Architecture: LF film vs. 35mm digital capture

[Rainer]

Which lenses do you use on your DSLRs to avoid ballooning distortion when shooting architecture?

I tried using my D100 with the 18-70DX but distortion is too strong with this lens.

You think any non-zoom lens would do better?

[Kirk Gittings]

" another issue with digital, in my opinion, is where is the truth?"

In architectural photography, at least as it applies to design competitions, digital manpulation is already a huge issue for obvious reasons. In all forms of "documentation" photography there is an ethical issue. In "art" photography (whatever that really is), well, art is art. In my opinion when photography strays too far from the actual world it loses its unique power as a distinct visual medium and merges with all other art forms.

[Neal Shields]

“a Better Light digital back will give better detail than even your 4x5”

I had a very long debate about that, via Email, with Mike Collette at Betterlight over about a month’s period of time.

We exchanged detailed sections of images: his from his back and mine from 12,000 dpi drum scans of Velvia.

He may have won.

However, one thing we noticed is that film will try to reproduce very fine detail at actual size. Pixels will take detail that is smaller than the pixel and magnify it at a sacrifice to contrast. A black power line, imaged at the sensor as 3 microns in diameter will be reproduced by a 9 micron sensor as a dark gray power line 9 microns in diameter with perfect sharp edges.

So if we stood side by side and photographed power lines several miles away you might see his power lines and not mine.

I have said that you will pry my film from my cold dead fingers, but if I was making a living at photography, I would own a Betterlight back. Especially if I needed perspective control.

P.S. If you haven't seen his stitched Golden Gate Panorama, it is worth a trip to his site.

http://betterlight.com/panoWideView.asp

I tend to see most digital as an assault on photographic quality, but you certainly can't say that about Mike and Betterlight.

[Emmanuel BIGLER]

A black power line, imaged at the sensor as 3 microns
Neal. Such a fine line @3microns width is extremely difficult to transfer with any lens stopped down to f/16 and beyond.
If the lens was perfectly diffraction-limited @f/16 the width of this line would be something between 10 and 15 microns (1.2 N lambda, with N=16 and average lambda ~0.6 microns), not 3 microns. I would not blame the sensor for broadening a 3 micron black line to a 10 micron grey line. Diffraction and aberrations take their part of the job before the sensor can do anything worse ;-);-)
However, checking how lines of 20-30 micron width are recorded on film or on a 9-micron silicon detector is an interesting experiment for which you certainly have the answer.

[Kirk Gittings]

I'm sure they have improved but a couple of years ago I demoed a Betterlite at Calumet in Chicago on a breesy day. Unfortunately the slow multiple passes of a scanner back did not do well with moving trees etc., which is often how I have to shoot architecture and simple to deal with on film.

I'm not trying to be hardheaded here (though I am told I am) but I'm talking here about real worlds architectural photography. I've made my living doing this since 1978. I am open to new ideas. My methods have changed dramatically over the years, but digital capture has to give something better or equal to what I have now and at a reasonable cost.

[Henry Ambrose]

Kirk, you're not being hardheaded, just realistic.

I am not an anti-digital Luddite. I've used digital capture since 1992 for studio product photography. Its absolutely wonderful for that application. The quality of the files produced by scanning cameras is amazing. But they are still not a practical architecture tool for everyday shooting. If anything moves the photo is ruined. (or you have hours of PS correction work) When you're outdoors something is most likely going to move.

The DSLRs, even the very best, don't seem to play well with wide or PC lenses. I see countless discussion about how to use Leica or Contax wides on somebody's new 1DsMII. The lenses that are available are not that good. The sensors don't work well with wides anyway. Same situation for MF backs as far as I can tell - and there is an even smaller selection of appropriate lenses.

I could see using a 22MP digital back on a MF camera with movements with wide lenses that worked. But they don't exist. (the Alpa page mentioned looks good and might work!) The angle of acceptance problem is still there. And then there's the cost for a camera back that you sure can't afford to keep a back up for or fix in the field with gaffer tape.

For lower resolution needs like small reproductions in print (if thats all they need or want) you can do it fast and good with any of several different DSLRs. But when the needs call for large prints or the very highest quality then a big film camera wins the race.

In commercial photography digital makes the most sense when you are involved with high volume shooting and or scanning - you can save (or make) lots of money replacing film with digital for those kind of operations. Digital can also be faster to deliver than film.

But, when you're photographing a building how much film can you use? How many scans do you have to make? Do they really need it tomorrow?
For me the answers lead me directly to film. A dozen (or maybe shooting extravagantly 24) sheets of film will easily me all the give coverage I need for a 3-4 view commission of a building. I can deliver my prints and scans on CD in a few days after I shoot the job. My costs for film and processing (less than $100) are a pretty low percentage when compared to my reasonable fees for such a job.

Why would I buy a $20,000 digital back or an equivalent load of DSLR gear for photographing architecture?

[simon warren]

Hi Kirk

I dont think that you are missing anything. Too many people have been too quick to take up digital. Film will allways be king in my book. I shoot architecture. Most of my pics are used very big. If there is a digital system out there that can match the quality I would like to know about it.!

Cheers Simon Warren

[David Luttmann]

Hey Simon,

Try a Better Light digital back. No Bayer processing, and resolution that will leave 4x5 in the dust. Even the latest 16 multi exposure 22MP backs exceed what is capable with 4x5. You get the equivalent of an 88MP image with no grain and color bang on. The results I've seen from the Better Light can pretty much match 8x10 film....which is something I never thought I'd say.

[Kirk Gittings]

I did look at these a couple of years ago. There are many questions to be answered? How big is the capture surface. How wide a lens do you need to use then to equal a 90 and 75 on 4x5? Do the sensors suffer from noise because of the angle of light at the corners? Do you need to be teathered? Is this a 25,ooo.oo investment that will be obsolete in 3 years?

Henry nailed the issue above. In the real world of arch. photography for those who make their living doing this, what makes sense now?

[David Luttmann]

Kirk,

Why would it be obsolete in 3 years? If the quality is as good or better than 4x5 or 8x10 film capture, then the photos you take with it in 3 years will be just as good as now. I always find it funny that the second digital capture is mentioned, people talk about it becoming obsolete. Just because Fuji comes out with Velvia 100 doesn't mean my photos now with all the Velvia 50 in my freezer are obsolete. This is especially amusing in a large format forum. Many people consider our large format cameras to be obsolete....but you and I keep using them.

As to the other issues at hand, noise and CA issues have proven minimal on the Better Light tests I've done. Sample 22MP photos I've worked with come extremely close in rez, and surprass that of 4x5 film when noise (grain) is factored in. When the 16 shot feature is enabled, the resolution far exceeds anything I've seen from 4x5.....something I never thought I'd say about digital capture.

Yes the backs are expensive. But when I factor in the film and processing costs on an annual basis, I'd have $25,000 paid for in about 6 years. It does come down to economics.