enhancements. For example, at the moment it is only possible to annotate uniforms and whole files, but not per vertex parameters. We had no need for this function as we only used fragment shaders which do not have per vertex parameters.
If you want to try out ASL for yourself, take a look into the doc
folder where
you will find a documentation of ASL.
The program should be platform independent. However, the final version was only tested with Mac OS X 10.6 (Snow Leopard) at the time of writing this.
- Not all shaders will work with Mac OS X 10.7 (Lion). It seems that Qt messes
up the initialization of the OpenGL context on that platform. Presumably
because Lion introduced OpenGL 3. The problem may be fixed with Qt versions
4.7. However, it still persisted in the beta version of Qt 4.8.
- Jan Gosmann
- Denis John
- Markus Kastrop
The project has been supervised by Matthias Heinrich and Prof. Dr. Mario Botsch.