Canon 5D Mk II versus 4x5 medium format

[Claude Fiddler]

A Canon 5D MK 11 versus 4x5 for 40x50 inch prints

I’ve been a large format photographer for 27 years and had 4 books published on my work www.wildernesslight.com. Here’s what I’m throwing out. I’ve been considering using stitching and a camera like the 5D mk2 to achieve 4x5 quality in a 40x50 inch print. And I mean a print you can stick your nose up to and say damn that’s sharp! I know one can do this with mega-mega pixel medium format but I’m not quite ready to sell body parts to get to that point. Also there area number of problems beyond price with m.f. I’ve heard countless quality problems with lenses, mis-aligned sensors, crappy cameras etc. Believe me I’m hearing this from the BEST photogs in the world. The guys who should be getting everything for free because they know what they are talking about. Example: Four Mamiya 80 mm lenses with only ONE that is sharp. Hasselblad lens literally falling apart, mis-aligned sensors..... Take note m.f. manufacturers your quality control isn’t .Claiming tolerances in the microns means consistently delivering this exacting quality. This means testing an m.f. system to death before one can be assured it is working properly. And how the hell can you test if you don’t have a comparison. I’m not interested in having to test a system costing tens of thousands of dollars. Plus the camera or the lens depending on manufacturer ends up being a big and clunky instrument. I can’t carry big and clunky 100 miles into the Alaska bush. I can’t have a lens fall apart. I can’t carry batteries for car starts.

So, is it possible and PRACTICAL to use stitching to make 4x5 quality prints from a 35 dslr. Think about this scenario- low after the sun has set light. A foreground that starts 5 feet away and goes out to infinity with detailed subject matter from foregrouncd to background. 2 rows of 4 frames, 2 focus points, one near, one far doing a focus blend with helicon focus. Photoshop’s does not work. Exposure blends are out as this totally complicates the equation into the realm of I’ll stick with some additional bulk and weight of 4x5 to avoid madness.

Does this work?. Not can this work, in theory this could work. My question is, does this work?

The comparisons I’ve looked at with an older 5D stitched print are close but still not 4x5. The phaseone p30+ did not measure up to 4x5. One glance at a 30x40 print was all it took to say not 4x5.

I’m looking for comparisons from photogaphers used to making large prints from large format and having made many, many prints. The comment here needs to be informed, not a guess like “I get great 16x20’s.....Not in the ballbark. 30x40 is a minimim size here.

Claude Fiddler

[Colin Southern]

Hi Claude,

First things first, welcome to the CiC Forums - it's great to have you with us.

As a point of reference, I shoot with a Canon 1Ds3 (same pixel dimensions at the Canon 5D2) I print (mostly 24" high and up to 88 inches wide, on canvas), and I'm occasionally involved in stitching multi-shot panoramas. So at least some of my experience is almost "in the ballpark" of what you're asking about.

I can see a couple of challenges in what you're trying to achieve, and the bottom line is I really don't know if they could be solved to your satisfaction.

First up, we really need to tie-down just what resolution you'd find acceptable - personally, I've never found much to be gained above 180PPI - so for a 40 x 50 inch image @ 180PPI you're going to need a 65MP image - so probably 4 to 6 shots, depending on overlap - that bit sounds quite doable. If you're wanting higher resolution (say up to 360PPI) then of course you'd need to quadruple the above number of images, which you'd probably find "problematic", at best (see below).

Other issue is in the stitching. Personally, my weapon of choice is Panorama Factory, but as of the last version I installed, it didn't support multi-row panoramas (there is a way to trick it into doing it, but I never managed to get satisfactory results). Thinking about it, if you were to shoot a single-row panorama (say a horizontal panorama using a series of shots taken vertically) then you'd be printing in the region of 140DPI (fine for canvas, but possibly not enough to qualify as "damn that's sharp" at standard photographer viewing distance (ie "nose length"), so you may well be committing to multi-row panoramas - and when I last played with some of these (several years ago) they weren't really up to a professional standard - keeping in mind that having to correct for the likes of barrel distortion must be a real nightmare for programs like this.

So sorry that I can't offer you anything more difinitive, but I hope that this helps in some small way. Happy to discuss this further with you if you feel it's of any benefit.

