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!

  • HAPPY THANKSGIVING TO ALL OUR AMERICAN FRIENDS!

Technical Camera vs View Camera

Abstraction

Active member
The technical cameras seem to be all the rage these days. They're compact, easy to transport and offer view camera movements, as well as interface support for various digital backs.

However, I was wondering why the view cameras don't seem to get any love? I could be wrong, but they could be made to interface with the digital backs just as easily and whereas they don't offer the same portability as the tech cams, they seem to offer greater movements. Furthermore, they're MUCH cheaper than tech cams and the lenses can be had for a song compared with tech cam lenses.

Is there a reason the view cameras have gotten the shaft?
 

GrahamWelland

Subscriber & Workshop Member
In a word, precision. Very few full sized view cameras, and I assume you mean 4x5 etc, have the required precision to support the focus accuracy required by digital backs. That's not to say that it can't be done but the lenses you might have been using with their film oriented design and tolerances don't really stand up so well when used with a digital back. Also some of the rear movements aren't ideal for use with a small sensor located in the middle of the rear standard.

Now if by view camera you mean the likes of the linhof techno or cambo actus db then I would say that they are indeed relatively popular but do require use of either GG or live view.
 
Most MFDB are smaller than 645, which is tiny comparing to a 4x5 frame. So if you do use a MFDB with a 4x5 (or larger), you will either have to be stitching, or just wasting a lot of real estate. Also, you will be using short lenses. Now supposedly some of the advantages of MFDB+Tech Cam over 4x5 are that it is smaller and easier to use, but I am not entirely sure. A lot of tech cam users seem to stitch and rely on liveview for focusing. In the end, I think the workflow is different and it is what you are more comfortable with. Subjectively speaking, the high end MFDB definitely has an edge in terms of sharpness and most other image quality metrics. OTOH, 4x5 + film is a visceral experience. I just exposed 41 sheets of 4x5 Portra in Death Valley (and 9 rolls of 120) in a whirlwind tour, and I am not sure a MFDB+Tech Cam would have been easier or faster to use.
 

torger

New member
There was a period when digital backs where used with 4x5" view cameras, without gearing. This works fine for table-top photography when parallelism is not much on an issue and you don't use wider angle lenses.

However when you want to move to applications such as architecture photography or landscape photography with wide angles parallelism became a big issue. Another issue is that digital sensors are tiny compared to large format film which make the ground glass tiny, so focusing becomes a challenge. Pixel-peeping and other digital cameras also raised expectations on how precise focusing should be.

The CCD digital backs did not have any live view at all or just a sad excuse for a live view. In addition, almost all digital backs of older age doesn't have a screen/onboard demosaicer which is good enough to actually see if the picture is sharp when looking at 100%. This means that if you focus on the ground glass and feel uncertain you couldn't really check in the field if the shot was sharp (unless you brought a computer and shot tethered, which indeed was not too uncommon).

The solution to all this was the "pancake cameras", especially those with high precision focusing rings. Alpa and Cambo can have them mounted, and Arca-Swiss RM3Di has extremely precise focusing built-in. Instead of focusing on ground glass you either use hyperfocal presets or use a Leica Disto to measure distance which you then just dial in. Arca-Swiss also have their own distance measurement gear which you can use. In addition to solving the focusing problem, they are extreme rigid and precise so parallelism with wide angles is not an issue.

However, there are also "digital" field view cameras. Linhof Techno, Arca-Swiss MF-two and Arca-Swiss Universalis, Cambo Actus. The Actus is very compact cannot carry sliding backs. Those have geared movements for all or most settings, and have better precision and parallelism than the 4x5" cameras. Still with the rich movements they're not as rigid as the pancake cameras and I'd recommend to not shoot wider than say f/11 -- at f/11 the depth of field masks the tiny residual errors you have.

