NicuB wrote in post #13473003
@tzalman good reference - i was aware of this, but i think it's contributing to the thread.
When I was referring to 'nailing the focus' I had in my mind : the printing size that normal (not pro) people were printing, and also the fact that nobody was pixel-peeping that hard so if you were let's say pretty close to that point you were good. Today most of the people do not print. I print mostly A4 (~12"x8") and with this size I cannot tell a difference for something that is slightly front-focus or back-focus. Another example would be : Robert Capa
@wimg: this is the main thing that I thought about: maybe I was moving. That picture where I observed was hand-held, the tests from yesterday - tripod. Very good information from your post. I'm not aware of what is different between my EOS3 and my digital when it comes to focus. Can you please elaborate on that?
Thank you all for contributing.
Regarding the difference between digital and film AF, film emulsion has a certain thickness which a sensor doesn't have, and this thickness is quite large compared to the CoC, namely 0.2 mm. That gives one two things: first of all DoF will be deeper than anticipated, and secondly the DoF fall-off will be much more gradual.
Now imagine an AF sensor, which does either contrast or phase detection. It has a rather sharp cut-off point for in focus or out of focus, and because an AF sensor essentially is a scaled down version of an imaging sensor, it will also have a zero thickness focusing plane. This implies that the curve for being in focus is steep, where that for film is more gradual.
I found that AF on digital is generally adjusted, certainly with bodies up to and including the 5D II, at a point just before reaching maximum sharpness. In film this is on the shoulder of the DoF curve, within DoF, but also not right at the sharpest point. However, DoF transitions are more fluid with film than with sensors, because, as mentioned, the sharpness/DoF curve with film is gentler.
Furthermore, especially with the lower MP sensors, there is a sharpness plateau, caused by maximum resolution of the sensor and the cut-off point of the AA-filter, something film obviously doesn't have.
The above causes the DoF curve in film to look like a nice, normal distribution or sinusoid type curve, while the DoF curve in digital looks more like a slightly rounded block wave type of curve.
If you look, f.e. , at pictures of brick walls with a lot of detail, taken at 45 degrees to the wall, using larger apertures, in digital you are likely to see 5 distinct sharpness zones as a result :
1. Very unsharp/smeared, OOF in front of the focused zone
2. Getting a little sharper, on the shoulder of the DoF curve in front of the focused zone, some details are almost visible
3. Sharp, the DoF zone
4. Getting less sharp, on the shoulder of the DoF curve behind the focused zone, some details are almost visible, but tapering off rapidly
5. Very unsharp/smeared, OOF behind the focused zone
Looking carefully, you might even be able to see the transitions, or beding points, between these zones, they seem quite distinct too, which is exactly what seems to point more towards a block wave type of DoF zone.
Now, since the AF sensor and AF system make the lens focus towards the top of the slope leading towards the actual DoF zone, this will look slightly less sharp with digital, especially when pixel peeping, whereas with film, due to the gentler slope, and much more gradual transition, this isn't a problem.
Note that the AF system works this way in order to maximize subject DoF.
The difference btween film and sensor DoF is also why I recommend focusing 1/3 into a subject, or important part of a subject, in order to get the entire subject as sharp as possible. Yes, DoF distribution is supposed to be 50% / 50%, but focusing that way will cause more in front of the focusing zone to be unsharp with digital, as opposed to film .
A few years ago I created a little diagram to demonstrate this, with a slightly exaggerated set of curves for film and digital, with the focusing point drawn in. I have attached this drawing below. In yellow the sensor DoF distribution and curve, in green the one for film, and a combined set, to see the difference more easily, and one of each with and without Dof zones delimited, so 6 little drawings in total. The curves do not reach the bottom of the graphs due to noise and/or spurious resolution, and/or minimum film density . Neither do they reach the top, but that is because I am imaging system rather than medium (film or sensor) or lens resolution. Top row depicts film system resolution, middle row depicts digital system resolution, and bottom row is the combined set.
Kind regards, Wim
please log in to view hosted photos in full size.