Warp <war### [at] tagpovrayorg> wrote:
> Meenal <mee### [at] studentumassedu> wrote:
> > I am creating 2D projections of a 3D object by shining a light source
> > through it. I want projections at various angles and therefore, I rotate
> > the object to get different projections. The projections look good. My
> > question is on the kind of interpolation that POV applies while creating
> > the projection, i.e. when in taking a projection of ab object from a
> > non-orthogonal views, you get non-integer values of pixels; what and how
> > does POV interpolate those non-integer pixel values? To their nearest
> > neighbor?
>
>   Your question is a bit unclear.
>
>   First of all, could you explain how you think POV-Ray is doing
> the "projection"? If your concept is wrong, it would be good to
> know it in order to correct the misconception.
>
> --
>                                                           - Warp

When I shine a light through an object, and then render the image, I get a
2D image. This image is a black and white image with white being the pixels
where the objects are. My question is that how does POV decide what pixels
will be white and what will be black? Say that a part of an object does not
completely cover an entire pixel - what is the value of the pixel
considered to be? Is the pixel interpolated to its nearest neighbor - i.e.
say morethan half of a pixel is occupied by an object, will the entire
pixel be considered an object? How does POV deal with parts of pixels? How
does it interpolate those?

Meenal

Post a reply to this message