Diary/Week 11

Saturday afternoon

Really, it's embarrasing. It's been seven weeks of this course, and I still haven't written anything more than my PM. I haven't read my material. And Tuesday is seminar 2. Let's take a look at my original time plan:

Feb 12 First seminar; presentation of introduction, hypothesis and plan

Mar 09 Researching the subject thoroughly, finding more models than the image model, hopefully making ray tracing unfeasible without having to implement it (someone must have researched it before)

Mar 30 Implement the image model

Apr 05 Evaluate the image model with profiling tools

So. Eight days ago I should've been done reading all my material at least once. Ugh, that means that seminar 2 actually *was* moved, which means the mail I sent to Olle a few days ago is even more embarrasing...

Thirteen days to implement the image model. I guess I can do it, but I'll have to cancel my projects outside school (horrendously stupid to try to work during the last term).

By the way, reading Savioja's paper, it seems that ray tracing is actually easier to use for a reflection approximation than the image model, since the image model is actually complete (I knew that from before, but just haven't thought about it) while ray tracing isn't (it's truly approximative). If I use the image model, I probably can't get late reverberation because the model gets too complex then. However, with ray tracing, it seems that one can even get late reverberation without the model being too complex. I really don't feel like changing my focus now so I'll keep working on the Image model...

So. What's my plan of attack? Hell if I know... Spend today reading all I can, spend tomorrow rounding off old non-school projects so my clients can stop bugging me, and spend monday reading even more, writing a little on my paper, and preparing a presentation for Tuesday.

Long term? I... don't... know...

Saturday evening

Okay, so in ray tracing sound, it's enough to do specular reflections. Which is very very much easier  than the diffuse reflections that Swartling said it would take -- linear instead of quadratic. Raytracing is definitely possible.

Sunday afternoon

Read Savioja's paper now. Interesting read. Lots of notes. Need to consolidate into my paper somehow.