Kind regards,

[Dave Humphries]

Welcome Claude,

Cards on the table; I have no experience of this, so this is a guesswork answer (your least favourite), but I think I spot another problem looming.

Part of what you wrote made me think you might be shooting in lowlight/twilight, but I could be wrong.

The problem you may still encounter is that in the time taken to recompose/focus/expose a camera on a tripod upto 8/16 times, the sky could have changed significantly.
In this time the sun may have set completely, so brightness variations will be evident, but even if shooting at midday with short shutter speeds, clouds could cause issues. I'm sure judicous cloning may help the latter, but it all adds up in Post Processing time.

I guess practice and a slick routine will minimise the issue; i.e. if you know how many degrees to pan or tilt each time and have two pre-marked focus points.

Hope that helps,

[McQ]

Hi Claude. To add (and repeat) some of the previous posts...

In a short answer to your question: yes, of course you could make stitched prints using a 5DMkII that would be extremely sharp even at 40x50". A stitch using a 5D Mk II could certainly rival the sharpness/contrast produced by a 4x5 film setup at 30x40", or even surpass it, depending on the number of stitched rows.

However, I think the real question might be: would the stitched digital setup enable you to actually capture better photos? That's a tough one. Ease of use, comfort with the setup and speed of capture ought to also be factors. How much experience do you have with stitching? Doing a multi-row stitch would likely require two different exposures for dynamic range, in addition to potentially varying the focusing distance for improved depth of field. This is certainly doable, but can definitely make it more difficult to quickly and easily capture a wide range of subject matter under different lighting conditions. As pointed out, taking all the exposures for a stitch in low-light takes time, and this may make some subjects prohibitive due to changing light or subject movement.

Something else to consider: combining images to increase depth of field can have an odd effect at 40x50" if the subject is distributed throughout the range of near to far distances (and is not layered). The DoF can look unnatural and jumpy if not done just right. Further, I think you will find that DoF issues are more of a problem with a stitched photograph than one would ordinarily expect based on a given focus distance, lens and sensor format. This is because stitched photos have to overlap 20-30%, causing inherent inefficiencies (see the "Photo Stitching Efficiency Calculator" at the bottom of the tutorial on Photo Stitching Digital Panoramas).

One thing to try would be renting/borrowing a digital setup to see how much stitching you feel is necessary to equal that observed in your current setup, and whether you feel at ease doing so. There's definitely a subjective component to judging image quality, particularly when it comes to noise levels and local contrast -- this aspect is hard for me to assess.

Overall though, I hope all of these caveats do not discourage you from digital altogether. There's many other inherent advantages to a digital setup which are not mentioned here...

PS: I know you've already seen most of this post via email, but I thought I would paste it here for inclusion in the thread.

[BillTexas]

The last time I went to England with my brother and his 4x5, I carried it for him. Simple solution for him and he got some great photos, some of which are enlarged and look great.

The bottom line is you are never going to get the quality you want by stitching together a bunch of digital images. And you are going to spend more time on the computer than if you had just taken the 4x5 with you.

There are backpacks made for this size equipment and they work great and still have room for some food and other stuff.

Every camera we use has limitations and that is just as true or digital cameras as film cameras. The Canon is a great camera, but it's not a 4x5 and that's the simple answer to your inquiry.

My brother also has an 8x10, but there are limits for the filial donkey. But he does have a backpack for it. And I have carried it around closer to home. I don't even want to talk about his tripod.

BillTexas.

[Claude Fiddler]

Thanks for the replies.

I know well there's no simple answer here. The best thing to do is as Sean suggests; rent a 5D and try to merge a photo with near,middle ground and infinity. Try a focus blend and see if this is a practical solution. See if one row works for a 30x40 at least. Try 2 rows and see if this works.

I'm well versed in carrying a 4x5 into extremely remote places in extremely difficult situations. I've thought about getting a lighter camera, but the Linhof is really hard to wreck. I tried in the middle of a 100 mile backpack in the Alaska Brooks range. Camera blew over in a wind, broke the ground glass and film back. I had to duct tape the film back on and off to use the camera, which I did for the rest of the trip.

I did not really miss any photos that I wanted to take because of 4x5, but then I wasn't paying attention to all photo possibilities. Use a 4x5 and it's easy to talk oneself out of making a picture. I did use a point and shoot on a recent Alaska trip and was excited by some of the more spontaneous pictures. This was probably the first time in 25 years of large format use that I thought about using a smaller, more versatile camera.

Well, onto an experiment and I'll let you know how it goes.

Claude

[Colin Southern]

Well, onto an experiment and I'll let you know how it goes.
Claude

Claude, thanks for the update - I'm sure that many of us are very interested to hear of your journey down this path. In fact, I'm sure I speak for everyone when I say that we'd all like to hear more of your journeys and see more of your work here; and again, if any of us can help you in that regard, please don't hesitate to ask - we value your contribution.

[Claude Fiddler]

Thanks for the compliment.

I should have FREE video downloads on my website later today. Alaska trips included.

I'll let you know.

Claude

[Colin Southern]

Thanks for the compliment.

I should have FREE video downloads on my website later today. Alaska trips included.

I'll let you know.

Claude

Thanks Claude, I'll look forward to watching them.

Is there anything that we might be able to help you with with regards to 35mm equipment (lens selection, processing etc)? (I shoot a lot of landscape myself, and am pretty familiar with the high-end Canon cameras and lenses).

[Claude Fiddler]

And thanks again

For my experiment I'll use an old 5D. A friend has one. Use Sean's rotate over the foot method to shoot panos and use helicon focus.

As far as possible future purchase I'm looking at the 5D mk2 with a 24-105 IS lens. This lens/camera weighs around 10 pounds with accessories. Weight is critical!!! At this time I'm carrying 85 pounds into the backcountry for a ten day trip. I've stripped my 4x5 set-up down to 15 pounds. I take one lens (a 150mm Rodenstock) and one box of Astia readyload film. Canon setup with stitching hardware ends up weighing ten + pounds. A medium format set-up with one lens weighs ten + pounds. I looked at the Leica m8 with three lenses but at ten megapixels I didn't see practically how I could get the resolution I'm after. Multiple Lieca lenses would be something of a plus Back-ups in case a lens breaks. I have to think of these things. There's no Fed-x 400 miles from the nearest human presence.

Ive looked at Alpa, Mamiya, Phase One, Hasselblad, Mamiya and Leaf back, Mamiya and Mamiya Back. If one digs deep or just scratches the surface all of the systems end up, for me with fatal flaws.

The ONE system that held some promise was the Leica S-2 but lack of a zoom lens and cost will probably (whenever this appears) spell F.A.T.A.L. with that system.

So as you can see I'm trying to solve a number of problems, weight, ruggedness, compactness, ease of initial use, softwear capabilities, and the ephemeral "this feels good in the hand" part. Claude

[Colin Southern]

Hi Claude,

It sounds like you're really pushing the boundries of the technology.

As you mentioned, it's important to rotate the camera around the nodel point (the common name that Sean points of isn't technically accurate) if you have forground detail. For this I use a 24" nodal bar (and panning head) from Really Right Stuff (www.reallyrightstuff.com) - in fact they have all the hardware you'll need for making multi-row panoramas, and everything they make is top-quality stuff. I haven't visited their site for a wee while, but you should be able to download a complete catalog (with a sample multi-row panorama off memory).

For stitching you might like to take a look a Panorama Factory (www.panoramafactory.com) - I see that V5 now supports multi-row stitching; it's a very sophisticated package supporting the correction of the likes of barrel distortion etc (free trial).

Hope this helps!

[Claude Fiddler]

Thanks

I've looked at rrs and know of at least one high end user. The machining is quite beautiful but their system adds almost 3 lbs to the tons of lightweight gear we end up hauling about. I'll look at the pano site. Also looked at nodalninja. Check this one gigapansytems.com.

I almost, correction I did laugh. Good Lord we are techno geeks!!!!!!!!!!!!!!!

C

[Colin Southern]

Hi Claude,

I guess that the days of Ansel Adams where he would disappear into the bush with a pack horse and 2 plates are long gone!

