> I began modifying the finish definitions of the standard metals.inc. The
> pigment were right to me, but they looked really too dark with the new
> finishes. So I normalized every color vector ( < r, g, b > / vlength (< r,
> g, b > ) ) and then I multiplied it by sqrt(3) in order to get every color
> vector as bright as <1,1,1> ( if a normalized vector is one unit long and
> you multiply it by sqrt (3) you have a color vector as long (i.e. bright) as
> <1,1,1> since vlength (<1,1,1>) = sqrt ( 1^2 + 1^2 + 1^2) = sqrt (3)). Then
> I lowered each value to get the right brightness. Here follow some (badly
> compressed :-\  ) examples. As an aside, if you want the whole .inc file
> just ask.
> Cheers,
>
> --
> Jonathan.
>

  OK ... I just don't see the difference in doing that and simply
multipling <r,g,b> by some factor greater than one to increase
brightness. I guess your method makes it more consistant . . . as
you begin with <1,1,1> regardless of the pigment shade and the
value you use - sqrt(3*x) - gives the relative brightness.
  If you will share your inc file, perhaps I'll be able to see what
you are doing by comparing a few of them and working with them
until I can "see" it ;)

Nice example scene . . .
Thanks . . .
--
Y

Post a reply to this message