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!

Arca Rm3di

Jae_Moon

Member
I discussed the challenge of making perfectly accurate focus using tech cameras with MFDB in previous post (#31).

A quick summary, at least in my case, the best possible method with the presently available technologies to achieve perfect focus is by using 'lens equation' with accurate object distance measurement and a precisely indexed focus ring. A lens equation, for a specific lens with a specific focusing system, can be developed mathematically. The 'lens equation' will provide the required angular rotation of a focus ring for a measured object distance.

First, I would like to 'rant' against all View Camera and Tech Camera manufacturers for their failure to promote (i.e. to educate their customers) properly how to use the features they are building into their camera. It is a shame that they let the customers to rely upon HF (hyper focal) or 'focus stacking' for proper focusing, or '35mm rule of thumb' for lens tilt calculation after spending a small fortune on equipment that boast 'multi mega pixels', '13 f-stop dynamic range' and 'lens with resolving power to see the sub atomic particles', etc. Why can they develop and provide more than a 'satisfactory' solutions to the challenges their customers face? Enough with ranting.

Let't talk about Tilt or Swing in Tech Camera application.



A simple geometric diagram for Tilt (both camera and lens) is shown above.

The Plane of Sharp Focus is defined by two points photographer select to be in 'sharp' focus, Near Subject (or Object) and Far Subject. Camera Height, Camera Tilt angle and a specific lens are also defined by photographer to compose the image properly. The lens Tilt Angle and the required Focus Distance are the results of these five input data (the location of Near and Far Subjects, Camera Height reference to the ground, a lens focal length and Camera Tilt Angle). In addition, Near and Far Wedges of Acceptable Sharpness (Angular DOF) are defined by a user selected CoC.

The real challenge of using Tilt or Swing with View Camera or Tech Camera is not finding a lens Tilt Angle but the required Focus Distance after a tilt is made. Harold Merklinger published two books (The Ins and OUTs of Focus, Focusing the View Camera) on the subject of Lens Tilt in details.

Once I was satisfied with the results of 'Lens Equation' in achieving a perfect focus with a measured Object Distance, I decided to tackle the challenge of Lens Tilt (or Swing). It is rather complex geometric calculations but doable using a spreadsheet, such as Excel or Numbers. It took me much longer than I expected (my mathematical brain atrophied significantly over last 40 years) but I completed the project and have been using since last Spring.


The picture above shows the setup I use for both leveled and lens tilted application. It maintains the exact geometric configuration between Rm3di and Disto 5 so I can always accurately measure the Object Distance or Near (Far) Object locations. Since all Disto measurement are offset from Lens Axis, Lens Nodal Point and DB sensor plane, proper corrections are made in the program.

Near (and Far) Object locations should be measured by their horizontal and vertical distance from the lens nodal point. Disto 5 has a function to measure them (Direct Horizontal Distance).

I made a quick shot this morning after reading "Torger's" comment. I used SK90, placed two $20 bills on the floor (1.96m and 3.22m away from the camera, and the nose of Andrew Jackson was the targets), camera was tilted 23.8 degree and the camera height was 0.96m. The required lens tilt was 4.82 degree and the calculated focal distance was 2.14m, therefore the index setting of Red 22.1. The aperture setting was 8.0.


Images are directly from C1 without any enhancement.

100% of Front US$20 bill.

100% of Back US$20 bill.

One additional picture I took while working on the program, Lens Swing. I superimposed 100% images of two blocks, front and back.






This is the one man's story to use a tech camera, Rm3di, for its full potential.

There are several mechanical designs of Rm3di which I wish they improve soon.

1. There are no indentations (mechanical click) for both vertical and horizontal shift center positions.
2. Thumb dial for Lens Tilt needs major improvement, for both accuracy of setting and ease of use.
3. Helical ring is too tight to rotate. I understand its need to stay at a set point but the friction could be much less.
4. Focus setting point should be a 'fine line' much closer to the index markers.


Jae Moon
 
Last edited:

torger

Active member
Impressive work. As the RM3Di allows for precise setting of a lens distance and a tilt angle, laser measurements with a special app that has the mathematics can indeed solve the problem which you prove. Nice. If I had an RM3Di I would certainly like to have such an app.
 

Jae_Moon

Member
what sort of tilt angle tolerance can one expect to set/hold and measure?
I am assuming that your question was for Rm3di. It has +/- 5 degree with markings for each degree. I can only reliably set at 0.5 degree accuracy, at the best. It's 'thumb wheel' design is a sub-standard one. I think Cambo's (also ALPA's) thumb screw design is easier to use. Since one can calculate the tilt angle very accurately and the magnitude of tilt angle increases with longer focal length lenses, all manufacturers should consider using a 'caliper' based tilt adjustment which allows 0.1 degree accuracy. So far, they haven't been hearing demands for improvement.

Jae M
 

f8orbust

Active member
Impressive work...
Indeed. However, I get the feeling that this is heading in the direction of a solution looking for a problem. I don't remember seeing any pictures of Ansel Adams, Paul Strand or Edward Weston with an abacus. With experience, this is all easily achieved 'in practice' using those two things either side of your nose and the 'app' running in the space between your ears :)
 

Jae_Moon

Member
Indeed. However, I get the feeling that this is heading in the direction of a solution looking for a problem. I don't remember seeing any pictures of Ansel Adams, Paul Strand or Edward Weston with an abacus. With experience, this is all easily achieved 'in practice' using those two things either side of your nose and the 'app' running in the space between your ears :)
Yes indeed, if I was using an 8x10 or even 4x5 with 'BRIGHT' GG which allows me to see the image, I wouldn't have bothered. With MFDB, we are flying blind. If one cannot see clearly while flying an airplane, avionics come very handy. No shame in relying on 'grey matter' between ears.

Jae M
 
Last edited:

torger

Active member
I don't think Arca-Swiss has taken tilt that seriously, and most RM3Di users don't either. In architecture, which I would say is the main application for RM3Di, tilt is not used that often at all. In landscape I see that tilt has lost popularity since the 4x5" view camera days, and actually also shift. Focus stacking is a new technique which seems to be increasing in popularity.

One reason for this I think is that shooting styles indeed have changed, more wide angle less of longer focal lengths, but also that cameras have changed, and that there's an almost unhealthy obsession with resolution. I think the DSLR competition has made many even more obsessed with resolution as it's the easiest spotted and most obvious advantage.

There are indeed new type of compositions you can do with focus stacking, but often I see stacking used in scenarios when you could focus conventionally, but someone has this fancy Rodenstock lens and IQ180 and wants to shoot at f/5.6 to minimize diffraction. My take is that if you drop below f/11 the depth of field becomes for 90% of the scenes unpractically short, so I will not get gear that gives me an urge to shoot at larger apertures than that :). There's nothing wrong with f/11 even on IQ180, but price must come down before I can shoot f/11 and not get the feeling I'm wasting money ;).

My normal working range is f/11 for 80% of the images, f/16 for 15% and f/22-32 for 5% of the images. If I had to drop down a stop or two it would be much harder to make images the way I want them to look (i e "in focus" all over).

Then we have the 645-fullframe with 90mm or even 70mm image circles, shifting margins are a bit smaller than before(?) or at least it's easier to detect sharpness falloff, and that has caused many to reduce the use of shifting.

The pancake cameras are more limited in movements, many don't even have tilt or have very expensive tilt mounts to select lenses, which further have contributed to making a shooting style with less movements popular.

RM3Di ultra-precise focusing precision is a natural part of this development, but I think it has now gone into overkill mode. Creative possibilities are crippled in the search for even more resolution and precision which is now at a level that don't contribute to anything valuable. It's more than a help to "focus in the blind", the precision is more than any 8x10" user could dream about. I think this "need of precision" is not coming from an actual need in creative imagemaking, but rather from an unhealthy race in resolution.

I have a Canon system too. My TS-E 24mm II has actually more flexible movements than an RM3Di - I can tilt in any direction and the compared to an IQ180/Rodie40 I have larger useful shift range too. To me a MF technical camera should not only be a carrier of higher image quality, it should also carry better creative possibilities in terms of movements. Resolution is becoming a less important differentiator for each year.

Then we have the usability of ground glass on digital. I don't agree that we are flying blind. It is a challenge though, and require some practice and skill, I admit that. As far as I can see it was not really easier on a 4x5" though. The corners on a 90mm on a 4x5" frame was as dark as a 35mm on digital, the incoming light angles are the same so how can it be different? Actually with Rodenstock you have f/4 and retrofocus (brighter corners) so that must surely be easier than it was on 4x5", and I'm sure Linhof's new bright ground glass is brighter than most used on their large format cameras. Sure, the format is smaller so you need more magnification on the loupe. That's the difference. Due to some grain-myth many stay with the same magnifications they had on 4x5" and then of course it becomes more difficult.
 

Frederic

Member
...
RM3Di ultra-precise focusing precision is a natural part of this development, but I think it has now gone into overkill mode. Creative possibilities are crippled in the search for even more resolution and precision which is now at a level that don't contribute to anything valuable. It's more than a help to "focus in the blind", the precision is more than any 8x10" user could dream about. I think this "need of precision" is not coming from an actual need in creative imagemaking, but rather from an unhealthy race in resolution.
...
Oh well... don't you think pancake camera users, as buyers of quite expensive gear, have done their homework and selected the equipment that suits their needs best too ?
It's like saying your Techno hampers creativity because it has to be shot on a tripod. Isn't that crippling your creative possibilities ? It isn't if it's not the way you shoot, right ?

My back lacking LV, focus masks or a usable 100% view, I certainly consider the focusing precision to be very valuable and not overkill at all.
 

timparkin

Member
One of the facts that drops out of the maths for tilt shift lenses is that the focus along the line of the lens axis remains just the same as whether the lens is tilted or not.

i.e. if you looks straight through the middle of the lens (rear tilt helps) you can use hyperfocal focusing, tilt will only change where the near and far DoF bounds converge to.

Things get more complicated with front tilt but they can be worked out.

If you want to see tilt in practice, have a look here.

http://static.timparkin.co.uk/static/focus/index2.html

If anybody would like a medium format version just let me know what CoC and lens range to use and I'll tweak the figures..

The maths assumptions to make things simple start to break down near 7 degrees of tilt. You get the idea hopefully though.

Tim
 

Geoff

Well-known member
This is wonderful! How hard is it to do a MF version? Something say with COC of .007 (?) and a 55 and 90 mm lens?

Is this assuming lens tilt is on axis with the center of the lens?

Thanks much -
 

JGR

New member
This discussion is very interesting, I looked to see if I could change the title to reflect the points being discussed but couldn't find a way.

Jae & Torger's diagrams and points have left me wondering why Arca haven't implemented these ideas and maths into the E-Module which would make it a very attractive purchase. Why give you a tech cam with hyperfocal distance confirmation/settings via it's E-Module without taking tilt into account during the design of this product. I imagine this may well be a future upgrade!?

Wait, hold that thought... Usable live view is around the corner!

I need to get my hands on the RM3Di to asses the pro's and con's of focusing using tilt on the ground glass. I also think Jae's use of the Disto 5 along with an app sounds interesting. It may sound like it's all a bit much but in practise I doubt it will get in the way of creativity. Knowing your focus is nailed would surely give you more creative freedom.

Each to there own I guess. Some may prefer the GG, some live view, others may prefer an app such as Jae's. Two of those are ready available, live view isn't too far way. I'm in the market now though and can't wait around for improved live view, I will consider that option when it arrives.

I haven't quite got my head around Jae's math and how I would use the Disto 5 should I decide on purchasing one vs the E-Module though. I think further explanation for my ageing grey matter is required (hint hint) :)

In the mean time I will be trying to wrap my head around f8orbust's maths for working out tilt which I'm sure will help me understand Jae's maths as they are essentially the same thing... I know I have read this by Harold Merklinger but it was some time ago and think I need to read it again, at least for my own sanity!
 

David Klepacki

New member
We have been using our own in-house focusing system for the RM3Di for some time now. It is amazingly fast as well as accurate. And, there is no eye fatigue that builds up over a long day's work as when we were previously trying to judge focus using loupes and ground glass.

I have attached some pics of the system. It consists of hardware that quickly attaches a Disto on/off the RM3Di using its own accerssory interface (typically used for the eModule). It has a rotating collar that allows the Disto to fluidly rotate up and down as well as lock into place (great for measuring tall buildings and tree tops, or for when shooting downwards from higher vantage points). Of course, our adapter fixes the measurement distance to the same focusing plane at the sensor for consistency. You can use this to finely calibrate your lenses for very high precision and consistent focusing accuracy with the RM3Di.

The software part is a smartphone app (iphone/ipod or Android device) that calculates optimal tilt angle and focus setting for a desired distance or DOF. A sample screen shot is attached below (btw, minus one is used to represent infinity distance). We also have adapters that convert the Arca-Swiss accessory mount to either Alpa or Cambo accessory mounts. This allows us much greater freedom in choosing accessories for the RM3Di. Shown below is an example of using our Alpa iphone adapter on the RM3di, and another image of what our Cambo adapter looks like.

If enough people are seriously interested, we can make it available commercially.
 
Last edited:

Jae_Moon

Member
The software part is a smartphone app (iphone/ipod or Android device) that calculates optimal tilt angle and focus setting for a desired distance or DOF. A sample screen shot is attached below (btw, minus one is used to represent infinity distance). We also have adapters that convert the Arca-Swiss accessory mount to either Alpa or Cambo accessory mounts. This allows us much greater freedom in choosing accessories for the RM3Di. Shown below is an example of using our Alpa iphone adapter on the RM3di, and another image of what our Cambo adapter looks like.
David:

It looks fantastic!! I am specially impressed how your Disto mounting bracket works, it is perfect that the Disto sensor can be located along the lens axis and at the DB sensor plane. I want to order one if the bracket fit on Rm3di.

I have several technical questions. (The grey matter between ears want to know :))

1. Does the program assume the Plane of Sharp Focus be always on the reference plane (=ground), or you have another input screen which takes (x,y) positions of both Near and Far Objects?
2. Does the program allow the tilting down of camera in calculating the lens tilt angle and focus distance?
3. The focus distance that the program calculates, does it include the effect of lens tilt, or just based on lens focal length, Aperture, Coc, Far and Near Distance?

If camera manufacturers neglect to meet the needs of their customers, maybe we, the customers should take charge and develop better solutions.

Jae M
 

David Klepacki

New member
David, that looks fantastic! Will you be making it commercially available?
If I can get a minimum number of people (ideally 50+), then I will make a production run of the parts and sell to others who can appreciate it. Otherwise, it is hard to justify the costs, and mostly my time. I put a huge amount of effort into this over the last year or so. The problem is that I demand perfection, even when I know it is not realistic. I have to work closely with the machinists here. I have rejected parts several times, when it was not exactly right, sometimes being less than a mm off. As another example, the coatings I use are a teflon compound since the standard anodization was not smooth enough for me, etc, etc. It took me a while to find people who are skilled at what they do and have the precision equipment to do it. And, it is US labor, so the costs are not competitive with what can be made in China, but frankly I don't care. But, I can honestly say that I am very happy right now, and it was worth all the effort.

I think the best way to proceed is to put this project up on kickstarter.com and see what happens. Unfortunately, I am in the middle of another project right now, but I will try to get to it with in the next two weeks.
 

David Klepacki

New member
David:

It looks fantastic!! I am specially impressed how your Disto mounting bracket works, it is perfect that the Disto sensor can be located along the lens axis and at the DB sensor plane. I want to order one if the bracket fit on Rm3di.

I have several technical questions. (The grey matter between ears want to know :))

1. Does the program assume the Plane of Sharp Focus be always on the reference plane (=ground), or you have another input screen which takes (x,y) positions of both Near and Far Objects?
2. Does the program allow the tilting down of camera in calculating the lens tilt angle and focus distance?
3. The focus distance that the program calculates, does it include the effect of lens tilt, or just based on lens focal length, Aperture, Coc, Far and Near Distance?

If camera manufacturers neglect to meet the needs of their customers, maybe we, the customers should take charge and develop better solutions.

Jae M
The bracket will fit any Rm3di, so am not sure of your concern. In fact, it will fit any of the R-cameras that have the slot to hold the eModule, since that is what I use.

To answer your other questions:

1. Right now, the program assumes a common reference plane, since that is how we shoot. And, it removes a lot of complexity for the photographer, since otherwise he/she would have to enter (x,y) locations for the planes of near and far focus. Our method of using a common reference plane like the ground, makes it a lot easier and does not sacrifice any accuracy at all. We have been shooting this way for many months now. But, we use primarily tilt as opposed to swing. For swing, it is a little harder to visualize a common reference plane, and I will probably end up adding the option to specify (x,y) locations.

2. Right now I assume a level camera, because that is how we shoot. However, all calculations are based on a local coordinate system that is aligned with the optical axis. It would be very simple to add the geometry to allow for a tilted camera.

3. Oh, of course! That is the whole point of constructing this tilt calculator. You correctly point out that the proper selection of focusing distance must be calculated using the tilt angle. You cannot use distance focusing tables generated for zero lens tilt. If you do, you will get images that are out of focus or soft at best.

As for you last comment, I am just grateful that such a camera exists like the Rm3di. For a precision fanatic like myself, it is a dream camera. Yes, there was a lot of work that needed to be added in order to extract its highest image quality, but at least we have the tool to do it.
 

f8orbust

Active member
David - really interesting to see what you've done. Given that distos are generally bomb proof, it looks like a really robust system suited for landscape photography; one that could easily be adapted to a number of camera platforms (e.g. Alpa, which has a really precise/fine-grained method for setting tilt angles).

The only question I have is to do with how you set the tilt angle accurately on the RM3D/i once you know what it is? Other than '0' which has a detent, I couldn't really set a tilt with a greater accuracy than 0.5 degrees(ish). In fact, I could never be 100% certain what, exactly, the tilt angle I set was. With a wide angle lens, even a small change (0.1 degrees) in the tilt angle can have a fairly dramatic effect on where the plane of focus falls.

Jim
 

jlm

Workshop Member
you might try this:

there is an I-phone app for level, so if you hold the phone against the lens, it will measure degrees off plumb. haven't tried it, just thought of it.

just tried a few, not all will work
spirit level does and reads to 0.1 degree. you hold the phone vertical with the side against the lens. if the camera is tilted, you could measure the back and the lens and figure it out
 

Jae_Moon

Member
To answer your other questions:

Q. Does the program assume the Plane of Sharp Focus be always on the reference plane (=ground)?

A. Right now, the program assumes a common reference plane, since that is how we shoot. And, it removes a lot of complexity for the photographer, since otherwise he/she would have to enter (x,y) locations for the planes of near and far focus. Our method of using a common reference plane like the ground, makes it a lot easier and does not sacrifice any accuracy at all. We have been shooting this way for many months now.
Thank you for your quick reply. I may be totally wrong since I am trying to understand what your program is doing based on your post and lots of my assumptions :). But I would like to ask a question.

Since you are calculating lens tilt angle and focus distance based on the reference plane (horizontal ground or table top, for example), why do you need the user input of Near and Far Object distances? The lens tilt angle with a reference plane can be calculated with two variables; the hinge point distance (camera height in your case, either from the ground or from the table top) and the lens focal length. And also, the focus distance can be calculated with two variables; the lens tilt angle and the hinge point distance.

For those who still can remember high school geometry,

alpha = arcsin (f/J), where alpha is tilt angle, f is lens focal length in mm, and J is Hinge point distance in mm (or camera height in this case)

Fd = J/sin (alpha), where Fd is calculated focus distance.

What am I missing?

Jae M
 

David Klepacki

New member
David - really interesting to see what you've done. Given that distos are generally bomb proof, it looks like a really robust system suited for landscape photography; one that could easily be adapted to a number of camera platforms (e.g. Alpa, which has a really precise/fine-grained method for setting tilt angles).

The only question I have is to do with how you set the tilt angle accurately on the RM3D/i once you know what it is? Other than '0' which has a detent, I couldn't really set a tilt with a greater accuracy than 0.5 degrees(ish). In fact, I could never be 100% certain what, exactly, the tilt angle I set was. With a wide angle lens, even a small change (0.1 degrees) in the tilt angle can have a fairly dramatic effect on where the plane of focus falls.

Jim
Here is what we use: Digi-Pas | Pocket-Size Digital Level | DWL-80Pro

You just need to be creative with some velcro and maybe a small support to temporarily attach it to either the lens or the R mount itself. Yes, it is a little tricky (not to mention ugly) with the shorter "pancake" length lenses, but not really a problem with lenses like the Rodenstock 32W. We are actually thinking of making a support ring to attach it rigidly to the R mount (the strongest piece that offers the most stability). The 80Pro has a precision of .05 degrees, but if you try to hand hold it while measuring, you will get nowhere near this kind of precision.

As Jae has mentioned, this is the weakest aspect of the Rm3di design. I too would like to see a newer model with vernier fine tuning of the tilt angle. But until that day comes, this is what you need to use. The good news is that this device is not only cheap but is so small and lightweight that you really don't mind. And, it has an integrated carabiner, which makes it easy to carry and attach to things (like your belt).
 
Top