On 11/12/23 09:49, Mike Miller wrote:
> There is no 'box' mapping in POV.
Not to PILE ON :-)... There are at least five (six in yuqk) and probably
more, but as Bill W. pointed out many require a special images for the
Somewhere, I have a couple of user_defined cameras for creating the
standard map forms for use with the inbuilt three box mapped methods.
The second, more complicated camera tries to create images which should
stand up better where users turn on interpolation.
Today, image interpolation at the edges of the standard box image
folding cut out often introduces artifacts(a).
(a) You have used other render tooling, yes? If so, do you know how
those tools handled interpolation of the images at the edges of the cut
out regions for different parts of a uv mapped mesh?
Do they not use interpolation at all?
Are the images cut out regions created with extra buffer pixels to
enable clean interpolation (smoothing) on the fly?
Or, do they just use really high resolution images for uv mapping so
interpolation doesn't much matter and so they don't use it? If yes, I
might guess by avoiding interpolation while rendering and on every 3D
reach into the image map, they are gaining some performance this way?
This last I ask because I see POV-Ray folks often using interpolation
with image maps used with uv mapping and I wonder if we should not be
creating per-interpolated versions at higher resolutions for repeated
Post a reply to this message