利用者:Mx/GSoC2008 Week11

GSoC 2008 - Week 11 report

This week

I successfully removed the lib3ds binding this week: Freestyle now uses the render database to import the scene data. As Freestyle inputs triangular meshes, the vlak-based shape representation was utilized for building up the geometrical data structures.

As of today, Freestyle runs natively in Blender without the use of any of its former libraries (Qt, SWIG, lib3ds). In this sense, we can consider the integration as finished. Of course, much more work will need to be done post-GSoC to implement the remaining work listed in the proposal.

Next week

As I have mentioned at the last meeting, I will be off to SIGGRAPH for all of next week so I will not be able to work on the last week of GSoC. I will not be present at the meeting this Sunday and there's a chance that I might miss the one next week too, traveling back home from the conference on Sunday August 17.

Issues

Only vlak-based shapes are supported. That means that strands or halos cannot be rendered by Freestyle yet. I am interested in any information about how to transform these structures to the vlak format.

Schedule

Looking back at the original project proposal, I have implemented phase 1 and phase 2 (without the integration with the internal renderer). Phase 3 will be completed post-GSoC. Fortunately, I have been able to meet the final project requirements set by Jean-Luc.

A lot more work will be needed to refactor the code produced during the summer. Style modules are not 100% compatible and some minor tweaks are necessary to achieve this support. More testing should be done to further validate the code produced during the GSoC.

Overall, I am satisfied with the amount of work I have put into the project. The main mistake I made is underestimating the amount of work needed to replace SWIG: this is why the project schedule had to be pushed back. Fortunately, the work paid off and Freestyle runs natively without any external library. After meeting with Ton and discussing the best way to integrate the stroke rendering directly in Blender's internal renderer, I'll have a clearer picture how the remaining work required to finish implementing the original proposal.