For David Millier

Started Feb 4, 2013 | Discussions thread
FDecker
Contributing MemberPosts: 665
Like?
Re: For David Millier
In reply to DMillier, Feb 7, 2013

Can you take a look at this image in another thread:

http://forums.dpreview.com/forums/post/50802149

which compares the output from two Merrill cameras with a NEX 7.

I see what I would expect to see: jaggies all over the place that I attribute to the absence of an AA filter and a nicel clean result from the NEX that I attribute to the presence of an AA filter. Maybe this is just unconscious bias on my part and if this is an incorrect conclusion to draw, would you mind explaining what is really going on in these frames.

Hi David,

you are absolutely right. The jaggies result from the discrete sampling and the fact that the Sigma sensor doesn't use an AA filter. For low spatial frequencies (no aliasing problems) this AA filter still works as a smoothing filter which hides the jaggies behind some "fuzziness" of the resulting image which is blurred by the AA filter so that any detail hits more than one pixel. Sharpening in firmware or software on the other hand partly surpresses this fuzziness.

Of course, you could do the same with a Foveon image. It would even be easier and more defined (could even be adaptive to the level of detail in a respective image area) to do this in software based on shard unblurred pixels. Unfortunately, you introduce the same loss of micro-contrast into the image this way. But, definitely, it is worth to start from "clean" crisp images. You can apply software algorithms to do this or that.... But you can't recover detail which is lost by an AA filter.

Regards,

Frank

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow