MFDB vs Large Format - where are we today?

[PaulSchneider]

Hi guys,

I know that I'm sticking my nose in a bee's nest, especially considering that it's my first post in this forum ...

... but here comes a sincere question from someone who is pondering large format photography as a means to getting into ultra high resolution prints. The last months I thought that medium format digital is the be and end all in terms of photographic quality, especially with those high prices in mind I thought it must be the best.

After a lot more reading though, I reckon that to achieve the highest quality possible analog is still the way to go?

My question is this: In 2010, with the p65+ and H4D-60 (soon, apparently) around as the best digital backs, how does medium format digital compare to large format photography?

Can one achieve similar print sizes or not with digital means? Or is ist still a completely different ballgame? After more than a year after the p65+'s introduction I'd be happy to hear some experiences?

What resolution can one squeeze out of a 8 x 10 negative with the best available hardware? somewhere I read that it is about 19500 x X pixels?

I know that here's a thread about this, but the this picture got me thinking .... :

http://www.gettyimages.de/detail/85399865/Getty-Images-Entertainment

...

Kind regards

Paul

[evan clarke]

I think the tactile quality of the print is way more important than resolution an ink jet print is still just an ink slick. If the viewer is making a decision by inspecting a print with a loupe, your subject matter has lost the war...EC

[BetterSense]

I agree with evan clark. If you have no desire for analog, use digital. If it's all about pixels, you might as well start with pixels.

[David de Gruyl]

Ok, is a 6x4.5 (cm) frame the same as a 4x5 (in) or 8x10 (in) frame? Even if you use the worst film in the world, you will have more information on the larger "sensor". I would even argue that the 6x4.5 digital back is less interesting than the fleet of 6x7 gear you could replace it with.

All that being said, the larger format has drawbacks, primarily in weight (for 8x10) and in limited depth of field. You may have more light striking more photoreceptive elements (silver halide molecules or whatever) but more of it has the potential to be out of focus. The larger the area, the smaller the DOF. This is why a, 8x10 portrait looks so much different from a MF portrait (in my mind).

Also, processing, handling and scanning a piece of 8x10 film is non-trivial. You might consider 4x5 as a better size for resolution gathering. In fact, I get better scans at higher resolution from 4x5 than from 8x10. I also don't feel stupid putting color film in the 4x5 (because I print color digitally, there is no benefit derived from the 8x10 for color).

This does not stop me from using the 8x10, but I do consider part of what I do performance art. The other part is making prints. So far, nothing beats a contact print. 8x10 is a much more reasonable size contact print than 4x5. It is still on the small side, but I am unlikely to get a bigger camera. No room for it.

[erie patsellis]

If resolution and large print sizes are your goals, you may want to consider a Betterlight scanback.

[Frank Petronio]

All that being said, the larger format has drawbacks, primarily in weight (for 8x10) and in limited depth of field

I see those as advantages ;-)

The question is moot, one can always do a better scan from the film and digital backs will continue to improve. For all practical commercial uses, digital rules.

But only with large format do you get the size, weight, and three-dimensionality that a smaller film or sensor size will never be able to capture. Not that you can't make great images with small cameras, but especially with shooting people with a big camera, it is a different experience in itself, which changes the way the subject responds.

Holding a H3 up to someone's face ain't the same as setting the Deardorff up!

[David de Gruyl]

Holding a H3 up to someone's face ain't the same as setting the Deardorff up!

Uh, exactly.

Or, for that matter, waving a crown graphic around has a completely different effect from any other type of camera. People tend not to care if you stick one up their nose, and fire 1000 watts in their eyes. (ignore the number, I have no idea what my metz hammerhead yields).

[Steven Barall]

First you have to define "image quality". If it's purely a math problem then a properly scanned sheet of large format film wins by a long shot. An 8x10 inch piece of film has many many times the amount of data than a MF digital back so that's that.

All of the other stuff is subjective. You have to use whichever tool fits the job best for you. Whether one is better than the other is up to you and your personal needs.

Which is better, a hammer or a screw driver?

