I wouldn't necessarily doubt what the sales guy said with respect to these higher-end backs being somewhere between 5x7 and 8x10 sheet film quality. Remember, a flatbed scanner will not be as good as a drum scanner in resolving capability. One may indeed have to scan better than 5x7 sheet film, on a Kodak, to achieve results as good or better than a 39MP back. A drum scanner's capability merely allows one to use smaller film when making a comparison.
The image one sees on a screen is only part of the story. We still need to print the picture.
One caveat with digital systems that many pros miss is the effect of pixel size. This effects the Canon 1DsIII as well as the newer flagship digital backs, but not the 4x5 film at the same print size.
On paper, with roughly the same size pixels, the 1DsIII and the ~35 MP backs have the same diffraction degradation beyond f8. Optimum f stop should be about f 5.6 to get full resolution without diffraction*). Just try at f32 and see the result. here . Here the white orchids have green edges due to seemingly through this phenomenon. (This does need to be tested at different apertures.)
I really don't think that the 1DIII can match a DB for dimensionality, clarity and presence of the delivered image. I believe that if one is not going to use film, a second-hand Phase one or other MF back would be far better, especially since the sensel pitch is way larger and therefore one can close down the aperture more to get increased depth of field. Unless one really needs more pixels, bigger pixels are better! The new backs such as the P45 have pixel pitches of 6.8 microns whereas the P25 boasts pixels of 9 microns.
The 1DsIII should be best reserved for fast action shooting where lots of long lens work is required and where one cannot afford the MF options of a camera under warranty.
Here however, we are more interested in low volume Landscape, Architecture, still life and portrait where 4x5 film still reigns supreme for at least the next few years. Charles Cramer estimated that his 4x5 film costs over 3 years was 65% of a Phase One P45 back!
If one wants a large print, and one is considering a DB such as the P45 or the new wonderful Sinar Hy6, here , then presumably one has the workflow to justify the ~ $44k expense of the system. However, people who use now use LF may, for the most part, never shoot enough to justify leaving film. A properly scanned 4x5 Velvia sheet of film will still barely edge out the best backs in resolution and color can always be profiled to your taste. Source.
So, for many people, LF is still a much more cost effective proposition for making large impressive prints. If one wants speed and less weight, A used MF digital back and your old Rollei, Contax, Hasselblad, Bronica SQ, Mamiya RZ or a Mamiya 645 AFD will do nearly as well as the flagship digital backs with ~30-35 MP. That will be a package that many more of us could afford around $5000 to $9,000.
My needs are far more limited so for my huge prints, 8x10 film will serve me well for some time and I don't have to worry about resolution!
Asher
*The light waves "scraped" by the small apertures convert detail (points of potentially observable detail at normal viewing distances) of the print from being smeared by ripples of spillage to adjacent pixels.
I have used a 1DS MKII for years. It is superb. I put the MKIII next to it in the studio and even the person trying to sell me the camera had to admit there was not very much in it. I also had a 22 mp Leaf. It is harder to compare this as you cannot use the same lens as the canon on comparisons but I am certain the leaf was loads better. I now use a 33mp Leaf. It is in a different league to the DSLR's. I tested the 39mp phase one and thought the Leaf better.
The economics of it are only part of the story. Turnaround time is a factor and other stuff as well when shooting commercial work. I use the Leaf on the new AFI for medium format. I have never seen sharper lenses than the shneiders I am using on that thing.
My love though is the new Cambo SC23 that also takes the Leaf back. The same back but I think of it as LF when using it. The same calm methodical approach and best of all movements. I use the schneder apo digitar lenses. They are great but not as sharp as the MF lenses. The movements more than make up for it though. This makes it more effective economically. It would be hard to justify the back if it was only used on the Cambo
The point is one back but it is at the heart of a MF and LF system for me.
