April 04, 2006
Aperture Downsampling Results
Most of the time we want more resolution in our images so they can be printed on bigger and bigger formats. However, for web publishing it's important to have quality downsized version of our images. A question about how Aperture does this was raised on the Aperture discussion group about image downsizing. I used a few test images from Oshyan Greene's website to find out.
Here are the results. It's hard to tell exactly what algorithm Apeture is using, but it's probably bilinear filtering since it's very easy to do using the GPU. However, many other kernels could be implemented using pixel shaders.
Please compare these to the other results on Oshyan's site. After all, once Aperture 1.1 comes out people will need something to freak out about other than RAW conversion. Oh, and remember that resolution test images are not always good representations of what the algorithm will look like with normal images. However, it does give you a good idea of the general behavior of the algorithm.
Please compare these to the other results on Oshyan's site. After all, once Aperture 1.1 comes out people will need something to freak out about other than RAW conversion. Oh, and remember that resolution test images are not always good representations of what the algorithm will look like with normal images. However, it does give you a good idea of the general behavior of the algorithm.
The result of downsizing a 445x446 test image to 25% original size in Aperture.
A real-life image reduced from 3072x2304 to 1024x768 in Aperture.
A resolution test image downsized from 9344x5842 to 1023x640 in Aperture.
Posted by sean at April 4, 2006 10:12 PM