[williamtheis]

I agree with erie patsellis: king of the hill is betterlight but it comes at a cost. Long exposure times and problems with artifacts if the subject moves (wind). I shoot 8x10 and scan with respectable scanner (Heidelberg Tango drum) but it does not approach the betterlight. Also there are no cords and batteries to film. there is a clarity and color purity not matched by focal plane sensor DSLRs. Having said that, I don't have access to a medium format back for comparison mainly due to expense. a used betterlight goes for $8K, far less than medium format but you have to add the camera and it must be a rigid one (I had to replace my wooden folder with an Arca Swiss). this also gives the opportunity to shoot both 4x5 and betterlight digital just by carrying a few holders.

[Mike Anderson]

This is almost an academic exercise:

http://www.imx.nl/photo/leica/camera/page164/page164.html

and it deals with small film vs small sensor, but it shows that certain film, with a certain technique, and a certain type of subject (b&w test pattern with fine lines and hard edges), and a certain output, analog can have better resolution than digital (same size sensor).

It's yet-another-film-vs-digital comparison, and a laboratory exercise, but still...

...Mike

[Jack Dahlgren]

Paul,

The ultimate in quality comes from stitching together multiple 20"x24" negatives.
Fewer shots are needed with that size. You could probably get by with 4 instead of the 12 8x10's that people usually are stitching together. But if you want to have smooth transitions with enough overlap you might want to bump it up to 6 or 8 per image. It is also important to use a slow film to capture all the detail with minimum grain.

[williamtheis]

of course, you can stitch Betterlight captures easily as well.... I want one of these but am doing it the hard way now by just pivoting the tripod head and overlapping

http://www.betterlight.com/panoWideView.html

[Ron Marshall]

Up to a certain print size, there will be very little difference in color prints from MF digital and LF.

"Little difference" is obviously subjective, and best judged by renting both a MF digital setup and a LF setup and having comparison prints made at the size you want.

Sure that is not cheap; but if you are seriously considering something in the ballpark of the P65 you are talking major bucks and it would be worth investing a relatively tiny, in comparison, amount of time and money to settle the question for yourself.

Of course, you may also have to get someone to prep your files in Photoshop etc., but it's worth the money to be sure.

[Jack Dahlgren]

of course, you can stitch Betterlight captures easily as well.... I want one of these but am doing it the hard way now by just pivoting the tripod head and overlapping

http://www.betterlight.com/panoWideView.html

That gives a lowly gigapixel or so. 6x20"x24" will give you closer to 16 gigapixels which will almost always suffice for larger prints.

[David de Gruyl]

Are there actually any printers that can handle an image like that? (serious question).

I get an angry phone call when my image files crash the printers at my lab. And that is only 30x40 @ 300dpi (apparently there is a problem with 16bit color, but wtf? put a filter on the damn thing).

[Ron Marshall]

Are there actually any printers that can handle an image like that? (serious question).



I think he is joking!

[David de Gruyl]

I think he is joking!

I figured as much, but apparently people do this sort of shenanigans with the betterlight motor pano thingy.

[erie patsellis]

I agree with erie patsellis: king of the hill is betterlight but it comes at a cost. Long exposure times and problems with artifacts if the subject moves (wind). I shoot 8x10 and scan with respectable scanner (Heidelberg Tango drum) but it does not approach the betterlight. Also there are no cords and batteries to film. there is a clarity and color purity not matched by focal plane sensor DSLRs. Having said that, I don't have access to a medium format back for comparison mainly due to expense. a used betterlight goes for $8K, far less than medium format but you have to add the camera and it must be a rigid one (I had to replace my wooden folder with an Arca Swiss). this also gives the opportunity to shoot both 4x5 and betterlight digital just by carrying a few holders.

I'm about as hardcore film as it gets, I have my own C41 processor and print all of my b&w work and about 85% of my C41 work myself. I bought a used Dicomed FieldPro and I'm having a blast, 20x24 at native size, more detail and sharpness than I ever thought digital was capable of (even my Phase One scan back doesn't compare). I have found I can upsample about 2.5 times before detail starts to suffer visibly, Making it roughly comparable to the interpolation a Bayer pattern sensor performs.

I've compared some of the MFDB (Sinar 54m, Hassy 39V, and my Megavision S3), the only advantage they bring to the table is the ability to use strobe. I was surprised that even the 54M in 4x multi pop didn't compare from an IQ standpoint. The only sticking point (well two if you include price) is the exposure times. A good comparision of Dslr, Betterlight and film is on the Betterlight website here (http://www.betterlight.com/rest_of_the_picture.html), orginally posted on the Luminous Landscape forum, with some comparision and notes by Mike Collette.

