[Soc-2008-dev] Week 12 report

Davide Vercelli davide.vercelli at gmail.com
Sat Aug 16 00:24:06 CEST 2008


Week 12
=======

This week
---------
The main goal for this week was to have a proof-of-concept
implementation of instant radiosity. This is not yet in a testable
state, as it is currently working only for area lights and there's no
color bleeding going on; I'm working on fixing these limitations and I
hope to have some interesting renderings to show by the sunday
meeting.

I fixed the broken Oren-Nayar shader: the previous bound was too loose
so that the algorithm would never be able to go under the error
threshold.This is interesting as it shows how important it is to have
tight bounds for each shader, and that's not always easy to achieve.

Even though the current UI is not meant to be "final", nevertheless I
simplified it in a "user friendly" direction: it is now possible to
directly specify the total number of point lights that will be used to
convert all the area lights in the scene. This was previously awkward
as it required to specify a non intuitive density per area unit; it
was also very easy to end up having absurd numbers.

Next week
---------
1 - bring the instant radiosity implementation in a more
"interesting"/testable state
2 - write a full "state of the project" report on the wiki (here:
http://wiki.blender.org/index.php/User:UncleZeiv/SummerOfCode2008/Status,
still empty). This will include details on current limitations, plans
on future action and possible future extensions. I will also improve
my integration document
(http://wiki.blender.org/index.php/User:UncleZeiv/SummerOfCode2008/Midterm),
but I still have to make up my mind on several issues

Issues
------
I'll list all of them in the final report.

Schedule
--------
I think I managed to deliver more or less what I planned to deliver at
the beginning of the project. I am aware that even though the code is
pretty stable, is not yet mergeable in trunk as is, partly because of
missing support to many internal renderer features, and partly because
it became apparent that it's not clear how to cleanly integrate it
from a workflow point of view. More on these topics in my reports.


More information about the Soc-2008-dev mailing list