Some parameters.

The test that I shot is not scientific. There are no controls against which I tested. The results are entirely subjective but suit my shooting and workflow style. The point was to see whether using my workflow I could obtain better quality for fine art prints from either my digital camera (a Nikon D700) or the Linhof. I already know that another photographer would obtain better results than I have gotten. Remember, not only am I new to LF, but I am using old equipment. The 90mm lens that I used isn’t even multi-coated...practically ancient in comparison to 28-105mm lens on the Nikon (again not the finest of Nikon’s lenses).

Then there is also the financial reality to my test. I purposely did not drum scan the slide. Rather I borrowed a friend’s Epson V750 and scanned the image on default settings using the Epson software (not even the far superior Silverfast Scan Ai). Moreover, I only scanned the image at 2400dpi, well below the 6400dpi output that the Epson is capable of. The reason I did all this is to see whether absolute ‘idiot’ settings would compare at all the image that I shoot with the D700 and which are accepted by Getty Images.

Why not just buy a digital medium format rig and shoot with that (much like is lauded over at Michael Reichman’s Luminous-Landscape site)? Here the financial reality hits home hardest. A medium format back and body with lens is going to set back the photographer some R150 000* (minimum) in South Africa. Even the Nikon D3x is close to the 100 grand mark. My clients are happy with the D700 files and frankly it would take me an excessively long time to make a Phase one or Hassleblad system to pay for itself. No, I’m looking for exceptional image without the insane price tag. Even at close to R100* a shot, the Linhof produces a fair number of images before it reaches anywhere near the cost of a digital medium format.

The shot

The image is a bog standard shot of Durban’s Golden Mile fom the pier at Ushaka Marine World. The digital image is actually a stitched panoramic of two horizontal images (I did say that the test wasn’t scientific and that it was to suit my workflow). The image in mind was a panoramic view of the Mile in the early morning light just before the sun rose.
Here’s the stitched digital image (1/4sec @f10 on 200ISO):

The LF image was shot with the 90mm lens on Agfa RSX 100 (not the best for scanning I know) with an exposure of 4sec at f32. Due to the slightly less dynamic range of the film there was a Cokin P Neutral Density Grad filter stopping back the sky by about 1 stop.  The final image was scanned at 2400dpi, and cropped. Some Photoshop tweaking has taken place, as well as a pass through Neat Image to clean up the grain.

First Thoughts

A number of things struck me working with the Linhof. First off, forget working quickly! The Linhof requires focusing on a ground glass screen (I’ve removed the rangefinder so as to shave some weight) and being mounted on a sturdy tripod. Focusing with the use of a dark-cloth takes some getting used to as well as the upside down reversed image. Then there’s the fact that the f8 aperture produces a murky image that is difficult to see on the ground glass. Focusing is aided with a loupe, and even then I suspect that my image suffers from some softness (this is the second image I have ever exposed in the camera). Another thing that took me by surprise was the interest that I gathered. Everybody who passed me wanted to know what I was up to. So forget candid photography – not that the exposure lengths would allow it. Due to the small apertures required exposures usually sink below the 1 second mark.

Once I had rushed off to the photolab I had to then wait to see any results. This is another thing that I’d forgotten, having shot digital for the last 3 years.  When you finally look at that massive transparency you are figuratively bowled over. It’s huge! And it glows on the lightbox! These sheets are truly impressive. No wonder they were the staple for advertising photogs in the 80’ and 90’s (and before). When a client sees these they are impressed.

An then onto the Scan

In reality all images have to be digital in the end. Recent changes to the stock industry essentially require the photographer to produce a digital file as the end product. In the past, image libraries insisted on keeping the original transparencies. Now they don’t want to even see them. The same goes for international salons like the BBC Wildlife Awards. Two years ago transparencies had to be submitted if shot on film. Now, scans are required instead.

What most people want to hear about is the resolution. Is it worth the added time, weight and cost? The answer is really a mixed bag. Again, the scan was at 2400dpi, so I know that greater resolution is feasible. At 2400dpi though, you obtain an 8bit file of about 250Mb. I did do a 6400dpi scan in the end, but the 8bit file (my computer refused point blank to process the 16bit file) comes to a whopping 2Gb!

	To the left is a 100% crop from the Digital file (from a section to the right of centre of the full image). The same crop at 100% for the 4x5 transparency is shown below.

The above crop of the trannie does look a little soft admittedly. This is my own fault. I'm fairly sure that if i take the time to set the scanner up correctly I'll get a sharper image (The V750 is known to require a bit of tweaking before you get optimal results). Correct use of unsharp mask will also improve matters (once more...I'm not looking for perfect technique in this test. Possibly i'll add another article in the near future with careful scanning and post-processing). Another aspect to note is that the transparency has less detail in the waves. This is a result of the longer exposure that was required for the small aperture. The digital camera with its smaller sensor size only needed an f10 aperture, while the 4x5 camera needed a f32 aperture for a similat depth of field.

To indicate what the digital image would look like were one to interpolate upwards to the same size as the tranparency, I've included these two crops below (at roughly 66%):

The differences are fairly obvious. Whether they are enough to warrant shooting in 4x5 is an entirely subjective decision. Incidentally the image loses some of its apparent softness at 66%, prompting me to suggest printing somewhere below the transparency's maximum printout size (92cm width at 300dpi from a 2400dpi scan).

Conclusion

Large Format photography isn't suddenly going to replace my digital photography. A simple look at the costs involved in producing a large format tranparency quickly eliminates it as 'bread and butter' photography. As a stock photographer i need to constanstly be adding to my library of images, and Large Format simply does not equate to the volumes of images of images that are now requiored to remain competitive. It is another thing entirely though to use it for Fine Art photography. the contemplative approach is wonderful and the potential prints that can be produced - phenomenal. My Fine Art prints which I often produce up to A2 in size will take on a new light. Finally I have more lee way with the circle of confusion and related print size.

Film and digital are completely different to each other. Holistically comparing them is like comparing apples to oranges. You can compare aspects, but they are different fruits entirely. I have really enjoying delving back into film photography and have already gotten great joy from my LF setup. It won't replace my digital equipment, but it will certainly add a tool to my photographic arsenal.

Update 15th april 2010

A reader contacted me to point out that a more conclusive test would be a) shoot something that could not be affected by atmospheric haze (newspaper print for instance), and b) it might be possible to obtain a higher resolution scan from a print, rather than from the transparency.

I certainly agree with a), however very few photographers spend their time photographing newspaper print so we need to find a way to test the real world pros and cons of large format film vs. digital. As to b), the reader has offered to collaborate on a test where we obtain a print from the transparency. I vertainly look forward to the exercise, so stay posted for the results.

On top of the valid comments that I have recieved is the fact that digital has marched inexorably onwards since I wrote the article towards the beginning of 2009. In one short year we have seen the proliferaion of 20mp+ 35mm DSLRs as well as the introduction of the soon to be shipped Pentax 645D. The introduction of the Pentax MF offering isn't so important in itself (afterall if you have the means there are already a number of options to choose from including Rollei, sinar, Hassleblad, Phase 1, and now Leica) as is it's price point, US$9000-10 000. Almost half the price of the competition and only marginally more expensive than the flagship Nikon and Canon models (comparatively speaking).