I played around with some of the original scans and digital files from that test. When Cramer processed the Raw images in Raw Developer instead of Capture One, he found the detail to be even better than originally thought. At this point, the P45 is every bit as good as drum scanned film on 16x20 and 24x30 prints. At 40" & 50" it appeared the film may hold a slight edge. How important that is is up to the user. Cramer found that he was perfectly happy with 40" prints from the digital file and found the detail to be visibly similar.Originally Posted by Kirk Gittings
The digital file holds acutance better than a film scan can, even with more pixels from the film scan. As well, the dynamic range of the digital file at slightly over 12 stops far exceeds what any scan can obtain from any transparency film.
Now as to the cost of the P45........
And this, IMHO, is a problem. The dot pitch of your monitor is way bigger than the pixel pitch of a well made print. A monitor can show you as much as 5x enlargement over a print. Then, if you are looking at anything less than 100% pixels it distorts smoothness and detail because you are at the mercy of an algorithm that is determining which pixels to show you and which to toss out. Also the monitor is a light source so it typically shows shadow detail that a print (reflective source) often doesn't have. I could go on but I think you get the idea.
Unless your final output is going to be a monitor screen, I think you'll learn a lot more making and comparing prints.
Bruce Watson
That’s exactly what I did as well. I found that in real terms, screen to print comparisons were closer to between 35% & 40% of what we see on screen….but even that is not clear. I ended up making some 8x10 prints from 20” & 40” crops of the scans vs P45 to come to my conclusions. For me, up to 30” the digital file maintained sharpness better. At 40” +, the film won….but it is very close. From a 3’ viewing distance, there was no difference.
But still, I can afford to work with 4x5…..I can’t afford the P45.
Really the only thing that matters is the economics. If you have the volume of work to justify spending $40K plus on a high end digital, then it makes sense and you're not going to split hairs over it versus drum scanned film because the workflow advantages of digital are so much better for commercial photography. And the commercial guys I know who have that kind of work also have TWO set-ups since they need a back-up.
Otherwise you're going to be using film. The high end digital backs aren't going to scale down in price ~ ever ~ because they don't need to, there's never going to be enough demand... right now a $2000 dslr provides amazing quality for most photographers, and it's only the fine-art hobby togs and the higher end commercial shooters who care about anything better.
I haven't had to make a print, other than trade show displays, for a commercial client in any recent memory, so I really don't care how it prints on an inkjet. I just want the client to like the files.
I really doubt Canon is ever going to need to increase the resolution over the 1Ds3 because so few people will need it -- instead they'll make those bodies lighter and cheaper.
Asher, Bruce and Frank all bring up really good points with this. Firstly that comparisons on any computer monitor, even the top of the line EIZO, just don't show you a reality for prints. One could just as easily take a pro level HD video camera and wow many people with on-screen images, yet those are far lower on the megapixel scale of things.
People who claimed a savings from getting a digital back likely had very poor business models, and quite likely still do. If you are shooting commercially and either not charging for your film expenses, or not incorporating those expenses into your overall billing to each client, then your business model is sadly flawed . . . or quite simply you are not charging enough. In a similar manner, if you are using a digital back for commercial work, then you should be charging appropriately, not dropping your fees, and factoring in replacement cost within 18 to 30 months. Many of the digital back makers have lease plans, often with upgrades possible, that can make more economic sense; so another thing to look into. Only needing a digital back on a more limited basis could be accomplished through renting.
Technical issues are somewhat bounded by manufacturing, colour filtration (Bayer), interpolation issues, and internal electronic noise. The current optimum of large CCD imaging chips is in the 6 µm to 8 µm (microns) range. To go larger means greater heat generation, requiring more aggressive noise correction. To go smaller means more noise and greater chance of interpolation issues, though Sony and Nikon seem to be working around this a bit more lately (Kodak too, though more to come on this in the near future) in current offerings with under 6 µm pixel sizes. So the current digital backs with larger pixel cell sites gain an advantage in light gathering, as long as they fall within the sweat spot of charge to noise. The other advantage is a bit less apparent, namely the Bayer filtration is easier to quality control at larger physical pixel sizes; the cleaner and more accurate makes Bayer interpolation better, or makes correction easier.
