Anti-Aliasing In Pictures

RightHereBro

Estimable
Dec 13, 2015
4
0
4,510
0
Is it possible using a software ,to make a picture better quality using anti-aliasing techniques like we use in games?
 

If it's a photo, the camera already has a physical anti-aliasing filter which should remove the jaggies. Necessary since most cameras use a Bayer layout (only captures a RGBG grid, and makes up the rest of the colors). There are a couple professional cameras without an anti-aliasing filter for capturing maximum detail, and letting you smooth it later in software.

As for smoothing, the easiest is just a guassian blur (gaussian = fancy way of saying random). It depends on your pic, but usually a 1.3-1.5 pixel radius blur works well. Follow this up with some unsharp masking and you'll usually get something which looks as good as the original but without jaggies. Photoshop's smart sharpen filter is a really good alternative to unsharp masking.

A more methodical way is to resize the picture using bilinear interpolation. That is, rescale it to 200%, then rescale it to 50% to get the original size. Some people argue you should do this twice (200% x2, then 50% x2). Others argue it's better to scale by 125% 5 times, then to scale by 80% 5x. Rather than argue about it, it's usually easier just to try it and see if you like the results.

https://en.wikipedia.org/wiki/Image_scaling

Note that all these methods destroy information - your picture will not have as many details as before. What you're doing is trying to destroy the high-frequency (detailed) aliasing information, which a minimal loss to other high-frequency information.
 

If it's a photo, the camera already has a physical anti-aliasing filter which should remove the jaggies. Necessary since most cameras use a Bayer layout (only captures a RGBG grid, and makes up the rest of the colors). There are a couple professional cameras without an anti-aliasing filter for capturing maximum detail, and letting you smooth it later in software.

As for smoothing, the easiest is just a guassian blur (gaussian = fancy way of saying random). It depends on your pic, but usually a 1.3-1.5 pixel radius blur works well. Follow this up with some unsharp masking and you'll usually get something which looks as good as the original but without jaggies. Photoshop's smart sharpen filter is a really good alternative to unsharp masking.

A more methodical way is to resize the picture using bilinear interpolation. That is, rescale it to 200%, then rescale it to 50% to get the original size. Some people argue you should do this twice (200% x2, then 50% x2). Others argue it's better to scale by 125% 5 times, then to scale by 80% 5x. Rather than argue about it, it's usually easier just to try it and see if you like the results.

https://en.wikipedia.org/wiki/Image_scaling

Note that all these methods destroy information - your picture will not have as many details as before. What you're doing is trying to destroy the high-frequency (detailed) aliasing information, which a minimal loss to other high-frequency information.
 
Thread starter Similar threads Forum Replies Date
F Apps General Discussion 6
K Apps General Discussion 2
U Apps General Discussion 1
A Apps General Discussion 1
G Apps General Discussion 1
john vitz Apps General Discussion 5
M Apps General Discussion 2
J Apps General Discussion 1
S Apps General Discussion 5
F Apps General Discussion 3
A Apps General Discussion 8
A Apps General Discussion 1
A Apps General Discussion 2
J Apps General Discussion 1
A Apps General Discussion 1
G Apps General Discussion 1
stevietrixx Apps General Discussion 19
V Apps General Discussion 3
N Apps General Discussion 2
huilun02 Apps General Discussion 1

ASK THE COMMUNITY

TRENDING THREADS