scott wrote:

> As Michael mentioned, look at Blender, it is biased more towards 
> animation that actual design work, but still allows all the things I 
> mentioned + lots of animation features.  Another example is modelling 
> the suspension of a car.  For simplicity let's say the only input is the 
> relative height of the wheel compared to the body.  In POV you would 
> need to do the math to align and rotate every component to fit 
> together.  In Blender (or another CAD/modelling package) you can tell it 
> which bits are "fixed" (using the mouse) and then it automatically does 
> all the transforms for you.

Well, there are physics simulations of various kinds for POV-Ray too. 
(And this kind of thing would be much easier of SDL supported 
reflection, proper control structures and data types, etc.)

IIRC I did look at Blender once (or was it Wings?). As far as I can 
tell, it only edits triangles. And as we all know, that's a very hard 
way to get anything done. (Just like editing individual pixels is a very 
hard way to draw a realistic RSOCP.)

>> I was thinking more that once you write a script to take a sequence at 
>> 250 frames per second, add motion blur, and convert down to 25 frames 
>> per second, you can then process lots of videos quickly.
> 
> Yeh, AVIsynth would be perfect for that, actually it's what I mostly use 
> it for (converting POV output to video, and sometimes overlaying it on 
> top of real video).

For fast-moving things, I often render at 2x or 4x the normal framerate, 
add some motion blur, and then reduce the framerate back to normal. It's 
meant to work as a time of temporal antialias. Seems to work moderately 
well. (For good results, you probably need more than 4x. But it takes so 
damn long.)

Similarly, if I'm generating frames from something that doesn't support 
spatial AA, you can always render large and resize... ;-)

[Actually, POV-Ray's adaptive AA can twinkle annoyingly if you don't 
turn the sensitivity up high enough.]

> The way I found AVIsynth is through the h264 codec, x264.  It takes 
> AVIsynth files as an input, which is perfect to convert from POV output 
> to .mp4 files in one step, even including motion blur if you like.

I prefer to keep all my videos losslessly compressed until just before 
they go onto the DVD, but yeah. ;-)

It's kinda frustrating that everything has to be at such a low 
resolution to go onto a DVD. But then, when you watch it on TV... well 
let me put it this way. I put a test render on there at half resolution 
once. It looked horrid on my computer screen. But on the TV, I honestly 
couldn't tell the difference...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message