With live view-capable backs the view cameras don't have the ground glass challenge any longer which have made view cameras more popular, and very slow super-precise focusing mechanisms like on the RM3Di less popular (Alpa/Cambo has a faster focusing mechanism that work well also with live view, a better choice in my opinion). However due to the wide angle compatibility issues the CMOS backs have not become the huge hit they could have been, at least not yet. But it has made a difference still, there's several view camera users with CMOS backs these days.

Then there are a few of use that actually use sliding backs and ground glass and can then use more or less modern CCD backs. Ground glass requires some skill and good eyes, and also a good quality ground glass and a good loupe. A problem has been that sliding backs have often had various precision errors in themselves (especially legacy ones), have had grainy dim ground glasses and the loupes have been too small magnification. In all there's been a high risk to get a very bad ground glass experience which has added to its poor reputation. However if you get the best sliding backs and best ground glass, like Linhof's recent bright ground glass with a high magnification loupe, there's a 14x that Linhof Studio has which is popular, I use a 20x myself actually, you can indeed focus at f/11 precisely in most conditions (if there's nothing to focus at or it's very dark you can't of course, but to me it has not been any significant issue, I do bring a laser pointer to help in those conditions but I haven't used it since I bought it).

I've have a Linhof Techno myself and I've looked a bit at Arcas view cameras. I think the Linhof, despite a few issues, is the best digital field view camera there is. The Arca-Swiss Universalis is also good but is messier to pack and lacks a degree scale on the tilt (which is very useful to have when tilting for wides, when you tend to use tables). Linhof doesn't have that good reputation on the forums, and I think one issue is it's poor representation in the US. In the EU Linhof Studio has made the camera quite popular. Anyway, don't take my word for it, the camera preference is a very personal thing.

Another reason field view cameras are less popular is that most that get a tech camera unless they are studio photographers, are interested in wide angles. Some even get just a single lens, a very wide angle. In that case you don't really gain much from having a view camera, instead you see more of the view camera drawbacks. The view camera is most useful if you use several lenses including longer ones. If you buy only one lens the view camera kit is often more expensive than a pancake camera, but if you buy several lenses the lens board vs helical focus cost difference becomes clear (about $1000 per lens), and if you use longer focal lengths the weight and space of the barrels compared to the flat lens boards also become a clear difference.

The reason I'm using a field view camera instead of a pancake camera like an Alpa is because I use many lenses, I shoot more often normal-to-longer lenses than ultra-wides, I appreciate to fit all those lenses in my camera backpack, and indeed cost is also a factor, with my seven lenses I've saved like $7k in lens mount cost. Would I say shoot architecture professionally I'd use a pancake camera, although there are those that use the Techno anyway, like Sean Conboy. Christopher Barrett is another architectural photographer using view cameras, Arca-Swiss, nowadays often with a Sony mirrorless as "digital back".
 

Abstraction

Active member
Thanks for those replies. I never would have thought that focusing would be an issue. Is the MF digital back's resolving power greater than 4x5 film? Otherwise, why does it require greater focusing accuracy?

As far as wide angles are concerned, people used to mount roll film on view cameras. Were there not wide enough lenses available to accommodate roll film backs? If there were, could these lenses be used? Why weren't they worried about parallel focus rigidity?
 

torger

New member
Thanks for those replies. I never would have thought that focusing would be an issue. Is the MF digital back's resolving power greater than 4x5 film? Otherwise, why does it require greater focusing accuracy?

As far as wide angles are concerned, people used to mount roll film on view cameras. Were there not wide enough lenses available to accommodate roll film backs? If there were, could these lenses be used? Why weren't they worried about parallel focus rigidity?
If you look for grain-free resolution 4x5" film is about 40 megapixels. If you see through the grains its considerable more though (look here: http://static.timparkin.co.uk/static/tmp/cameratest-2/800px.html). But the surface is also larger. In a way the digital backs are large format resolution shrinked down to a smaller format, and with everything smaller you need smaller movements. It all comes down to depth of focus, with 4x5" you shoot often at say f/22, with MFD you shoot f/11, hence smaller depth of focus, and hence more precise focusing required.