On the plus side, any reasonably good lens with an aperture will work quite well with a scanback, the resolution requirements are far less demanding with a larger sensor.

The MFDB has numerous drawbacks, Joseph Holmes does a good job of running down the issues here (http://www.josephholmes.com/news-medformatprecision.html)

All in all, if I'm going to go through the hassle, I might as well get the best possible image I can, and I typically shoot at least one sheet of film as well as a digital capture.

[Eric Leppanen]

Take a look at Jack Flesher's posts in this thread. Jack is a former member of this forum and shot a lot of 4x5 and 8x10 film before migrating to MFDB:

http://luminous-landscape.com/forum/index.php?showtopic=36661

[PaulSchneider]

Thank you very much for your comments, so basically if one really wants to go big, stitching and 8x10 should do the trick. But what interests me to know is, where is digital's finest today compared to analogue film?

I read an article on luminous landscape that basically said that the p45 is nowadays about equivalent to 4 x 5 film. So I guess the p65 is a little better than that.

Does anyone concur with this assessment? Or was Mr. Reichmann a bit over enthusiastic?

Would theoretical resolution would one need to achieve in a digital back to approach the resolutions extracteable out of 8 x 10? 200 Megapixels? 300?

Furthermore, for someone who has no idea with the costs involved in large format photography.

How much does a good (not the best) 8 x 10 setup cost? 10k? and how much is the total cost of one exposure? 30 USD?


Thank you very much for your comments!

[PaulSchneider]

Take a look at Jack Flesher's posts in this thread. Jack is a former member of this forum and shot a lot of 4x5 and 8x10 film before migrating to MFDB:

http://luminous-landscape.com/forum/index.php?showtopic=36661

Wow, great link! Hadn't seen thos thread! Thank you!

[Jeremy Moore]

of course, you can stitch Betterlight captures easily as well.... I want one of these but am doing it the hard way now by just pivoting the tripod head and overlapping

http://www.betterlight.com/panoWideView.html

They are really fun :D

[David de Gruyl]

8x10, used, less than 1500 dollars for camera, lens, filmholders, and tripod. You can certainly go much higher by getting a decent camera. The camera, however, matters very little. (by the way, this is real world pricing for me, with the most money sunk into the lens).

Scanners vary, and the best are very expensive.

For color film, figure 15 for e6 plus developing. B&W is much less.

[Eric Brody]

I'm not at all sure this is a discussion with a conclusion. Be aware that LL also said a 10mp Canon DSLR surpassed a Mamiya 6x7 (pure baloney that I'm sure MR regrets) in the "old" days, a few years ago. Very few people who buy prints look at them with a loupe.

If you're photographing for a client, the client will dictate the quality (or chose another photographer if yours is insufficient); if your working for yourself, the sky is the limit and your bank account determines how high you can fly. People who photograph for themselves also can determine what they enjoy doing. Some love the darkroom, some hate it; some actually enjoy the incredible control available in Photoshop. I recently had the privilege of watching John Sexton make a print in the wet darkroom. The man is a magician. If I could come close, I'd still be working in the darkroom. I cannot and actually enjoy the computer so am having a ball with Photoshop and digital files.

Good luck sorting this one out. It's way more than the math.

Eric

[Jim collum]

Pano adapter for the Betterlight is a *lot* of fun.

Here's a color IR image from Garapatta Beach, south of Carmel, Ca.

http://www.jcollum.com/fm/waves.jpg
original 6000x58000


http://www.jcollum.com/fm/angkor1.jpg
Cambodia
original 6000x14000

http://www.jcollum.com/fm/angkor2.jpg
Cambodia
original 6000x26000

[Chris Strobel]

I print with Epson wide format inkjets at Epsons native 360 res.It doesn't seem to matter if it was scanned 8x10, 4x5, or stitched 5DmkII (hell even stitched A640) As long as nothing is uprezzed or interpolated, all my color prints look the same, like Epson inkjet prints.A couple weeks ago down at Crysta Cove state beach, for giggles I snatched my wifes 6mp hundred dollar p&s and shot this rock.When I got home I printed it on the 4800, Harman FB AI, 12"x15", and the end result made me wonder why the hell I'm breaking my back lugging around my C-1 and A-100 to make inkjet prints.

http://farm5.static.flickr.com/4003/4509052789_9526516c74.jpg

[David de Gruyl]

As long as nothing is uprezzed or interpolated, all my color prints look the same, like Epson inkjet prints.

