 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I have a problem obtaining nice sharp edges.
I have tried several antialiasing methods and huge resolution renderings
(something like 8000x6000). I still get jagged edges. Mostly when overlapping
shapes. If shape is directly over background, the edge is quite nice.
Please take a look at attached samples.
I have used INI settings:
Green spikes image: http://img229.imageshack.us/my.php?image=spiky2zo7.jpg
Antialias=On
Sampling_Method=2
Antialias_Threshold=2
Antialias_Depth=2
Test_Abort_Count=400
Green blob: http://img223.imageshack.us/my.php?image=002005gx6.jpg
Antialias=On
Sampling_Method=2
Antialias_Threshold=5
Antialias_Depth=1
Test_Abort_Count=500
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Queen soft nous apporta ses lumieres en ce 2008/02/09 14:04:
> I have a problem obtaining nice sharp edges.
> I have tried several antialiasing methods and huge resolution renderings
> (something like 8000x6000). I still get jagged edges. Mostly when overlapping
> shapes. If shape is directly over background, the edge is quite nice.
> Please take a look at attached samples.
> I have used INI settings:
> Green spikes image: http://img229.imageshack.us/my.php?image=spiky2zo7.jpg
> Antialias=On
> Sampling_Method=2
> Antialias_Threshold=2
A very high treshold value. Almost every pixels will stay under that difference.
> Antialias_Depth=2
> Test_Abort_Count=400
Over the white background, the difference is large enough to trigger the
antialias process, but elsewhere, the difference is just not large enough.
You probably want something more like 2/256 (0.0078)or 2/(256*3) (0.0026).
>
> Green blob: http://img223.imageshack.us/my.php?image=002005gx6.jpg
> Antialias=On
> Sampling_Method=2
> Antialias_Threshold=5
An absurdly large treshold value, NO pixel will exeede that value.
> Antialias_Depth=1
> Test_Abort_Count=500
Here, the treshold is so large that antialias never have a chance to kick in.
You probably want something more like 5/256 (0.0195) or 5/(256*3) (0.0065).
>
>
>
>
In POV-Ray, the rgb values range from zero to 1. The difference between 2 pixels
is calculated as the sum of the absolute values of each channels. That
difference normaly stay in a 0 to 3 range.
In your examples, you probably want a difference of 2 on a scale that goes from
zero to 256.
The "default" antialias treshold is of 0.3.
Set Antialias_Threshold to something like 0.1 and you will get much beter results.
--
Alain
-------------------------------------------------
You know you've been raytracing too long when you've tried to scan your face for
a texture.
Quietly Watching
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Queen soft wrote:
> Test_Abort_Count=500
Did you actually read the documentation about this value? - It has
absolutely nothing to do with anti-aliasing at all!!!
Also, your other anti-aliasing settings are not very sensible as they are
out of range to get good results (they practically disable anti-aliasing).
Reading the documentation might be a *very* good idea...
Thorsten
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain <ele### [at] netscape net> wrote:
> Queen soft nous apporta ses lumieres en ce 2008/02/09 14:04:
> > I have a problem obtaining nice sharp edges.
> > Sampling_Method=2
> > Antialias_Threshold=2
>
> A very high treshold value. Almost every pixels will stay under that difference.
indeed! I usually go with the command line +AM2 (method) +A0.06 (threshold) or
lower...
but the AA method which I find gives the best result is camera blur. Just a
slight aperture and enough blur_samples gives excellent results, at least if
you're going for photorealism...
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I usually use threshold 0.3 (the default). However, on green areas
there's still a lot of noticeable jaggies (the human eye is more
sensitive to green, afaik), and I have to lower the threshold (like to
0.1) in those cases.
Like the rest said, anything above 1.0 is a lot like disabling
antialiasing...
Now, for a final render, some people use thresholds even lower (0.01).
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nicolas Alvarez napsal(a):
> I usually use threshold 0.3 (the default). However, on green areas
> there's still a lot of noticeable jaggies (the human eye is more
> sensitive to green, afaik), and I have to lower the threshold (like to
> 0.1) in those cases.
>
> Like the rest said, anything above 1.0 is a lot like disabling
> antialiasing...
>
> Now, for a final render, some people use thresholds even lower (0.01).
JPEG compression does the AA for me ;-)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Jan Dvorak wrote:
> Nicolas Alvarez napsal(a):
>> I usually use threshold 0.3 (the default). However, on green areas
>> there's still a lot of noticeable jaggies (the human eye is more
>> sensitive to green, afaik), and I have to lower the threshold (like to
>> 0.1) in those cases.
>>
>> Like the rest said, anything above 1.0 is a lot like disabling
>> antialiasing...
>>
>> Now, for a final render, some people use thresholds even lower (0.01).
>
> JPEG compression does the AA for me ;-)
:-) Actually, did you know that anti-aliased images slightly improve
compression and quality when using JPEG?
Thorsten
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thorsten Froehlich wrote:
> Jan Dvorak wrote:
>> JPEG compression does the AA for me ;-)
>
> :-) Actually, did you know that anti-aliased images slightly improve
> compression and quality when using JPEG?
More than slightly, depending on the specifics of the picture. AA also
greatly increases the quality and compressibility of MPEGs, too.
Regards,
John
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
John VanSickle wrote:
> Thorsten Froehlich wrote:
>
>> Jan Dvorak wrote:
>
>>> JPEG compression does the AA for me ;-)
>>
>> :-) Actually, did you know that anti-aliased images slightly improve
>> compression and quality when using JPEG?
>
> More than slightly, depending on the specifics of the picture. AA also
> greatly increases the quality and compressibility of MPEGs, too.
Indeed, because AA "decreases" the frequencies the DCT (discrete cosine
transformation) needs to represent.
Thorsten
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thank you very much for your quick responses.
I don't know why I was increasing the threshold !!! With values like .01 and
..001 it works just fine !!! Thank you again.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
