|
![](/i/fill.gif) |
"the_ajj" <nomail@nomail> wrote in message
news:web.43749ec0ba58206ca43155f60@news.povray.org
> Hi,
>
> I got this comment regarding one of my images 'The Chess Game'
> which can be seen in the images newsgroup:
>
> "so, did you render an 8bit image out of povray?
> Ideally you'd render a float precision image so you can get all the
> out-of-focus hilight shapes just as bright as the in-focus
> hilights."
>
> I have no idea what this means or how I do it. Any ideas?
The problem is, if you apply focal blur as a post process step (as you did),
it cannot take into account any surfaces that are brighter than white in the
image file. This will make a difference around bright highlights etc. For
example, suppose that the focal blur algorithm determines that 5% of an
output pixel should be from the white surface, and 95% from the (black)
background. Using the standard image formats, this pixel would end up at 5%
grey. If you had used floating point precision, then it would be able to
take into account that the white is actually far brighter than white, and
the resulting pixel would be brighter.
Unfortunately POV does not output float precision images. It's easy to
patch it to do this, but then it's likely not going to be compatible with
your post processing.
One idea is to render the scene at (say) 10% brightness, then do your focal
blur algorithm on this image, then make it 10x brighter. You could then
merge just the highlights on top of the original or something.
Post a reply to this message
|
![](/i/fill.gif) |