The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

focus distance accuracy

jlm

Workshop Member
i had a bit of a diversion, so:
regarding the alpa lens shimming, most of the concept is a red herring, since all you are doing is ensuring that when the lens helicoid is set to the infinity stop, the camera is actually focused at infinity. very important, to be sure, but after that, focus at any intermediate distance is entirely up to how accurately you can estimate that distance (maybe to a few inches, if you are lucky) and then, how accurately you can set the distance scale (and ultimately the image distance) on the helicoid to that number.

Here the arca leads the way with the accurate and fine pitch helicoid, set by working from a chart, I believe. Alpa offers add-on scales with intermediate markings as useful aid.

for example, using a 28mm lens with a subject distance of 10,000mm (10meters or about 33ft)and comparing that to a subject distance of 10.2 meters ( 33ft-8inches)

to give you and idea of the dimensions, the lens to image plane distances are:

inf: 28.000mm
10M (33ft): 28.079mm
10.2M (33ft-8inches): 28.077mm

so at about 33ft, 8" subject distance change means about .002mm distance to sensor change

the Arca has five turns for it's helicoid and my estimate is it moves about 7mm (6mm is what my 120mm lens has to move to get from inf to 2.6M (8ft). that works out to about .004mm movement per degree of rotation and for a 3" dia lens adjustment ring, one degree is about 1/2mm on the barell, or less than 1/32"

it would take 1/2 of one degree for the Arca to focus between 10M and 10.2M with a 28mm lens

for a 43mm lens, about .004mm difference
for a 70mm lens, about .010mm difference between 10 M and 10.2M
for a 120mm lens, about .029mm difference

using 1/f=1/I + 1/O


any discussion?
 
Last edited:
G

Garcia

Guest
With a 28mm @ f/5.6 focused at 10 meters, DOF is so large that you can afford a good bit of focus error.
 

Anders_HK

Member
With a 28mm @ f/5.6 focused at 10 meters, DOF is so large that you can afford a good bit of focus error.
Yes, but afford minor error since in particular for wides the critical is the tolerance on depth of focus (not depth of field).

Alpa and Arca addresses this by two different means; shimming and fudge factors.

The rest is about tolerances of individual components of the systems and how they add up (and adjusted/shimmed in factory). Under assumption that Alpa's tolerance of their system is suffice small, then there should be no issue to accurate focus with an Alpa helical compared to Arca. A weakness of the Arca in regards to tolerances is the tilt that is built into the body since for sharpest optimum focus without tilts, and in particular for wides, the lens need to be perfect parallel to sensor (within 0.02mm). This is possibly a reason to why Alpa is providing tilt adapter for 80mm lenses and longer. An Alpa also avoids table look ups for each focus, as is Arca's method...

For the Schneider 24mm Digitar as example the total focus travel from 1.5m to infinity is only 0.3mm and if instead an error causes defocus by 0.1mm one will focus on 5m instead of infinity. Thus the large depth of field would not be suffice helpful to achieve optimum sharpnes of image... Now compare that with if the tilt neutral position is out of similar tolerance... or not perfect adjusted for where the plane of focus is intended at tilt.

Regards
Anders
 
G

Garcia

Guest
Good points, Anders.

Now I realize that Arca helicoid moves equal amount no matter which focal lens you are shooting with, but Alpa or any other traditional helical mount displaces the lens as a function of its focal length so you get more accurate focus setup thanks to fine pitch thread for wider lenses.

So, in wide angles you get more precise lens placement per degree of focus ring turn with a classic helical mount than with Rm3d five turn mount albeit Arca system is more precise for focusing teles.
 

jlm

Workshop Member
measured travel of my lenses:
43mm: 2.5mm travel, about 180 degree rotation, close focus limit 0.4M, pitch 5mm/rotation

70mm: 8.3mm travel, about 270 degrees rotation, close focus 0.8M, pitch about 11mm/revolution

120mm: 6.1mm travel, about 270 degrees rotation, close focus 2.6M, pitch about 8mm per revolution

I don't have an arca to compare, but if it travels enough to cover the longest travel range of these lenses, say 8.3mm in five revolutions, the pitch would be 1.6mm per revolution, about 3 times finer pitch than my widest, the 43mm lens
 

cng

New member
Alpa and Arca addresses this by two different means; shimming and fudge factors.
I'm not an Arca user, but can we all please stop referring to the Arca focusing system as using a "fudge" factor. It's an incorrect description and does a disservice to the R camera system. It implies that it's applied randomly and/or by guesstimate differently in every situation, whereas the Arca focusing offsets are actually very precisely measured for each individual lens and applied consistently.

