Focus->recompose NOT a perfect approach

[paulbod3]

Wonderful, but very few lenses are to be considered flat field, some macros and the ones for copying ; in most cases the focus is to be mantained on a nearly spherical surface that mimics the front element curvature.

 

[paulbod3]

One typical example of nearly spherical focus plane as follows :

http://www.photozone.de/8Reviews/lenses/tamron_1750_28/index.htm

also, what is called the barrell distortion can be considered as another proof of non flat focusing plane.

 

[ljclark]

in most cases the focus is
to be maintained on a nearly spherical surface that mimics the front
element curvature.

Is that an accurate rule of thumb? I have a 35mm lens (for 35mm FF) that has a CONCAVE front element. What is that supposed to tell me?

Regardless, any of these "calculations" that don't take the individual lens into consideration seem a little dodgy to me.

--
If you don't talk to your cat about catnip, who will?

 

[John Bean (UK)]

in most cases the focus is
to be maintained on a nearly spherical surface that mimics the front
element curvature.

Is that an accurate rule of thumb? I have a 35mm lens (for 35mm
FF) that has a CONCAVE front element. What is that supposed to
tell me?

So have I, along with two 35mm lenses with small, "bulbous" convex front elements and one with a quite large, almost flat front element. None of them appear to have radically different flatness of field.

Regardless, any of these "calculations" that don't take the
individual lens into consideration seem a little dodgy to me.

Quite.

--
John Bean [GMT - is there any other?]

PAW 2007 Week 10:
http://waterfoot.smugmug.com/gallery/2321711/1/134422693/Large



Index page: http://waterfoot.smugmug.com
Latest walkabout (4 April 2006): http://waterfoot.smugmug.com/gallery/1348582

 

[av_photo]

in figuring out how much distance to the subject changes when you focus and then recompose. Obviously, you can't do a calculation on the fly, but doing a few examples will give you a feel for it.

 

[paulbod3]

Agree it is not such an accurate rule of thumb, the back element shape should be taken in consideration too as well as all the intermediate surfaces, and their positions varies depending on zoom position and focusing distance.
The shape of the front element anyhow is quite important.

The OP is absolutely correct in underlining the importance of accurate focusing with shallow dof, and, from a tripod, with time available, live view can give a very accurate approach.
Handhold, a larger choice of focusing points is surely an improvment.

 

[paulbod3]

Personally, I feel Oly has concentrated on live view, where the next development might be full time af allowing any point of the sensor to be chosen.

At this stage, live view af (or af+mf) is too slow (1 sec delay) for sports etc, and suitable mostly for professional work on tripod for macro, portrait, panorama.

With a moving subject, any af, withone, 3 or 49 points will have a slight misalignment due to the time delay beetween measurement and actual taking.

Maybe Oly thinks that a 49 point af to be used only mirror down will not be necessary when full time af live view is available (less than one year ?). Anyhow other brands use an hybrid (contrast + distance) af that is not as reliable as contrast only, so overall deuce...

 

[paulbod3]

http://www.olympusamerica.com/cpg_section/lens/dea/products/lens/50_20M/index.asp

The ZD 50/2 looks to me convex too, it is supposed to be flat field and suitable for copying works.

Actually a surface the size of a lens all at the same distance from a plane should be concave....

 

[Raist3d]

If you know very well how your depth of field works, you can definitively get things in focus. If the objects are in an arc more so. I don't think it's as hard as you mention in actual use, I have been able to get by recomposing just fine.

• Raist

Not very long ago I finally realized why I need more than 3 focus
points, or at least 3 points more separated than today on Olympus
dSLRs.

The focal plane is literally a PLANE, not a sphere! When you focus
at e.g. 3 meters you have NOT focused on everything that is 3
meters away from the camera. You have put a plane in focus, a plane
that has its closest point to the camera at the intersection with
the cameras optical axis. All other points in the focal plane are
further away from the camera.

When you recompose after focusing, the focal plane will turn and
end up behind the subject that focus was locked on.

An example: If you focus on the eyes of a model 3 meters away and
recompose by tilting the camera down by 0.5 meter (on the subject),
focus will end upp 4.1 cm behind the eyes. This is a lot if you
work with short depths of focus.

The solution is of course to use an auto focus point that lets you
focus and compose at the same time. That is why I really hope
Olympus will deliver more than 3 auto focus points in the E-1
successor.

Andreas

--
Raist3d
Tools/Gui Programmer - vid games industry, photography student