Getting the Correction Factor right...
Posted: 24.04.2015 17:35
I've been playing. This is what happens when you're a geek and don't have a life.
All my examination of the TIF output of the stacks was in Bridge CC with no sharpening, at full screen, at 100% on my calibrated 30" monitor. Images are 7360 x 4912 px.
I took my Nikon D810 with the Tamron SP 90mm macro. I initially tried a grid of lines and placed it parallel to the sensor, and shot as the manual said at f2.8. At CF1 I got no out-of-focus (OOF) bands. At 5.6 none. At 8 none... I finally decided I couldn't have the target parallel to the sensor and set up so the camera was 45 degrees from the target.
Started shooting the 10-shot stacks with CF1 and went through 8. At CF5 I could JUST start seeing a slight banding, but my total focus movement was so small it was difficult to tell. So, I changed to f5.6 instead of f2.8 and changed my target to one of the back-focus targets with markings from +2 inches to -2 inches and added vertical lines every 1/4 inch so it would be easier to see the OOF bands. I don't shoot stacks at f2.8, so 5.6 was more realistic for me - I generally use 5.6, 8, or on occasion f11.
Eventually, after looking at a lot of stacks, I decided that at 90mm a CF of 4 was the highest I could go without seeing any OOF bands. This gave an interval of 10.
THEN I switched to my 28-300 and did the same thing, starting at 90mm again. Got the same results, so I decided any lens on my body at 90mm would provide a CF of 4. Then I went to 180mm and shot with CF1, 2, 3, 4, 5, 6, 8, 10, 12, 14,16, and 18. And got twice the CF before I started getting bands. With CF between 8 and 10, things looked good, with 10 just starting to show OOF bands. At CF12 they're quite visible. On my camera, a CF8 also equals an interval of 10.
Switched to 270mm, which if I understood what was happening would move my CF another 50% (presuming it went something like 90mm CF4, 180mm CF8, 360mm CF12)... Tested again using CF 8, 10, 12, 15, 16, 18 and found that indeed a CF of 10 appeared to work. CF11 might also work but at CF12 I just started to see very slight banding.
I expect that if I tested at 45mm I'd get a CF of 2, but since I've never shot stacks at a focal length that short, I didn't test it.
BUT, I didn't get FEWER images when doing a "real" stack because the interval at different focal lengths changed. For example, to cover +2" to -2" at 90mm and f5.6 might give me a A - B interval of 50. Doing the same thing at 180mm gave me a higher value for interval between A and B (maybe 60 or 70). So, my CF went from 4 to 8, but the "DOF in interval values" stayed the same (10), and the number of shots increased because the overall interval increased and each focus adjustment was relatively smaller as the focal length increased. At least that's what I looked like to me.
In the end, the short answer is that for my D810, with a fixed 90mm I use a CF of 4. With a zoom, I have a linear graph that lets me pick a CF that's right for the focal length at which I'm shooting with CF4 at 90, CF8 at 180, and CF10 at 270mm...
THEN I tried it with a friend's Canon 5D Mk III with a Canon 100mm macro. TOTALLY, COMPLETELY DIFFERENT. With the Canon "Size of focusing steps" set to Big, we got OOF bands at CF1. Switched to "Small" and CF1 was ok, but we'd get OOF bands at CF2. So, for that system, leaving the CF at 1 and just using the interval provided by the DOF function appears to work better.
We tried the SAME thing with a Tamron 28-300 on the 5D MkIII. Got the same results. Values were different, but the bottom line was that for that camera, it appeared that leaving it at CF1 worked best. At any higher value the intervals got large enough to see visible OOF bands.
Needless to say, your results may vary DRASTICALLY from mine - I'm certainly no testing expert and was simply trying to get some empirical data to help me be more systematic when I'm sitting in a bog.
My next effort will be testing my Nikon 70-200 with the extension tubes. I frequently use this combination for close-up/macro work because it's a lot more flexible than the fixed 90mm macro. It'll be interesting to see what effect the extension tubes case with respect to the focal length....
Now if the stinkin' weather will just improve enough for a spring flower to stick it's head out of the ground, I can try it for real!
Stas - just curious. Does what I've been doing make sense in terms of getting correction factors for zoom lenses? I didn't change the Nikon "Size of focusing steps" in the preferences from 15 other than to try one test at 150 and got GIGANTIC steps between shots. So, I put it back to 15 and left it.