I must admit to feeling a little "long in the tooth" for that kind of effort, prefering "heli-tramping", the concept of which was brought home to me when a paramedic friend of mine once had to be airlifted out of the bush after getting infected blisters 3 days into a tramp. Took him 3 days to get in ...

... took the helicopter 20 minutes to get him out. There's something inherantly lazy, but none-the-less satisfying about that idea

[Policar]

Sorry to up an old thread, but what did you choose to do? I'm curious since I'm debating trying lf.

I believe the current stance from the non-purists' camp is that a 5D II compares well with 4X5 when printed large and even competes with 8X10. Mathematically, this doesn't quite hold up, as the 5D II should really only manage a critically sharp 300dpi print at 12''X18''. But a 40 lp/mm (not the theoretical maximum but a fairly sharp print for lf after scanning) 4X5 scan is about 70 megapixels of real resolution, which translates to a 24''X30'' 300dpi print. Those who shoot digital claim the far superior tonality and lack of grain more than negate the difference in resolution, which many believe doesn't even exist. So while your 5D II images may-or-may not look as good close up as lf, from a reasonable distance they should look just as good. Using intelligent upscaling, sharpening, and adding a fake "film grain" overlay can increase perceived acutance and push the 5D II to possibly slightly better than 4X5--but it might look fake to a film purist.

Anyhow, the few shots you have on your website look incredible so why not stick with film? What do you want out of digital? Stitching and focus stitching both look terrible to me and are a pain and I would just live with the slight loss of resolution rather than use them...you get intermediate focus areas that never look right and then there's the need to correct for the major distortion on retrofocus wide angles before stitching--and even then I think there are issues.

I think the next great nature photographers, those who will make names for themselves as more than imitators and luddites, will make heavy use of digital-only technology and techniques: hdr, tonemapping, digital dodging and burning, high-iso image stabilized handheld, etc. Unfortunately current hdr generally looks absolutely horrible, like someone applied "shadows/highlights" in photoshop alongside a wide-radius USM to increase local contrast. But most non-connosieurs far prefer average quality hdr to goood 4X5 professional prints. So why not buy a digital camera and use it like a digital camera instead of a replacement for lf?

[Colin Southern]

Perhaps I'll start printing my work at a maximum of 1.5 inches x 1 inch - should be able to get incredibly high resolutions that will really stand up to inspection under the microscope ...

... literally!

[Policar]

I don't think you need to limit yourself to 12''X18'' for digital unless you like staring at the prints close up. From a distance, the better acutance and local contrast should make a 5D II look sharper than 8X10, though I have not tested this.

I've studied 1X1.5 inches (35mm slides) under a microscope. Not actually that sharp! But point taken.

[Inkanyezi]

I've been looking through this thread and wonder whether it really makes sense. To me it seems simpler to get a Phase One or Sinar with a 65 megapixel back and just shoot away. Maybe it's a bit more expensive, but it will do the job without stitching.

Real world problems with digital imagemaking go far beyond pixel count, and sharpness has many facets. Dynamic range is one of those issues, and undeniably, the larger format has nicer tonal quality and better dynamic range. And you need no stitching for a 65 megapixel shot, although I would prefer to switch the back down to 16 megapixel for low light, decreasing noise and increasing dynamic range, improving tonal quality in the darker regions of the image. The photon wells are bundled 4 and 4 to accomplish this, so the 65 megapixel back becomes 16 megapixel. Both Phase One and Sinar do this.

[Colin Southern]

Maybe it's a bit more expensive, but it will do the job without stitching.

I suspect that "maybe" is a wee bit of an under-statement!

[Inkanyezi]

I suspect that "maybe" is a wee bit of an under-statement!

Understated, yes, you read it right, but I thought that your pictures might be worth it, considering also other expenses. A state of the art tool can be a good investment.

[Colin Southern]

Understated, yes, you read it right, but I thought that your pictures might be worth it, considering also other expenses. A state of the art tool can be a good investment.

Well Claude's pictures perhaps, not mine -- I shoot with a 1Ds3 which has more than enough pixels for my humble 22 x 44" prints

To be honest, when I compare the stats of the likes of the Phase 1 and 1Ds3, the Phase 1 somewhat underwhelms me.