I don't have experience of the rollfilm+analog wides so someone else has to comment on that. In general though I think digital pixel peeping has raised the expectations of precise focusing, plus the smaller format requires higher degree of parallelism due to the smaller depth of focus.
 

tjv

Active member
Like Torger, I use a Linhof Techno with a Credo 60 and make use of the sliding back and ground glass. Before getting the digital back, I used the Techno for three years exclusively with a roll film back. It's a brilliant system and I've never had any problems with rigidity etc–let alone encountered focus problems–save for lack of light to find a focus point (where by design a pancake camera would be very easy to use.) Thankfully I don't tend to use my Techno to photograph mine shafts and black cats in said mines, etc...

I can only echo what Torger has said above, that it's a very versitile and very much underrated system and I would certainly buy it again, especially because I appreciate the process of using it–more so than the way that one uses a pancake camera to focus and assess composition. That's my very subjective opinion formed from careful consideration of the work I do, so YMMV. It's also worth noting that on a full frame DMF I shoot between 40mm and 90mm, never wider or longer. Again, as Torger has said above, any wider and the pancake cameras tend to win out.

In terms of cost of entry, I found at the time the Techno, which I imported from the UK to NZ, was vastly cheaper than a comparative Alpa STC kit (never checked out the Cambo or Arca.) The lenses can be bought mounted in standard Technika lens boards vs. needing special helical mounts, so for the two lenses I bought when first ordering the kit it would have meant I'd needed to have stumped up 50% extra cash for a less versatile set of movements if I went for the Alpa. Things may have changed since then and Cambo especially offer some very attractive pancake systems, it's just that they don't interest me now because I'm happy with the Techno for what I need it to do.

Now with the advent of excellent CMOS digital backs, I think the Techno is coming into its own. People can dispense with the sliding back and GG and be left with a very easy to use, small package. Again, lenses are still cheaper due to the mount, etc, etc. With CMOS and good live view it seems to me that the super precise helical mounts are a solution looking for a problem. Of course there are also attractively priced packages in the new Cambo Actus and equivalent Arca if one is using a CMOS back, but handing an Actus the other day I wasn't impressed with the build and rigidity compared to the Techno. Then again, the Techno is a lot more money and I'm probably biased towards what I know well!
 
If you look for grain-free resolution 4x5" film is about 40 megapixels. If you see through the grains its considerable more though
I sold my LF camera years ago, but funnily enough, I just had a sheet of 4x5" (admittedly using the very sharp and grain-free Acros 100) that was drum scanned to 350mb yesterday, and was having a good look at it last night.

At 50x40" at 300dpi, grain is almost invisible (even with a lot of sharpening), and I was surprised at just how much more real detail is in that 4x5" file ..... compared to a 38mp camera I've played with recently (a Leica S) at that sort of 50" wide print.

I think if I turned a 38mp Leica S vertically on its side, however, and rotationally stitched 2 or 3 images together to produce a 50"x40" print, it would probably all look a lot closer to the 4x5" drum-scan -- at that stage, the "long side" of the 38mp sensor is being asked to achieve 40" in print terms, which isn't a stretch.

I hope this isn't veering off track, but what do you see as benefits from rotational / nodal stitching (eg, using a DSLR), versus in-camera stitching (eg, using a tech cam like an Alpa STC)? Note, I have NOT bought a nodal plate yet, but am trying to simulate a nodal stitch as best I can handholding the camera, and every time the 3 files are merged, it create a shape like a "Bow Tie" ..... eg, I have to cut off the top and bottom of the Bow Tie on both sides of the aggregate file. Is that what I should expect to occur with nodal stitching, EVEN when I eventually get everything perfectly aligned using a nodal slide??

And in comparison, what does one get with something like an Alpa STC for 3-stitch effort (middle shot, then left- and right-stitch?) - presumably everything is perfectly lined up, with no bow-tie effect?
 

torger

New member
I sold my LF camera years ago, but funnily enough, I just had a sheet of 4x5" (admittedly using the very sharp and grain-free Acros 100) that was drum scanned to 350mb yesterday, and was having a good look at it last night.

