|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi,
What is the maximum level of AA that can be used when rendering? The
standard in the .ini file is 0.3 but can it be higher?
Thanks.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"the_ajj" <nomail@nomail> wrote:
> Hi,
>
> What is the maximum level of AA that can be used when rendering? The
> standard in the .ini file is 0.3 but can it be higher?
>
> Thanks.
The value is a threshold, expressed as a ratio ranging from 0 to 1. The
lower the value, the higher the "level" of AA. With a value of 0, each and
every pixel in the scene is antialiased, whatever its contrast against its
neighbouring pixels. With a value of 1, only a perfectly black pixel
against a perfectly white one (or the opposite) would be antialiased.
If you have a strong pattern in your scene, such as wood, you might want to
soften moire artifacts by using jitter_amount greater than 1. Basically it
tells antialiasing to work not on one pixel at a time but to an area
slightly larger, thus "blurring" somewhat possible artefacts.
Just to be clear, please note that AA is not at all a post-process blurring
or softening process a la Photoshop. It actually makes Pov Ray compute more
information to avoid jagginess.
Hoping this helps!
JYR
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
the_ajj <nomail@nomail> wrote:
> What is the maximum level of AA that can be used when rendering? The
> standard in the .ini file is 0.3 but can it be higher?
JYR's answer was good but incomplete.
The 0.3 specificed with the antialiasing parameter is a threshold value.
It means, AFAIK, that if the current pixel color differs 30% or more from
any of the neighbour pixels, then antialiasing is performed to that pixel
(by shoothing more rays).
If you want antialiasing to be calculated more easily you have to lower
that threshold value. For example 0.1 means that the difference has to be
only 10% for antialiasing to be calculated. 0.0 naturally means that
antialiasing is always calculated for each pixel (this can be very slow,
but in some cases it's actually feasible).
Now, the threshold is not the only thing you can tune.
Firstly, there are two antialiasing types from which you can choose,
and secondly, you can specify the number of extra rays traced for
antialiased pixels.
Antialiasing method 1 (the default) simply traces a fixed amount of
extra rays (which you can specify), averages the results and that's it.
Antialiasing method 2 adaptively traces more rays (down to a certain depth,
also specified by the same option). There's a great graphical explanation
of antialiasing method 2 here:
http://povray.org/documentation/view/3.6.1/223/#s02_01_02_08_04
The number of extra rays traced with method one, or the maximum subdivision
depth of method two can be specified with the Antialias_Depth or +R option.
With method 1 the square of the given value is used as the number of extra
rays. The default value is 3, which means that 9 rays are traced. If you
specify eg. 4 it means that 16 rays will be traced. The URL above explains
well what this value means with method 2.
There's one antialiasing option left: Jittering (+Jn.n or Jitter_Amount).
If you specify this, the extra rays will be randomly jittered by the given
amount.
--
- Warp
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Warp wrote:
> It means, AFAIK, that if the current pixel color differs 30% or more from
> any of the neighbour pixels, then antialiasing is performed to that pixel
> (by shoothing more rays).
Actually those 30% are misleading. The criteria is if
Colour_Distance_RGBT() is larger than the aa threshold and
Colour_Distance_RGBT() calculates the L1 norm of the difference between
the pixel colors.
Christoph
--
POV-Ray tutorials, include files, Landscape of the week:
http://www.imagico.de/ (Last updated 31 Oct. 2005)
MegaPOV with mechanics simulation: http://megapov.inetart.net/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Actually those 30% are misleading. The criteria is if
> Colour_Distance_RGBT() is larger than the aa threshold and
> Colour_Distance_RGBT() calculates the L1 norm of the difference between
> the pixel colors.
Yes, I was thinking this as I read the above replies. This means that useful
values for the threshold are between 0 and 3, or 0 and 4 if alpha output is
on.
- Slime
[ http://www.slimeland.com/ ]
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
That's great - thanks a lot everyone !
Adrian
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |