Back-focus Images

[David Solow]

Here are some images of my tests with my Linhof MT3000 and my Cooke 229mm Portrait lens. Thank you to everyone who tried to help me figure out what was wrong. I thought I would include the LensAlign test plus a shot of a friend.

With the friend. I had him hold up a rigid plastic card with very defined text next to his eye. I had another friend standing beside him, making sure the plastic card was in the same plane as the front of the model's eye.

For the LensAlign tests, I cropped the images and put them through a high pass filter in Photoshop, in order to see the numbers more accurately.

I scanned all of the film (Kodak E100G) at the highest resolution on a Hasselblad X5 scanner.

Here is the LensAlign test from a distance of 54 inches:

[David Solow]

Actually, the first one is from a distance of 54"; the next, 75"; the last, 150". All measurements are from the front of the lens to the plane of focus.

In the first image, 54" from the subject, it looks like the plane of focus is at the 10 or 12 of the smallest numbers.

In the next image, 75" from the subject, it looks like the plane of focus is at the 12 to 16 of the smallest numbers.

In the next image, 150" from the subject, it looks like the plane of focus is at the 20 to 28 of the middle set of numbers.


Here are the images of my friend:

As you can see in the close-up of the eye, while the model's eyes are not in focus, the hair around the side of his head and even his ears are in greater focus.

In the close-up of the face, his chest hair is in greater focus than his eyes.

I thought I would share this with everyone helped me before.

any more comments or suggestions are welcome.

David

[tlitody]

Here are 3 sets of figures. one each 54in, 75in and 150in.
All based on CoC of 0.1mm.

Check the depth of focus and depth of field figures. Methinks that what you are seeing in your tests are such miniscule variances that you will never see them in a finsihed print. Its only because you are scanning and then pixel peeping that you are seeing them. But I may be wrong and someone will correct me.

You can get the software from QIOptiq I think.

[Paul Fitzgerald]

David,

on your targets it looks like standard DOF spread, there is always more DOF behind point of focus than before.

on your portraits it looks like lens design, sharper in the center, spherical aberation.

You really should try a sharp normal lens to test this again at 36", closer reduces the DOF considerably.

Good luck with it.

[tlitody]

the link for the pre-designer software is:

http://www.winlens.de/index.php?id=70

it will give you some idea of the tolerances you are working too.

[David Solow]

Thanks, Tiltody and Paul, for responding.

I neglected to look at my images, after I uploaded them. They are much too small here to see the smallest numbers clearly. Allowing for the fact that there is greater DOF behind the plane of focus than in front, the actual plane of focus is well behind the "0". Even using the most conservative number in the ranges I listed above, there is considerable back-focus for the kind of photography I'm doing. In the finished prints, one is clearly able to see that the chest hair and hair around the side and top of the head are in much greater focus than the eyes.

All of the finished portraits and still lifes I've shot clearly show back-focus.

I like shoot most of my portraits from a distance of between 54" and 75" from the subject at f/4.5. The DOF from those distances with my 229mm is approximately 1" (at 54") to 2" (at 75"). I need to be spot on to get the portrait I want. A shift of 2 inches back-focus is the difference between a shot I can use and one I can't. For some still lifes, I like to shoot even closer with the lens wide open, with a DOF of less than 1 inch. I need the lens/camera combination to be much more accurate than it is.

Thanks again,
David

[tlitody]

Let us know how it works out in the end and whether you had the GG moved or not.

[Frank Petronio]

You mean how much he moves it.

[tlitody]

You mean how much he moves it.

yes that too but he will never be sure unless he measures it himself before sending it to be adjusted (if he does).

[tlitody]

Here are 3 sets of figures. one each 54in, 75in and 150in.
All based on CoC of 0.1mm.

Check the depth of focus and depth of field figures. Methinks that what you are seeing in your tests are such miniscule variances that you will never see them in a finsihed print. Its only because you are scanning and then pixel peeping that you are seeing them. But I may be wrong and someone will correct me.

You can get the software from QIOptiq I think.

also note that your lens design affects the depth of field and depth of focus values. A quick check suggests to me that your lens may be a double gauss type in which case the depth of focus figures in the charts I posted above will be around half of what you see in my attachments. So for 54in at f4.5 it would be around +or- 0.264mm (which is close to rdenny figure of 0.008in (0.2mm) ).

Caveat is I'm not entirely sure about your lens design but you can play with settings in the free software to find what tolerances you want/need to work within.