At 50x40" at 300dpi, grain is almost invisible (even with a lot of sharpening), and I was surprised at just how much more real detail is in that 4x5" file ..... compared to a 38mp camera I've played with recently (a Leica S) at that sort of 50" wide print.

I think if I turned a 38mp Leica S vertically on its side, however, and rotationally stitched 2 or 3 images together to produce a 50"x40" print, it would probably all look a lot closer to the 4x5" drum-scan -- at that stage, the "long side" of the 38mp sensor is being asked to achieve 40" in print terms, which isn't a stretch.

I hope this isn't veering off track, but what do you see as benefits from rotational / nodal stitching (eg, using a DSLR), versus in-camera stitching (eg, using a tech cam like an Alpa STC)? Note, I have NOT bought a nodal plate yet, but am trying to simulate a nodal stitch as best I can handholding the camera, and every time the 3 files are merged, it create a shape like a "Bow Tie" ..... eg, I have to cut off the top and bottom of the Bow Tie on both sides of the aggregate file. Is that what I should expect to occur with nodal stitching, EVEN when I eventually get everything perfectly aligned using a nodal slide??

And in comparison, what does one get with something like an Alpa STC for 3-stitch effort (middle shot, then left- and right-stitch?) - presumably everything is perfectly lined up, with no bow-tie effect?
The film-vs-digital resolution comparison is very coarse and depends on many factors, film type a major aspect of course. Personally I think film have an advantage of looking good even if over-sized. No-one likes seeing pixels when nosing a print, grain has a more pleasing appearance, and indeed it's quite common to add grain simulation on digital prints to have them look more pleasing in huge format. Film also resolves monochrome detail way past the grain which gives it a special quality. Saying that 4x5" is about 40 MP digital is if you want to feel good about digital, but if you dig past the grain into the true max resolving power you get up to 380 megapixels out of a 4x5", again using Tim's test as a source: https://www.onlandscape.co.uk/2011/12/big-camera-comparison/ , but that is a bit theoretical.

Anyway on to the stitching question.

The bow-tie effect is normal, and that is because you convert your stitch into a rectilinear projection. If you have shot an ultrawide lens sometime you have probably noticed the stretched perspective along the sides of the frame. When you stitch with a longer lens and want to make the same field of view as a shorter rectilinear lens you need to make the same stretch, hence the bow-tie effect. If you make say a cylindrical projection instead (which can work well for landscape panoramas) you don't need to cut away as much. In any case getting a nodal plate won't change anything of that.

When you stitch inside an image circle the sensor is already seeing a rectilinear projection so there is no projection remapping required, and this is how you typically stitch on tech cams.

An advantage of nodal point stitching is that you always use the sharpest part of the lens, the center, even to render the "corners" of the final image, and of course you don't have any limit on how wide angle view you want, just shoot more frames. Nodal point stitching can replace both shifting and wide angles, but it means more computer work and thus a less effective workflow. I have only used Hugin as stitching software (highly competent software, but not user friendly), but there should be better alternatives, still I hear often than people have trouble in recreating the proper perspectives (which they succeed with when using shift on a tech cam). In theory it should not be any issue, so I think it's related to software. Anyway, the point is that with nodal point stitching there seems to be more of a challenge to get the end result you want, but if you do get deep into it it's possible.

With my DSLR I have a nodal point stitch head and I made some stitches. You can get a huge amount of megapixel in any projection you want. However time-consuming and I did not find it a very pleasing way to make images, just too mechanical and the reprojection made the composition a bit unpredictable. With the tech cam the workflow is typically a little bit more efficient, less overlaps, less shots and you don't need to reproject the images so the composition is more predictable. Still I personally don't stitch even on my Linhof as I'm a fan of the one shot image shooting experience, I rather use a wider angle and sacrifice some resolution than stitch, but that's me.

