Canon Line Offset Noise

Started Oct 20, 2009 | Discussions thread
Steen Bay Veteran Member • Posts: 7,418
Re: That'a not what I mean

GaborSch wrote:

Steen Bay wrote:

So, the demosaicing process in both DPP and ACR creates the maze artifacts in the attempt to maximize the amount of detail, and those artifacts makes NR (which decreases the amount of detail and artifacts) necessary, because otherwise the sharpening will make the pics look terrible. Is that it?

What I meant is, that the very normal noise appears as maze if sharpened. Look at hulster's http://forums.dpreview.com/forums/read.asp?forum=1019&message=33442216 (and read my answer to that).

What I meant is, that normal random noise doesn't appear as maze if sharpened. The sharpening doesn't create anything that isn't already there, it just increases the contrast and makes it easier to see. The maze must be created by the demosaicing process. But it's correct that the maze is much less visible in Hulkster's new DPP pics without sharpening. It's also interesting to see that the ACR pics (without sharpening?) now have more detail/sharpness than the unsharpened DPP pics.

Now, that I have DPP 3.7, I could experiment and I see, that DPP eliminates the banding very well, and I don't see maze at all (though I have not tried all images I collected). I don't know if DPP uses some "secret knowledge" (most of the 45 KB MakerNote is unknown outside Canon) for the correction. It is easily possible, that the camera records some data for the correction. Another possibility is, that DPP applies a demosaicing algorythm, which is less prone to mazing but yields inferior result; I don't believe that Canon would do that.

Post (hide subjects) Posted by
xxD
xxD
xxD
xxD
xxD
xxD
xxD
xxD
tko
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow