stitching camera

[hovis]

Hello

I'm an artist, I'm in the process of building a stitching camera to make high resolution images of my work for printing. I've already spoken to an engineer about building a one-off but I'm curious as to whether there would be enough demand for me to invest in the tools to make them myself.

I'm aware that similar things already exist. What I'm proposing is a low cost kit which would require some work from the buyer to make fully functional. You would need to make a bag bellows, and to find and attach a lens mount for whichever camera you use.

Not being an engineer, the camera will not have the finesse of say, a Sinar camera, but I'll be using it on my own work so it will be accurate and robust. It will be made from aluminium, left raw to keep the cost down, and the XY sliders will have thumb screws to hold their position. I don't have a definite figure yet but I'm estimating I will be able to make them for around £300. I'd be very grateful for any feedback.

Please view the camera at www.jonathanpauldavies.com/camera.html
The blue shape represents an slr camera and would not be part of the kit.

Kind regards

Jonathan

[David Grover / Capture One]

Issues to think about...

- Maintaining electrical contacts between body and lens
- Image circle(s) of proposed lenses.  Enough coverage for your requirements?
- Focus issues once lens is moved away from its normal mount (infinity focus lost)
- ...?

Might be easier to make a jig so the whole camera can move to cover / 'scan' a whole artwork.  Then you need to think more about accurate stitching, even lighting...

Hope that helps.

David

[hovis]

Hi David

Thanks for your reply, I think I need to be more specific. I designed the camera shown around an El-Nikkor 135mm and Sigma SD14 combination. the two diagonal lines represent the viewing angle of the lens and the recessed plane on the blue box is the sensor plane, I think it should, however, work with a range of combinations of large format lens and slr bodies. I'm trying to find out if there would be enough people out there interested in a low cost/low tech but well made stitching camera to justify me investing in the tools to make them.

The front plate with the hole is where you would attach a large format lens, this in turn is attached to a focussing rack, if someone needed to maintain an electrical connection between lens and body then this idea wouldn't work for them, I'm simply offering something which I was unable to find on my budget as an artist, which is a means to get very high resolution images of stationary objects.

Regards

Jonathan

[yaya]

Check this one out

Yair

[DavidP]

Why don't you check out Kapture Group, they have some nice stitching backs that fit on view cameras. They have one that does 4 in a rectangle.

[hovis]

Hello again

YaYa and DavidP, thanks for your reply, unfortunately I think something is being lost in the description of my project.  I already have a design (shown in the image) which I hope will work well. I'm just trying to find out whether there would be any interest for such a device from others, to make it cost effective to invest in the tools and make them myself.

The Bulldog camera is simply a large format camera (it has no inbuilt stitching capability), it would require the additional expense of a high res' capture device. The Kapture group device while very nice is $5000, again this is outside my budget and again I would also have to buy a large format camera to attach it to.

My proposal is a low cost stitching device for an SLR camera and large format lens.

Hope this helps

Warm regards

Jonathan

[marcwilson]

Hi Jonathon,

Does your camera work by moving the lens or the camera...as presumably you would want to avoid any parallax issues with any 3d elements to your work, or is it all flat?
Also how much shift can you get out of it...or is that only limited by the lens you choose to use?
And presumably it will only work with a certain minimum focal length of lens to allow for focussing...or are you not concerned with infinity focus at all due to the nature of your needs?


Marc

[BJNY]

Hello Jonathan,

Your design made me think of some components from Really Right Stuff
http://reallyrightstuff.com/rrs/items.asp?...amp;Tp=&Bc=

in particular, a variation of this
http://reallyrightstuff.com/rrs/Customkiti...il%2DPk&eq=


Also not too different from your design is Novoflex's ProShift+
towards the bottom at http://www.novoflex.com/english/html/fr_ema5.htm

Billy

[BJNY]

Jonathan,

Just for reference, here's a different design not requiring an optical rail:
http://www.gottschalt.de/de/kameras_04.html

Billy

[marcwilson]

I think what Jonathon's design has in its favour, as well as the obvious cost of course, is that by using large format lenses the amount of shift available is huge so multi shot stitches and very large files will be attainable, which is of course essential if using the small dslr sensor. This is it's bonus over say the gottschalt or zork psa designs.

I see it as a simplified and of course much cheaper cambo xpro.

Depending on focal length of usable lenses it could provide a very high resolution file dslr stitching option for a landscape or architectural shooter.


Marc

Jonathan,

Here's a different design not requiring an optical rail:
http://www.gottschalt.de/de/kameras_04.html

Billy

[Ron Steinberg]

I think what Jonathon's design has in its favour, as well as the obvious cost of course, is that by using large format lenses the amount of shift available is huge so multi shot stitches and very large files will be attainable, which is of course essential if using the small dslr sensor. This is it's bonus over say the gottschalt or zork psa designs.

I see it as a simplified and of course much cheaper cambo xpro.

Depending on focal length of usable lenses it could provide a very high resolution file dslr stitching option for a landscape or architectural shooter.


Marc

Using a dSLR with a lot of shift will create a problem with shadowing from the mirror box. Otherwise the design looks sound. For the cost of a custom made bellows I'd also look at a helical mount for your large format lens and build something more along the lines of a Cambo WideDS but have all of the shift on the back.