I think the point is made: so long as it has enough pixels to begin with, it is fine.

The trouble begins when someone says that a full frame 35mm dSLR has the same number of pixels as a 4x5. I can see that the P65 might have enough for many prints, but you can squeeze twice/eight times as much out of 4x5 (depending on the film / scan back).

[Chris Strobel]

I think the point is made: so long as it has enough pixels to begin with, it is fine.

The trouble begins when someone says that a full frame 35mm dSLR has the same number of pixels as a 4x5. I can see that the P65 might have enough for many prints, but you can squeeze twice/eight times as much out of 4x5 (depending on the film / scan back).

True, even with my 5DmkII I still have to either uprez the file or let the printer interpolate just to even get a 16x20.Of course its just so ridiculously easy to shoot several frames and stitch these days it seems pointless to spend for a P65+ unless your rich or can justify it in a professional workflow.

[Daniel Stone]

PaulSchneider,


as a working(starting) assistant here in LA, most working pros are digital-only now. But if you're looking at shooting for art, or for personal work, you most likely won't have a deadline to meet time-wise. so IIWY, I'd still look at film, especially LF, as a viable source of base capture.

I've seen plenty of p45+, p65+, leaf digital 56mp files, hell, even 10 shots stitched from p65+ files, to make 1 shot after plenty of PS work.

personally, I still think that film(8x10 in this case) is still the best way to go if you want the best quality if you're planning on printing big(over 30x40 really). especially if you're dealing with a single-shot situation(like with people, portraits, etc....)

but in the end, the technology for both 4x5, 8x10 or digital MF, you'll be looking to spend a nice bit of cash to get the best image quality out of the film or digital sensor.

this will involve most likely a GEARED camera, a heavy/super stable tripod, and the best glass(most modern to maximize the film's potential). this combined with a great drum scan will insure you terrific quality.

-Dan

[rdenney]

True, even with my 5DmkII I still have to either uprez the file or let the printer interpolate just to even get a 16x20.Of course its just so ridiculously easy to shoot several frames and stitch these days it seems pointless to spend for a P65+ unless your rich or can justify it in a professional workflow.

Hmmm. I've stitched images on a number of occasions. In the most recent attempt, I scanned both ends of a 6x12 transparency in a Nikon scanner. It took me an hour to align those images, and the Photoshop tool for doing so could not deal with the image size on my computer (which is admittedly a bit long in the tooth, though it still has 2 GB of RAM and no shortage of space for a scratch disk).

http://www.rickdenney.com/images/Niagra_cannon_scan19-20_lr.jpg

At sample spacings of about 6 microns, the negative holder for the Nikon is nowhere near precise enough to even make sure both images are aligned in rotation, making the stitching process more than just lateral translation. And if I'm not going to align it to the pixel level, what's the point of scanning at high resolution? Photomerge had errors noticeable even in a full-screen image, and I had to downsample the scan files to 2400 just to get the software to run. Based on how difficult it is to scan two ends of the same transparency, where the images are truly identical on the stitch line, I'm thinking that stitching images that will vary in details will certainly not be any easier. Maybe there's a software tool that makes it easier.

And as much as I am a card-carrying member of the Pretty Rocks school, things change from one image to the next in a stitch. Clouds move, trees sway, even the Sun moves. The lens performs variably across the image--there is vignetting, varying patterns of out-of-focus rendering (or even in-focus rendering), and maybe even distortion. Making all the repairs necessary to make those stitch lines look natural is a lot of work, and I admire those who get good results doing it. Given the importance of timing in much of photography, or the importance of how a lens renders a scene variably across the frame, I can hardly defend stitching as THE answer.

I once used Panoview to stitch a dozen or so images of the ridge line visible from the Mount Ranier visitor center. It worked, but I had to engage in quite a bit of "computer art" to blend the clouds. (Not linked here because it wasn't made with a large-format camera). That one took a couple of hours, too.

On the Betterlight scanning back: For some reason, images made using that device seem so obviously made with it that even screen displays of them draw the immediate response, "that was made using a Betterlight scanning back". There is something about the way it renders color that is characteristic. The images are still pretty amazing, but if things move in the image, the results are nearly surrealistic. The waves in the example above illustrate this--I've never seen waves that look like that. That in no way diminishes how striking that panorama is, of course. But it's a characteristic look that is different than one would get with a frame-wide exposure, and that difference may or may not align with one's artistic objectives. So, a scanning back also seems not to be THE answer.