All my examination of the TIF output of the stacks was in Bridge CC with no sharpening, at full screen, at 100% on my calibrated 30" monitor. Images are 7360 x 4912 px.
I took my Nikon D810 with the Tamron SP 90mm macro. I initially tried a grid of lines and placed it parallel to the sensor, and shot as the manual said at f2.8. At CF1 I got no out-of-focus (OOF) bands. At 5.6 none. At 8 none... I finally decided I couldn't have the target parallel to the sensor and set up so the camera was 45 degrees from the target.
Started shooting the 10-shot stacks with CF1 and went through 8. At CF5 I could JUST start seeing a slight banding, but my total focus movement was so small it was difficult to tell. So, I changed to f5.6 instead of f2.8 and changed my target to one of the back-focus targets with markings from +2 inches to -2 inches and added vertical lines every 1/4 inch so it would be easier to see the OOF bands. I don't shoot stacks at f2.8, so 5.6 was more realistic for me - I generally use 5.6, 8, or on occasion f11.
Eventually, after looking at a lot of stacks, I decided that at 90mm a CF of 4 was the highest I could go without seeing any OOF bands. This gave an interval of 10.
THEN I switched to my 28-300 and did the same thing, starting at 90mm again. Got the same results, so I decided any lens on my body at 90mm would provide a CF of 4. Then I went to 180mm and shot with CF1, 2, 3, 4, 5, 6, 8, 10, 12, 14,16, and 18. And got twice the CF before I started getting bands. With CF between 8 and 10, things looked good, with 10 just starting to show OOF bands. At CF12 they're quite visible. On my camera, a CF8 also equals an interval of 10.
Switched to 270mm, which if I understood what was happening would move my CF another 50% (presuming it went something like 90mm CF4, 180mm CF8, 360mm CF12)... Tested again using CF 8, 10, 12, 15, 16, 18 and found that indeed a CF of 10 appeared to work. CF11 might also work but at CF12 I just started to see very slight banding.
I expect that if I tested at 45mm I'd get a CF of 2, but since I've never shot stacks at a focal length that short, I didn't test it.
BUT, I didn't get FEWER images when doing a "real" stack because the interval at different focal lengths changed. For example, to cover +2" to -2" at 90mm and f5.6 might give me a A - B interval of 50. Doing the same thing at 180mm gave me a higher value for interval between A and B (maybe 60 or 70). So, my CF went from 4 to 8, but the "DOF in interval values" stayed the same (10), and the number of shots increased because the overall interval increased and each focus adjustment was relatively smaller as the focal length increased. At least that's what I looked like to me.
In the end, the short answer is that for my D810, with a fixed 90mm I use a CF of 4. With a zoom, I have a linear graph that lets me pick a CF that's right for the focal length at which I'm shooting with CF4 at 90, CF8 at 180, and CF10 at 270mm...
THEN I tried it with a friend's Canon 5D Mk III with a Canon 100mm macro. TOTALLY, COMPLETELY DIFFERENT. With the Canon "Size of focusing steps" set to Big, we got OOF bands at CF1. Switched to "Small" and CF1 was ok, but we'd get OOF bands at CF2. So, for that system, leaving the CF at 1 and just using the interval provided by the DOF function appears to work better.
We tried the SAME thing with a Tamron 28-300 on the 5D MkIII. Got the same results. Values were different, but the bottom line was that for that camera, it appeared that leaving it at CF1 worked best. At any higher value the intervals got large enough to see visible OOF bands.
Needless to say, your results may vary DRASTICALLY from mine - I'm certainly no testing expert and was simply trying to get some empirical data to help me be more systematic when I'm sitting in a bog.
My next effort will be testing my Nikon 70-200 with the extension tubes. I frequently use this combination for close-up/macro work because it's a lot more flexible than the fixed 90mm macro. It'll be interesting to see what effect the extension tubes case with respect to the focal length....
Now if the stinkin' weather will just improve enough for a spring flower to stick it's head out of the ground, I can try it for real!
Stas - just curious. Does what I've been doing make sense in terms of getting correction factors for zoom lenses? I didn't change the Nikon "Size of focusing steps" in the preferences from 15 other than to try one test at 150 and got GIGANTIC steps between shots. So, I put it back to 15 and left it.