| Author | Thread |
|
|
11/17/2015 12:03:33 PM · #26 |
Neil, I get some different numbers for DoF (I used DOFMaster on-line calculator). I used 105mm FL and f/10 with a subject distance of 20 feet.
Originally posted by DOFMaster:
Subject distance 20 ft
Depth of field
Near limit 13.9 ft
Far limit 35.6 ft
Total 21.7 ft
In front of subject 6.1 ft (28%)
Behind subject 15.6 ft (72%)
Hyperfocal distance 45.2 ft
Circle of confusion 0.08 mm |
Note that I use two times the pixel pitch for the CoC. This is much more representative of actual results than the "rule of thumb" CoCs that are still based on film.
|
|
|
|
11/17/2015 12:05:58 PM · #27 |
Originally posted by dtremain:
Just a thought. If the lens was telling the camera it was at 103mm, but was really at 250mm, would that confuse the autofocus? I guarantee I wasn't 20 ft. from the car - more like 60 - 80 or more. And I don't think 103mm would be that tight on the car from that distance - even with the crop sensor. |
Well your lens reported the focus are at between 46 to 60 feet away. And that's at f22 or whatever your minimum aperture is. If you were indeed much more than 60 feet away, then the focus area was significantly in front of the subect (and out of the area we can see in the picture.
So I'd say it just misfocused. I'm sure that's not totally uncommon. I've seen a camera/lens hunt and then settle completely wrong. |
|
|
|
11/17/2015 12:10:53 PM · #28 |
Originally posted by kirbic:
Neil, I get some different numbers for DoF (I used DOFMaster on-line calculator). I used 105mm FL and f/10 with a subject distance of 20 feet.
Originally posted by DOFMaster:
Subject distance 20 ft
Depth of field
Near limit 13.9 ft
Far limit 35.6 ft
Total 21.7 ft
In front of subject 6.1 ft (28%)
Behind subject 15.6 ft (72%)
Hyperfocal distance 45.2 ft
Circle of confusion 0.08 mm |
Note that I use two times the pixel pitch for the CoC. This is much more representative of actual results than the "rule of thumb" CoCs that are still based on film. |
There was no Canon 70D in the list of cameras so I picked 60D. I did use 110mm instead of 103, because I did it from what I remembered reading and didn't double check. But using 103 and the 60D, I get:
I don't know specs on Canon cameras anymore. |
|
|
|
11/17/2015 12:19:34 PM · #29 |
OK, I fat fingered it, LOL!
Originally posted by DOFMaster:
Subject distance 20 ft
Depth of field
Near limit 19.1 ft
Far limit 21 ft
Total 1.83 ft
In front of subject 0.87 ft (48%)
Behind subject 0.95 ft (52%)
Hyperfocal distance 432 ft
Circle of confusion 0.008 mm |
What I did originally was erroneously select 0.08mm (80 microns!) as the CoC. I should have selected 0.008mm (8 microns). The take-=away here is that the numbers should be *smaller* than what DOFMaster will "normally" report. The DOFMaster defaults are based on 30 microns for FF and 19 microns for APS-C, which is an assumption that is based on 35mm film (!).
Message edited by author 2015-11-17 12:20:18.
|
|
|
|
11/17/2015 01:07:35 PM · #30 |
Originally posted by dtremain:
I uploaded the original (JPG)
I started editing with the RAW image, but don't know how to upload that so you can see it.
Don't remember the editing steps, but I think I left it pretty much alone coming from RAW, then did levels, straighten (to tilt it a bit more), high pass sharpen, and Nik Color Efex 3.0 Pro (darken / lighten center). I use PaintShop Pro X8, rather than the Adobe Photoshop. |
I'll let you and the guys nerd out on distances, lenses, and all that, but what I can tell you is that the noise was something you added in post somehow. Your levels adjustment could have started you down the noisy road if you pushed your slider too far. When it comes to balancing histogram levels, I only ever bring the end sliders in to the base of the mountain (so to speak). The middle slider gets you in trouble real quick, so I usually leave it alone. The high pass sharpen probably then sharpened the noise, and the Nik preset may have brought out a little more or at least made it more obvious by brightening it. |
|
|
|
11/17/2015 01:20:34 PM · #31 |
| Do some tests. Select a stationary object at a good distance, and try both auto focus, and then manual focus, both through the viewfinder, and by using the LCD and zooming all the way in, as Waddy suggested. Then compare them after downloading to your computer. Zoom way in and look at them side by side. You'll know right away if the issue is the autofocus. |
|