[Bob Smith]

While the design looks like it could work well... I think a better and potentially simpler approach is to keep a good camera/lens stationary and perfectly square to the flat artwork while incrementally moving the artwork.  I've done this with very good success using a 5D with 180mm Macro.  In order for this to work, the art has to be mounted to something rigid and perfectly flat.  Or a canvas on a good set of stretcher bars that result in a very flat finished piece.  I've rigged a system that can allow a mounted piece of art to slide across a wall.  The camera is mounted to a sturdy studio stand.  I've thought about and experimented with designs similar to what's being proposed here.  Moving the artwork itself has proven a better method for my needs.  This system means you can handle any size of artwork your system can move.  And you can capture as high a resolution as you want depending on how close you place the camera and how many frames you take to encompass the entire piece.  You're always using the sweet spot of a good lens and not having to rely on the edges of the images circle.

[Thomas Krüger]

Here is another project with an excellent controller: http://www.emrichs.de/gigapanbot/index-en.htm

[elf]

While the design looks like it could work well... I think a better and potentially simpler approach is to keep a good camera/lens stationary and perfectly square to the flat artwork while incrementally moving the artwork.  I've done this with very good success using a 5D with 180mm Macro.  In order for this to work, the art has to be mounted to something rigid and perfectly flat.  Or a canvas on a good set of stretcher bars that result in a very flat finished piece.  I've rigged a system that can allow a mounted piece of art to slide across a wall.  The camera is mounted to a sturdy studio stand.  I've thought about and experimented with designs similar to what's being proposed here.  Moving the artwork itself has proven a better method for my needs.  This system means you can handle any size of artwork your system can move.  And you can capture as high a resolution as you want depending on how close you place the camera and how many frames you take to encompass the entire piece.  You're always using the sweet spot of a good lens and not having to rely on the edges of the images circle.

I have to agree with Bob.  It would be far less expensive to move the artwork.  The PTAssembler forums have quite a few threads on orthographic stitching that highlight some of the issues to solve: http://www.tawbaware.com/forum2/search.php?mode=results.

Making crossslides with the necessary tolerances will not be inexpensive and I expect you'll have trouble meeting your target costs.

[marcwilson]

posted elsewhere...

[BobDavid]

Leave the camera static and move the artwork around ala copystand setup. I've done this often and the results are fine.

[hovis]

hello everyone

thanks for all your responses and I'll try to answer any questions. I can't give exact figures yet on the amount of shift possible, before as Ron points out shadows cause a problem, this will depend to some degree on the model of SLR used. From my research though I am hoping that it will be possible to get somewhere in the region of a 5 x 7.5cm stitched sensor. The way I plan to work though (I am a one man operation), using a manual milling machine I see no reason why I wouldn't be able to adjust the design by request and fit, for example, a MF back which would have no problems with shadows. In this instance the only restriction would be how high the lens could be reasonably held above the focussing rack to get the full image circle of some much longer focal length lenses.

There shouldn't be any limit to minimum focal length other than the depth of the mirror box plus the thickness of the sliders. In the design I've shown the pinion block can simply be turned around to get closer. The only reason it faces the way it does at present was my own convenience, I only have one lens at the moment (an El-Nikkor 135mm) which in the image is placed at about infinity.

Thanks for your suggestion Bob but I'd rather keep it on topic and not get into a discussion about the merits of each technique, I've been researching and experimenting with photographing my artwork for quite a long time now and this seems like the best approach for me, it also seems to be quite widely recognized within the fine art reproduction community that one image as your source produces the best results, when moving the artwork any problems (lighting irregularities and lens distortions) contained within each frame are multiplied by the number of frames which make up your final image.

Lastly, I really don't think it's going to be that big a deal to get good enough tolerances in the machining, the price may change once I've actually made one, as I may find that things need to be done a little differently. For example, since yesterday I decided to change the way I intended to attach the bellows. At the moment I'm just trying to get an idea of whether there will be enough demand for this, my aim though is definitely to produce an affordable device which works well.

Kind regards

Jonathan

[asf]

i also think it's likely you will have issues with tolerances in your design

perhaps if you get one made that works and can show results people would be more able to tell you how interested they are
at that point you would also be able to know how hard it was for you to make and if you can repeat it easily enough
durability of the system would also be a concern - if you can achieve results within tolerances how long will they stay that way?

as it stands if what you have designed works (within your costs) you'd have legions of artists wanting one
many i know would love such a device

[hovis]

Thanks everyone

I've decided I'm probably going to take the plunge and invest in the tools, they'll be useful for other aspects of my artwork so I think it'll be worth it. It may be a little while but as soon as I have a working model I'll post some photos.

Regards

Jonathan

[BobDavid]

Just out of curiosity, Jonathan, what kind or artwork are you going to be photographing? I'm not sure a 135 EL Nikor and Foveon sensor are appropriate for fine art repro. The 135 EL Nikor is not an APO lens, which is highly recommended for digital repro. Looking at your XY axis configuration, I don't get the sense that there will be a high degree of stability and accuracy. It's pretty tricky to get optical planes perfectly aligned for parallel and perpendicular trueness. A cheap and accurate way to accomplish your goal would be to buy a used Horseman LE on Ebay for a few hundred dollars and buy an adaptor so that you can attach your camera to the back standard. Nikon and Canon mounts are readily available on Ebay for short money. I haven't seen anything for Sigma. In fact, you could probably get a used Canon 5D, a used Horseman LE, a view camera lens, and the Canon mount for under $2K.