Here's an example from my DSLR days when I made a cylindrical stitch for a huge panorama (disclaimer: this image is so much not my style these days :) ). These type of perspectives does not hurt from cylindrical projection, I'd say they even gain from it as you don't get stretch effects which can be a bit ugly at times. When shooting architecture or other stuff which straight lines in it you need to use rectilinear projection though. I *think* there's software today where you can mix cylindrical and rectilinear in the same image, for example if you have an extremely wide panorama with a bunch of smaller buildings in it, you can make sure that each building has straight lines but the overall panorama is cylindrically projected. Again, lots of post-processing work in that case.

cylindrical.jpg

And here's an example of that bowtie effect, before cropping down. To make the lines render straight a rectilinear projection is required and then stretching on the sides is necessary:

bowtie.jpg

And finally the exact same image material projected cylindrically, before cropping. As you can see you will loose much less, but you get that "fish eye effect". However you only get bends in the horizontal direction, vertical lines are rendered straight (eg tree trunks), this is what makes cylindrical projection often work well for landscape panoramas.

cylindrical2.jpg
 
Last edited:

jlm

Workshop Member
was in a remarkable location, so about 10 nodal stitches, slight bit of goofiness on the top/bottom, processed in autopano giga 2.6 (photoshop is generally worse, and even choked on the gigs) shot with the IQ160, 43SK, vertically, about a 270 degree view. took some tweaking to get the sky a bit more uniform
 

torger

New member
was in a remarkable location, so about 10 nodal stitches, slight bit of goofiness on the top/bottom, processed in autopano giga 2.6 (photoshop is generally worse, and even choked on the gigs) shot with the IQ160, 43SK, vertically, about a 270 degree view. took some tweaking to get the sky a bit more uniform
Looks like a cylindrical projection, and a great demonstration that it actually does work on architecture if the conditions are right. As each building takes up quite narrow range in the horizontal direction you don't get any significant bends.

We got off topic in record time, but still interesting subject :)
 
That's all really helpful, thanks everyone. Looking now at the alternative you posted Torger, I personally prefer the image shaped via the "bow tie" -- I will now contentedly continue to crop off the extraneous bits on the left and right hand sides!
 

Abstraction

Active member
We seem to have gone off on a tangent with the stitching discussion. I'd like to get back on track a bit.

I am still at a loss why focusing would be a bigger issue with digital than with film? Let's assume you have a back with live view or you're tethering, why would a view camera focus more coarsely than a tech cam? Why can't you focus it the way you want? I understand the wide angle issue, and it's a big one, but as far as focusing, I don't understand why it's fine to focus on a film plane and not on a digital back? Furthermore, the parallelism issue doesn't make sense to me either. Even if the front and rear standards aren't perfectly parallel, it would be less of an issue with a smaller format, such as that of a digital back. There has to be an incredible diversion among the planes in order for it to matter to a much smaller format mounted centrally within a 4x5 field of view.

The DOF argument doesn't quite cut the mustard here because the f22 DOF on a 4x5 format would be roughly equivalent to f11 on a 645. Furthermore, you can always stop down if you need greater DOF.
 

jlm

Workshop Member
usual response to this is that the film grains are distributed through the thickness of the emulsion, whilst (!) the sensors on DB are more in a true plane
 

Geoff

Member
its a tolerance issue. You are working with much less, forward backward. Its not that you can't get it to work, its just that you can also miss.

Torger mentioned parallelism - you'd think front and back standards are parallel. Take a view camera, shoot with a lens wide open on a digital back (not even across the full 4x5, and check side to side. You can often find where one side is sharp, the other isn't. This is something we rarely found on film.

Many times, the standards are out of parallel (by just a wee bit), and sometimes the lenses are out too. If you shoot at f11/f16, on a sensor (smaller in size than 4x5) you can usually get away with it, but not wide open. And yes, f22 on a big piece of film covers up a lot of sins...

Simply put, the whole "digital back on view camera" requires a lot more precision than film. Abut 10X less tolerance.
 