The problem with the medium-format backs is their size. I love wide-angle lenses, but finding a lens with a focal length less than half the image diagonal when using those reduced-size digital backs is no mean feat. Finding one at a third of the image diameter is impossible without custom fabrication. That has been an issue with small cameras, too. The typical medium-format sensor might be 33x44, which has a diameter of 55mm. The only lens I can think of with sufficient coverage for that sensor that is less than half the diameter is the Canon 24mm TSE lens, which would be a challenge to mount and use with a sensor like that. Canon users stitch with that lens, but then we are back to that problem again. (I once stitched two images made by shifting a Canon 24mm TSE lens, and even that was a challenge when working at the pixel level--Photoshop couldn't do that one accurately, either, though the distortion of the earlier Canon lens didn't help.)

The shortest large-format lens I can currently afford is a 47mm Super Angulon, and that's adequately wide on the 6x9 format, though not as wide as I would like in some cases. I'll be all ears when there is an affordable 6x9 non-scanning digital back. But if they can make that affordable, could 4x5 be that far behind? I suspect the market would be pretty limited, making it a real challenge to justify the investment required for such a sensor.

The 35mm Grandagon digital, or whatever (I don't keep up with these unobtanium lenses), might be sufficiently wide for the largest current digital backs, but we are just getting to the point where it's even an option, and it still limits us to a focal length of half the image diameter.

So, if we define quality in terms of what sorts of photographs we can make in addition to image performance, then I would say that medium-format digital backs still do not accomplish my goals as a photographer, even without considering the expense. I have nothing against digital photography--when the technology is there and I can afford it, I'll go that way. And I'm looking forward to it. But it's not there yet.

Rick "thinking one has to define one's photographic goals before answering this question" Denney

[Struan Gray]

I use LF because it's cheaper.

But, one reason I don't get too hung up about it is the graceful way film handles extremes of lighting. The page on the Betterlight site that Erie linked to illustrates it nicely (http://www.betterlight.com/rest_of_the_picture.html). I love the clean look of digital colour reproduction, but hate the video-y blocking, artifacts and fringes that crop up when the contrast gets too much.

Jim C.: it would be interesting to see two shots of Garapata beach panned in opposite directions. I'm guessing your photo had the camera panning against the wave motion.

[Oren Grad]

The problem with the medium-format backs is their size. I love wide-angle lenses, but finding a lens with a focal length less than half the image diagonal when using those reduced-size digital backs is no mean feat. Finding one at a third of the image diameter is impossible without custom fabrication.

Rick, it's closer than you think technologically, if not financially. The backs typically preferred for landscape work have 36x48mm sensors, with a 60mm diagonal. And there's now a 23mm Rodenstock HR-Digaron S which will cover. Don't bother looking up the price, though. :)

I've dabbled just a bit in medium format digital capture. Apart from irritations like moire and the behavior with high SBR as Struan mentions, even the best files I've worked with still have an "electric picture" look when printed, which is distinctly different from the way film renders. "Image quality" can be fighting words, so maybe it's safer just to say the image *character* is different. Either you like it or you don't. I still mostly don't.

[ic-racer]

MFDB vs Large Format - where are we today?

Setting up a large format darkroom has never been so affordable. I don't even know what MFDB is but I'll bet it cost a lot today and will not be worth anything tomorrow.:D

[erie patsellis]

I'm pretty sure it's not an either/or situation for most of us.

[Ivan J. Eberle]

Horses for courses.

Jim Collum's pano Garrapata Beach is interesting to me because it demonstrates why pan-stitch doesn't work for many situations:

Waves simply do not break like this. (No, not even on Garrapata Beach-- this I know from it being just minutes away, being one of my most photographed areas). While it might be a lot of fun for the maker, large print images like this have a half-life of about 10 seconds. If forced to stare at that as print of for any longer, I'd get a headache. (To me, it's reminiscent of the "When Pigs Fly" syndrome of early PS campaigns from a couple of decades back.)

That said, and in fairness to Jim for graciously sharing these shots with us, the stitching technique works really well for the Angkor Wat shots, and I like them very much.