Part of the cost of medium format digital backs is the quality control. All the current companies do more quality control of their 22MP chips than Canon does on their 22MP chips, which is reflected in the cost differences. A great example was an article on PhaseOne in which a worker mapped and verified every single pixel in each back; they are not unique in this, Leaf, Sinar, and Imacon/Hasselblad perform very similar quality control. I think you can often see this in final prints with the colour control and fidelity possible from most MFDBs, though arguably the difference from these to a 1DsMarkIII is subtle.
I really don't get the dynamic range argument used as an advantage. When I look through Lürzer's Archive 200 Best Ad Photographers, or even CommArts, I rarely see higher dynamic range in what those magazines consider the best images. To take this to an extreme, using HDR software would seem like the ultimate, yet I have yet to see any images that didn't look gimmicky . . . or just plain sucked. So a digital chip has more dynamic range than colour transparency . . . big fluffy dog (BFD)!
Another item is resolution. Make a really large print, and people step back to see it all. Make a smaller print, and people get closer. After I saw an Edward Burtynsky exhibit at MOPA, and his 40" by 50" (roughly) prints from 4x5 film, I realized that even 4x5 is much more than enough for most printing needs, and should never fail to leave an impression.
A few white papers several years ago showed the development direction of future chips. In one of the Kodak white papers it was suggested that a practical limit is a roughly 645 sized chip at about 100MP. If you scale that down, you are near 40 MP on 35mm sizes. Scaling up is tougher, since chip rejection rates are higher on larger chips. This brings up a different issue: composing an image while using the camera. Using a good prism finder on a 645 sized camera is fairly comfortable, and gives one a brighter or larger image than with a 35mm sized D-SLR. Composing on a 4x5 ground glass is easier for some than using a smaller ground glass area; and I think this is especially true for using movements. Remember, we still need to make use of these tools to get images, and as long as the images do not contain flaws, the end viewer will often have no idea what gear was used.
Ciao!
Gordon Moat Photography
The dog is pretty big and fluffy when in order to max shadow details you must blow out highlights when using a transparency film....even a low contrast one like Astia. Thus, having more DR will allow you capture it all in a single shot as opposed to multiple shots.
As to HDR, they can be made to look gimmicky or not. An HDR image used to maintain details in shadows and highlights, and used properly, will not look unnatural. If you're only seeing unnatural shots.....you're looking at the wrong shots!
I think you missed my point David, using a higher dynamic range does not imply a better image. Quite to the contrary, the examples I see out in the real world of commercial imaging show a more compressed dynamic range. I have no need to maximize shadow details, and I have yet to see a compelling colour image that was more compelling due to greater shadow detail. Quite simply, I see no need for a greater dynamic range when the top level shooters just are not doing things that way.
Perhaps art landscape imaging is different? I have to wonder if photographers are pushing for more of this technology, rather than the opposite of it being something end purchasers of images really want. I know people who rave about HDR, but they all seem to be other photographers into this stuff, not advertising people, not art buyers, and rarely any gallery owners nor patrons of the arts.
So what is the point of greater dynamic range when it doesn't matter whether or not one uses it? Perhaps that gives greater options later to compress the dynamic range of the final print, though anyone who handles exposure accurately enough can control that, or simply bracket and choose the most appealing end result.
There is more than enough dynamic range in Fuji Astia 100F to make compelling images. This is quite easily possible without blowing highlights. Why would I want anything more? I could just as easily shoot C-41 films instead of E-6, yet using E-6 films meets my needs, and more importantly meets my clients expectations.
Ciao!
Gordon Moat Photography
Posting Permissions
Reply With Quote
Bookmarks