On 12/05/2010, at 01:22 , malefico wrote:

> In 2.4x, you could animate the material of a particle, let's say, Halo 
> Size, Alpha, etc, and this was mapped to the particle life. So if you 
> animate from frame 1 to 100, you would have that range mapped into the 
> particle life. This is great to make particles fade out when they are 
> about to die, etc.

It's a limitation of the animation system, or rather, the old method was a bit 
of hack, providing some nice functionality, but not fitting in too well. From 
memory, the old method evaluates the curve separate to the rest of the 
animation system and doing it per particle, rather than the entire material - 
although it was handy in that case, it's inconsistent, prevents the use of 
drivers etc and can cause more problems now in a more comprehensive and well 
integrated animation system.

An idea that came to mind previously is to do present this functionality as a 
texture, or probably better, a shader node - outputting all kinds of particle 
data (lifetime, velocity, mass, etc) as variables into the node tree. This is a 
lot more consistent with other types of data (eg. other per-point/face geometry 
data such as UVs) and would be a lot more flexible, you could do whatever you 
want with it, not to mention what you could construct with nodes (colour ramp 
widgets for example) would be a lot easier in many cases than RGB f-curves. In 
the proposed new shading system, it could come bundled up in a default 
'particle shader' too, so it's all just there waiting.

Anyway, probably best to tackle in the new shader system imo. I do think that 
going down the old path of trying to wedge this into the animation system is 
probably not a good idea.

cheers,

Matt
_______________________________________________
Bf-committers mailing list
[email protected]
http://lists.blender.org/mailman/listinfo/bf-committers

Reply via email to