Last edited:

Abstraction

Active member
its a tolerance issue. You are working with much less, forward backward. Its not that you can't get it to work, its just that you can also miss.

Forger mentioned parallelism - you'd think front and back standards are parallel. Take a view camera, shoot with a lens wide open on a digital back (not even across the full 4x5, and check side to side. You can often find where one side is sharp, the other isn't. This is something we rarely found on film.

Many times, the standards are out of parallel (by just a wee bit), and sometimes the lenses are out too. If you shoot at f11/f16, on a sensor (smaller in size than 4x5) you can usually get away with it, but not wide open. And yes, f22 on a big piece of film covers up a lot of sins...

Simply put, the whole "digital back on view camera" requires a lot more precision than film. Abut 10X less tolerance.
I'll take your word for it, but I don't understand why that would be the case. You're working with the same plane one way or another. Unless there's a tremendous difference in resolution, one plane should be as good as another. In other words, why wouldn't you see these issues on a film plane?
 

torger

New member
We seem to have gone off on a tangent with the stitching discussion. I'd like to get back on track a bit.

I am still at a loss why focusing would be a bigger issue with digital than with film? Let's assume you have a back with live view or you're tethering, why would a view camera focus more coarsely than a tech cam? Why can't you focus it the way you want? I understand the wide angle issue, and it's a big one, but as far as focusing, I don't understand why it's fine to focus on a film plane and not on a digital back? Furthermore, the parallelism issue doesn't make sense to me either. Even if the front and rear standards aren't perfectly parallel, it would be less of an issue with a smaller format, such as that of a digital back. There has to be an incredible diversion among the planes in order for it to matter to a much smaller format mounted centrally within a 4x5 field of view.

The DOF argument doesn't quite cut the mustard here because the f22 DOF on a 4x5 format would be roughly equivalent to f11 on a 645. Furthermore, you can always stop down if you need greater DOF.
"Depth of focus" is not the same as "Depth of field", I was talking about the former. Depth of focus is the depth of field behind the lens, that is in the film/sensor plane and the formula is DepthOfFocus = 2*FNumber*CircleOfConfusion. That is the larger f-number the larger depth of focus. So f/22 is twice as thick as f/11.

Concerning parallelism on the smaller format, you're extracting a similar resolution out of a smaller format, if you make the same print size you are enlarging the smaller format more. The more you enlarge, the more you enlarge any error. Think of taking your 4x5" camera and shrinking it. Every movement must be smaller. Ground glass grain needs to be smaller, and loupe magnification to larger if you want to view with the same precision as the 4x5" ground glass of course.

Smaller format means shorter focal lengths means more precise tilt and swing angles. When focusing wide angles I set fractions of degrees, for example to get a hinge distance of 1.25m on a 35mm lens I set 1.6 degrees of tilt. On a 4x5" that would be a 90mm lens set to 4 degrees of tilt.

I think there is also a large psychological factor that with digital and pixel peep it's much easier to precisely inspect focus on your computer, and thus people have got more picky about focusing.

But the key technical reason is smaller camera, similar resolution. With film if you wanted more resolution you shot larger film, the whole camera became bigger. Now we make smaller pixels and sharper lenses. I guess you could say the new IQ3 100MP presents similar resolving power as 8x10" film (grain-free), but to make prints the same size you need to enlarge almost 5 times more so thus the precision requirement is higher for the smaller format.
 

torger

New member
usual response to this is that the film grains are distributed through the thickness of the emulsion, whilst (!) the sensors on DB are more in a true plane
I've heard this too, but I don't understand how that would work. It sounds like the film would "choose" only to respond to the best focus wherever it is in the emulsion, and I don't think film is that smart. But I don't know...

Anyway the explanation that to make the same print size the smaller format needs to be enlarged more is the simplest explanation. Say if you have an 8x10 digital sensor with 100 megapixels on it and a 645 digital sensor with 100 megapixels on it, both sensors are capable of the same resolving power and can make the same sized prints, but the 645 requires smaller everything, and thus larger precision.
 
Top