Multiple slices of time don't capture with any technical accuracy anything with motion, and that includes many, many other (if not the vast majority) of scenes of the natural world around us. So what good is the added resolution? Unfortunately, LF film doesn't arrest motion very well in any kind of aesthetically-pleasing natural light, either-- not with the color emulsions we've got available today, anyhow.

Medium format film still fills a niche for me for the larger apertures and higher shutter speeds; ultimate print size doesn't wow me if the resulting image fails to capture the mood of a scene due to other technical limitations. Hi-end MF digital still seems rather absurd for landscape unless/until one's print business is sufficiently monetized that the depreciation becomes a writedown (that actually helps offset the print income on the tax return), and the inefficiency of a scanning workflow becomes onerous-- or all the additional stitching with 35mm FF is just too much bother.

(The ~$6500 Pentax 645D looks to be a step in the right direction, but will it be worth 3X what a Sony a850 can do?)

What will really be the game-changer, the one that will stop all these arguments dead in their tracks, will be a considerably larger-than-cropped-645 sensor with relatively large photosites, ultra-high resolution, high ISO capability and 16 bit or greater capture per channel.

Better yet, give me all that on a sheet or roll of a digital film with the same form factor as traditional film so I can use it in my traditional cameras. Polishing my crystal ball, I see this happening within the decade-- for under $100 :-D

[Jeremy Moore]

Horses for courses.

Jim Collum's pano Garrapata Beach is interesting to me because it demonstrates why pan-stitch doesn't work for many situations:

Waves simply do not break like this. (No, not even on Garrapata Beach-- this I know from it being just minutes away, being one of my most photographed areas). While it might be a lot of fun for the maker, large print images like this have a half-life of about 10 seconds.

Not everyone agrees as I really like them, but as you said "Horses for courses". Part of the joy in photography, for me, is discovering the limitations of how my eye and mind perceive the world. Jim's photograph does a great job of showing how waves break when panned across. I know how they break in a straight photograph, I've seen it and photographed it myself. Jim is showing me something I can't see--this is exciting.


That said, and in fairness to Jim for graciously sharing these shots with us, the stitching technique works really well for the Angkor Wat shots, and I like them very much.

Jim's shots aren't stitched. They are Betterlight panos. The Betterlight has the option of working with a mechanically rotating panoramic head--think digital Cirkut camera.

[Chris Strobel]

Hmmm. I've stitched images on a number of occasions. In the most recent attempt, I scanned both ends of a 6x12 transparency in a Nikon scanner. It took me an hour to align those images, and the Photoshop tool for doing so could not deal with the image size on my computer (which is admittedly a bit long in the tooth, though it still has 2 GB of RAM and no shortage of space for a scratch disk).

Nice image Rick, but your doing it the hard way.Just get yourself a little Canon Rebel, a used Nodal Ninja 3 head, and Autopano Giga.I guarantee things will go much smoother :)

[Andre Noble]

There is WAY too much emphasis on pixels count, and not enough emphasis on improving dynamic range performance.

[Andre Noble]

Nor is there enough emphasis on print substrate.

There is currently an exhibit at A&I labs in Hollywood of a series of 4x5 tranparencies of colorful flowers. The 4x5 sheets were drum scanned and the resulting digital file printed out onto huge sheets of Kodak Metallic paper. With all due respect, Christopher Burketts Ilfochromes, these are not. http://www.christopherburkett.com/home.html

[rdenney]

Nice image Rick, but your doing it the hard way.Just get yourself a little Canon Rebel, a used Nodal Ninja 3 head, and Autopano Giga.I guarantee things will go much smoother :)

I was trained both in architecture and engineering. Your approach appeals only to the engineer side.

Rick "demanding elegant solutions in addition to merely fulfilling requirements" Denney

[erie patsellis]

There is WAY too much emphasis on pixels count, and not enough emphasis on improving dynamic range performance.

Any competent digital back made in the last several years already far exceeds the dynamic range of chromes by at least a few stops. For offset reproduction anything over 5 stops is troublesome, and reproducing 5 stops requires a great deal of care in the prepress side. The norm in the chrome days was 4 stops max. While it's nice to have toys, these products are made for working pros, whose images are being reproduced in offset printing. Any other use (fine art repro, landscape, etc.) is a bonus, but not the intended use, with the exception of Betterlight products.

erie

[mrladewig]

But we are not limited to shooting chromes nor is our output choice limited to offset printing. Color negative is not that hard to scan and it still easily bests the available dynamic range of digital without the weird looking HDR artifacts. When I choose a film for a particular shot, I consider the character of the film and select the best choice. Sometimes that choice is Velvia, other times its Portra.

