Resize Image: Resampling methods

Cataratta · Post by **Cataratta** » Sat May 19, 2007 3:15 pm

In this menu there is a resample method menu.
(Neares,bilinear,hermite,gaussian.....)
OK the first is the worst , but someone can tell me the main difference of various methods ?
For example if a method is better than others for make smaller a photo. or another is better for enlarge etc..
Tanx for support.

helmut · Post by **helmut** » Sat May 19, 2007 4:20 pm

In the forum there are various topics and posts, already:

- Info about Resampling algorithms
- FAQ "How to resize images with text (screenshots)".
- Sample page of Lasse Kolb
- Image Resampling by Jonathan Sachs (link provided by Olivier)

Guest · Post by **Guest** » Mon May 21, 2007 10:13 am

Tanx a lot

Drahken · Post by **Drahken** » Mon May 21, 2007 10:09 pm

Quick answer: Lacnzos and hanning give the best results. Nearest neighbor gives the worst results because it doesn't actually use a filter at all. However, if you're resizing a 256 color image (like a gif or some pngs) and want to keep it 256 colors, or if you are resizing an image with transparency and you want to keep the transparency, then you need to use the nearest neighbor one any other one will change the image too much.
The other filters are pretty much useless.

Fore more detailed info, visit the links already posted.