> racing manually or coding ai to race?

   Manually: making racing robots should be very fun, but I don't have the
skills at all...

> I've some old code where I was using povray to visualise my (very) 
> primitive ai's idea of the racing line.
> 
> Each time the ai was started it would generate an include file containing
> the geometry of the track and a couple of iteration's of the generated
> racing line.

   That's very interesting...  I will kill for a TORCS robot that would
generate POV-Ray splines: is this what you did?

> Your images are as usual stunning - what's the render time like - would 
> video clips of races be possible?

   It renders decently fast if you don't trow focal blur and area light at
once... in the end, everything are meshes. Let's say from 15 to 30 min for
most the images. But what is stopping me from making an animation is not the
render speed, but the lack of telemetry data to control speed, acceleration,
braking, etc... and of course, the skills to manage it all and translate it
to SDL... to not talk about the free time needed. :(

--
Jaime

Post a reply to this message