I think making an assumption in today's market that any shot made on digital is intended for offset printing is folly. As magazines shut down and content delivery moves to the web, its just as likely the intended output will be an uncalibrated LCD as a magazine page. More likely, it will be intended for an ad campaign on both.

[David Luttmann]

There is WAY too much emphasis on pixels count, and not enough emphasis on improving dynamic range performance.

Depends on what you shoot with. If it's transparencies, then any DSLR or MFDB already has FAR more dynamic range. If it's color neg, then yes, the film can in some cases capture more.

But I agree, far too much is spent on pixels. The latest MFDB achieve resolution and dynamic range that even at 24x30 and 32x40 matches the quality of 4x5. Now as to the cost though, well, that's another issue :D

[Eric Leppanen]

As magazines shut down and content delivery moves to the web, its just as likely the intended output will be an uncalibrated LCD as a magazine page.Web ad delivery is already exceeding traditional print media:

http://www.forbes.com/2010/03/07/advertising-web-ads-digital-business-media-outsell.html

[Thebes]

On dynamic range, whenever I read a digital vs film comparison someone throws in the lower dynamic range of most transparency film, yet many never shoot it. I consider chromes to be largely a hold over of last decade's art directors and a general inability to quickly read a color neg. The color stuff I shot a bit for a newspaper back just before digital was scanned color neg and this seemed to be the norm... greater dynamic range and little need for ultra saturation on newsprint.

For B&W negatives, you can pull out "HDR like" dynamic range with a single shot (please note hdr does not necessarily imply tonemapping) I have recently shot with ice in the sun and a figure in a shadowed cave in the background, there is still detail in her dark hair and other shadow areas as well as the ice's highlights... no can do with a single digital shot. Such negs can be easily printed straight on azo or lodima or pt/pd, or printed with a bit of effort using split contrast techniques in an enlargement... or even scanned.

With output, I don't think its just the substrate. I tell something is "wrong" with a giclee, in terms of smoothness, at about a couple feet away, at about a foot I take off my glasses and can see the dithering patterns with my unaided eye. Don't get me wrong, I prefer digital output for color work, but with a nice B&W contact print they have a certain look, and if you loupe one vs a digital print you will see that this look is largely due to a healthy excess of detail which your eye can not "see" but can "feel". My 8x10 enlargements from 645 show more detail than I have seen in any "carbon pigment print", and all the digital c prints I've seen seem just a touch soft, as if the enlarger were a bit unfocused. I suspect this is the reason giclees are so often made on water color paper, and digital prints are made large far more often than enlargements of the past.


Choose the right tool for the job. For a large studio shooting for magazine ads, a digital back and a backup and all the nicest lenses optimized for the sensor makes complete sense. For someone shooting B&W landscapes film probably makes more sense. Color field work varies but it seems the majority of high volume pros are moving to digital.

BTW, a large studio doesn't shoot for ad banners on the web, the blogger, or webmaster, or whoever just grabs their p&s and takes a quick snapshot.... that is increasingly how it is, and the general quality of photography has suffered tremendously for it.

[Jeremy Moore]

For B&W negatives, you can pull out "HDR like" dynamic range with a single shot (please note hdr does not necessarily imply tonemapping) I have recently shot with ice in the sun and a figure in a shadowed cave in the background, there is still detail in her dark hair and other shadow areas as well as the ice's highlights... no can do with a single digital shot. Such negs can be easily printed straight on azo or lodima or pt/pd, or printed with a bit of effort using split contrast techniques in an enlargement... or even scanned.

Agreed.

[David Luttmann]

Agreed.

Ditto

[Jim collum]

Jim C.: it would be interesting to see two shots of Garapata beach panned in opposite directions. I'm guessing your photo had the camera panning against the wave motion.

The waves were coming straight towards the camera... as the camera panned from left to right, the leading edge of the wave got closer.. making them look as though they were breaking to the right. A reverse direction would just make them look as though they were breaking in the opposite direction.

[Jim collum]

Horses for courses.

