The 80/1.7 GF on the GFX 100S has *more* DOF than calculators say

Started 2 months ago | Discussions thread
MOD Hamiltionian Contributing Member • Posts: 929
Re: At f/2.8...
1

My instinct is to suggest that the order of the blur and sharpening operations could matter, but after playing around in photoshop I can't demonstrate that in any noticeable way. Sharpening unlikely to be the issue here.

The order in not important in linear convolution, just like in multiplication. In fact, convolution in the space domain is equivalent to multiplication in the frequency domain. However, the demosaicing process as performed by Lr is nonlinear. I’ll do some tests.

Gotcha thanks. Haven't spent time on this before, which now seems an oversight. It wasn't immediately obvious to me that sharpening was a linear convolution. And then yes, linear convolutions commute. For others following along:

https://en.wikipedia.org/wiki/Kernel_(image_processing)

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

Of course, I didn't have image of what the lens was putting down on the sensor at 600 nm sampling resolution. I just had the image captured at 3.76 um sampling resolution. I did enlarge that to 900% before doing the CoC simulation, and I used bilinear interpolation for that enlargement. Maybe I lost some sharpness there.

Maybe, seems pretty reasonable to me though and unlikely to cause a significant loss in image sharpness.

-- hide signature --

Stay Calm and Carry Cameras
https://thecentercolumn.com/

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