POV-Ray : Newsgroups : povray.binaries.images : Ratty text, further explored, with superfluous big images : Re: Ratty text, further explored, with superfluous big images Server Time
9 Aug 2024 17:22:59 EDT (-0400)
  Re: Ratty text, further explored, with superfluous big images  
From: Slime
Date: 6 Jan 2005 22:18:05
Message: <41ddff6d@news.povray.org>
> 1) 320x240 AA 0.3.  The original image, the way I had been doing things
> before. There is a rattiness to the box which detracts from the
> professionalness of the product.
> 2) 320x240 AA 0.3 -J.  Anti-alias jittering turned off.  The problems with
> the box have gone away, but I believe it detracted from the quality of the
> images, especially in showing of round edges. (Yeah, this image here seems
> to make me a liar, but I started this when I saw a jagged edge on another
> field of view.) The bumpy texture of the guy's yellow shirt looks off, and
> I'm guessing this would be sloppy in animation.

Although turning off jittering slightly decreases the quality of a still
image (since randomness is good for our eyes), it will *increase* the
quality of an animation. Ever seen an animation of a checkered plane? Look
at the horizon: even if the camera is still, if anti-alias jittering is on,
it will look like static. This is because the random jittering changes from
frame to frame. So, if animation is what you're heading towards, try it in
an animation before you write it off as a bad idea.

Besides, the guy's shirt looks almost exactly the same to me in the first
two images.

> 3)  640x480 AA0.3, then resized in GIMP.  The text box looks great, and
the
> characters aren't hurting.  The nusiance of this approach is of course
that
> it requires an image resizing.

Except for color clipping (which is irrelevant in a scene without very
bright objects), this is equivalent to 320x240 with +A0.3 +R6. (The default
is +R3, +R6 is like that, four times, averaging the result - which is just
what you did in GIMP.)

> 4) 320x240 AA 0.3 +R3 +AM2, increased AA depth with antialias method 2.
Text
> box is great, but we've "erased" the poor lady's eyebrows!!

+AM2 starts by sampling the *corners* of pixels instead of the centers.
Probably her eyebrows pass through the center of pixels but not the corners,
so they're missed. This is a common problem (happens a lot with ropes or
other thin objects in the distance) that happens with *both* AA methods, and
has no good solution at this time.

> Any advice here?

Based on what you were happy with and on my experience, I'd suggest going
with +A0.3 -J and something between +R3 to +R6.

> B) What are the possibilities of a pov feature that turns off AA for
> specific objects:

Personally, I think it would be more convenient to have an object property
that forces *higher* AA (more subsamples) for the current pixel when it's
intersected. This could be used to solve the eyebrow issue.

 - Slime
 [ http://www.slimeland.com/ ]


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.