Jim Collum's pano Garrapata Beach is interesting to me because it demonstrates why pan-stitch doesn't work for many situations:


You're right about the waves... what's being shown is an artifact of elapsed time. You'll get artifacts with film as well.. those depending on how long you hold the shutter open. The cotton like water and smooth glass-kenna-style shots.. aren't real.... just the way that film captures the waves over time. We're all used to seeing those artifacts, and what the scanning backs capture doesn't jive with what we're used to. I've grown to like the effect (which doesn't work in all cases). Personally, I still prefer the more traditional rendition of moving water over time that a single capture with shutter open gets.

[Jim collum]

Ditto

and ditto again :)

[Struan Gray]

The waves were coming straight towards the camera...

I hadn't considered that - I was assuming the panning speed was closer to the wave speed. In retrospect it makes more sense than the hideous cross rip you would need to make the waves break sideways like that.

In general, I am fascinated by different ways to capture time and motion in a still image. Cartoonists and artists have more tools than photographers, but slit and scanning cameras are fun. That the waves look 'odd' is the entire point. Hokusai wasn't exactly literal either.

[rdenney]

Personally, I still prefer the more traditional rendition of moving water over time that a single capture with shutter open gets.

In this image, the scanning effect is at least intriguing. But the point is that they produce different effects, and one will not always do for the other.

As I said In another thread, how we render subject motion is a gross effect, and freezing is also often an abstraction. All approaches can support an artistic visualization. Or not.

Rick "noting famous examples that demonstrate the focal plane shutter effect" Denney

[gnuyork]

This topic is opening up a can of worms for a heated debate. But as I mentioned in another thread, I saw an exhibit of a landscape photographer that had most of his images captured with 8x10 velvia or Astia and scanned with a color getter drum scanner and a few images of them were captured with an expensive high end digital back (I can't remember the specific model). He had notes below the prints as to which camera/lens/film or DB was used. All were printed on fugi chrystal archive.

I could call out the digitally captured images with success before reading the notes for conformation. I saw the difference, and for the purpose of fine art I personally prefer the film captured images.

I was also recently at an exhibit at Jackson Fine Art in Atlanta of the works of Andrew Moore. He had large (40 x50) C prints made from 8x10 color negatives. Again the look of his prints were incredibly detailed, and his work was quite inspirational.

I keep hearing that the Ilfochrome process is the ultimate way to go for slide chromes. I have used this process in college with great results, but I only used 35mmm at the time. This is encouraging because somewhere in the next few weeks, possibly even this coming weekend, I am being given a machine to make Ilfochrome prints, though shooting 4x5 chromes is completely new to me, and will be a learning curve I am sure.

[D. Bryant]

I print with Epson wide format inkjets at Epsons native 360 res.It doesn't seem to matter if it was scanned 8x10, 4x5, or stitched 5DmkII (hell even stitched A640) As long as nothing is uprezzed or interpolated, all my color prints look the same, like Epson inkjet prints.A couple weeks ago down at Crysta Cove state beach, for giggles I snatched my wifes 6mp hundred dollar p&s and shot this rock.When I got home I printed it on the 4800, Harman FB AI, 12"x15", and the end result made me wonder why the hell I'm breaking my back lugging around my C-1 and A-100 to make inkjet prints.



Try having your film drum scanned. You will definitely see a difference. If you don't something is wrong.

Don Bryant

[Brian Ellis]

I'm not into making the biggest prints with the most pixels or their film equivalent so I don't pay much attention to the various devices mentioned in this thread for doing those kinds of images. They're out of my price range anyhow. But since people have been talking about panoramas, stitching, gigapixels, etc. I thought someone should mention the Gigapan device. http://www.gigapansystems.com/.

I recently saw a copy of a panorama made with a Canon 1DS Mark III camera using this device and making three exposures for each "slice" in the panorama. I forget the total number of exposures but it was in the range of a couple hundred I believe. Each slice was then adjusted to taste using Photomatix or something similar, the results were stitched in Photoshop or the Gigapan software, and the portions of the resulting print that I saw, which was measured in some multiple of feet rather than inches, looked pretty darn good.

I can't compare quality of images made this way with film or any of the other devices mentioned here, I have no basis for comparison. I just thought it was an interesting device and if gigapixels are your thing, this seems like a relatively easy and inexpensive way to get them.