I think others have been referring to it as using a "lens offset" factor.

(I'm not aiming this specifically at Anders, it's just something that irks me.) :poke:
 

jlm

Workshop Member
I'm guessing here, but i assume the arca helicoid is calibrated in arbitrary units, probably quite finely, and then for each lens, they produce a chart that relates the helicoid scale to the actual distance for that lens?

so i look up say, 10M distance, for a 120mm lens on the chart, then set the helicoid scale to that number?
 

Shashin

Well-known member
Hi. I have been visiting Getdpi for a while, but have never joined. I work in an imaging center dealing with scientific, technical, and applied imaging.

Just to throw my 2-cents i here about image plane tolerances. All you need is to calculate the depth of focus. You only need two things to work it out: permissible circle of confusion and the f-number--focal length is not a factor.

Depth of focus = 2 x circle of confusion x f-number

The constant of 2 is that the depth of focus is in front and behind the plane.

If you take the pixel pitch for the circle of confusion, that will give you a number to work with. With a sensor with a 5 micron pitch (0.005mm) with a lens at f/5.6 would result in a depth of focus of 0.056mm. That would be the tolerance at the focal plane.
 

jlm

Workshop Member
interesting, working the equation the other way:

using .056mm as the image distance spread for an object distance or 10M (32ft), for a 28mm lens at f5.6, the object distance range is 10M to 5.8M

using a 70mm lens at f5.6, same .056mm, a target 10m away would have a 1M depth of field


these seem overly optimistic

feel free to check my math, i used 1/f=1/I +1/O,

or O=1/((1/f)-(1/I))
 

Shashin

Well-known member
Depth of focus is the total tolerance of the area--0.056mm. To put it another way, +/–0.028mm from the focal plane is the tolerance for the sensor. So depth of field should be on both sides of the 10m mark.

Sorry, I am out the door and so I may not be spending the time to do justice to your comment.
 

jlm

Workshop Member
Comment:

jack stated in his IQ review that the use of focus mask might neuter the need for shimming your back. I think what he means is that you don't need to be limited by the ability to first, accurately measure the distance to the subject and then second, set the lens to that number.

you just guess the distance, keep taking shots and chimp using focus mask until the LCD satisfies you. Sounds good to me.

you will still have to shim to make sure the lens can mechanically get to inf however, but this will only pertain to shots taken at...infinity
 

Anders_HK

Member
I'm not an Arca user, but can we all please stop referring to the Arca focusing system as using a "fudge" factor. It's an incorrect description and does a disservice to the R camera system.

I think others have been referring to it as using a "lens offset" factor.
Fudge factor was in no ill meaning :D

"Lens offset" is not correct term since the factors applies to "fudging":eek: /adding factors to focusing in order to compensate fabrication tolerances for camera, lens and back.

Perhaps we can hereby instead name them "adjustment factors"?? :rolleyes:
 

GrahamWelland

Subscriber & Workshop Member
Comment:

jack stated in his IQ review that the use of focus mask might neuter the need for shimming your back. I think what he means is that you don't need to be limited by the ability to first, accurately measure the distance to the subject and then second, set the lens to that number.

you just guess the distance, keep taking shots and chimp using focus mask until the LCD satisfies you. Sounds good to me.

you will still have to shim to make sure the lens can mechanically get to inf however, but this will only pertain to shots taken at...infinity
Exactly.

So long as you can actually focus to infinity it becomes an iterative process where you can basically treat the lens barrel distances as merely a guide and then fine tune using focus masking.
 
Top