11. June 2003 (rewritten with new examples : 7. July 2003)
Digital versus film. Still a hot topic, although there seems to be a growing acceptance that digital belongs to the future while film is a thing of the past.
In this article I'll show you what I saw when comparing one of the newest D-SLR's with my scanned 35mm film.
This article is the first in a small series of articles comparing a high-end D-SLR with scanned film.
I'm comparing the Canon EOS 1Ds with pictures from my normal workflow.
[Update 11.07.2003] The third article, comparing prints from all three formats (digital, 35mm and medium format) can be found here.
[Update 17.06.2003] The second article, comparing to scanned medium format, can now be found here.
Warning!
By the very nature of this article - exhibiting image quality and details - this document contains a large number of fairly big pictures and it will therefore take a while to load if you are on a modem connection. Please be patient.
Why and how?
So why actually bother with comparing film and digital? Can it be compared at all?
Well, I have to say, that my main reason for comparing digital with film is simple curiosity. How good has the D-SLR's become?
There are a large number of differences between using film and digital that I won't bother to dig into here. I'm mainly focusing on and interested in image details here.
So basically, since I've seen a lot of examples on-line that suggested that digital could outperform scanned film including scanned medium format, I wanted to test that myself!
Please keep in mind, that none of these tests are "scientific" in any way. The sole purpose here, is to compare my normal standard of scanned film with the output one can get from a high-end D-SLR these days.
It's NOT a review of the Canon 1Ds. Please note, that the lenses used are not directly comparable. I'm using my own gear - the equipment that I know - to compare with the 1Ds.
So, before I go on, I would like to stress, that there are a number of factors that I do NOT take into consideration. These includes, different lenses, different lens-design, exposure time might differ slightly, focus might differ slightly etc. It might also be, that another film and/or another scanner would have produced a different result. So, again this article compares the output I saw from the 1Ds to "my normal way" of making pictures.
A very reputable store in Copenhagen - Photografica - would let me shoot with a Canon 1Ds and compare it to 35mm and medium format.
So here is the list of cameras I used to conduct the test:
Canon EOS-1Ds with a Canon Zoom EF 28-70mm f2.8 L set to exactly 50mm.
Nikon F90x with a Nikkor 50mm f1.4
Rolleiflex 2.8GX which has a Rollei Planar 80mm f2.8 HFT lens build in.
(Results from the Rolleiflex will be shown in a coming article).
Expectations
What did I expect to see?
Well, from what I've seen online I expected the 1Ds to outperform my 35mm scans, but not quite meet the levels of details of my medium format scans.
This article focuses on the 35mm scans.
The test
As detailed above, the test was conducted with three cameras - Canon 1Ds, Nikon F90x and Rolleiflex 2.8GX.
Settings on the Canon 1Ds
ISO : 100
Quality : RAW (plus jpeg).
White Balance : Auto
Color Matrix : 4 (AdobeRGB)
Sharpness : 0
Firmware : 1.0.1
The RAW files are converted to TIF using a 30 day trial version of Capture On DSLR 1.1.
I used Provia 100F for both film cameras. I now that Fuji Velvia is capable of resolving slightly more details, however I'm more used to the Provia and it's easier for my scanner to handle.
Remember that I wanted to see, how the 1Ds compared to the image quality I get from normal "workflow".
The 35mm frames are scanned at 4800dpi using my Minolta Dimage Scan Multi Pro. This produces a print of approximately 13x19 inches at 360dpi.
The 1Ds files are enlarged to match the size of the scanned film using Photoshop bicubic interpolation. Each picture is manipulated as I saw fit (levels adjustments, sharpness etc.) to produce what I found to be the best possible picture.
In other words, a bit of image processing has been made to optimize the output from each format and to simply reflect a normal workflow in which one would optimize a bit here and there. I didn't bother cleaning the scanned film from dust though.
The results
OK, enough talking, what did I find?
The first example is this picture:
Not exactly art, but still with a lot of details to compare. The red circle above shows the approximate focus point.
1Ds versus 35mm
|
Canon 1Ds. 50mm f8 1/100sec. RAW file.
|
35mm Provia 100F. 50mm f8 1/60sec.
|
The 1Ds picture exhibits more details, not at least more shadow details than my Provia scan. The total lack of grain and noise on the 1Ds picture looks very impressive. The 1Ds picture is enlarged to match the size of the scanned film.
Color balance is not identical although very close. I assigned the Canon Daylight profile that came with the Capture One software which seemed to give the closest match to my Provia scans. I shot in a RAW + jpg mode on the 1Ds in which you get both the RAW image and a jpg image. Here's the jpg image that went along the RAW file:
Still lots of details but a few jpg artifacts can be seen especially if the image is sharpened more (the picture here has received some sharpening, but is NOT processed via Capture One, like the others). Enlarged to match the size of the scanned film.
Two more example from the same picture:
|
Canon 1Ds. 50mm f8 1/100sec. RAW file.
|
35mm Provia 100F. 50mm f8 1/60sec.
|
|
Canon 1Ds. 50mm f8 1/100sec. RAW file.
|
35mm Provia 100F. 50mm f8 1/60sec.
|
We see the same trend as above. The images are close, but it's impossible to find a single spot in which the 35mm frame shows more detail than the 1Ds is capable of. The total lack of grain in the 1Ds picture makes it look sharper and less "muddy".
The 35mm picture is "second generation" - it has been through my scanner, which introduce some shadow noise.
Let's look at a different picture:
Details:
Very slightly in front of the focus plane:
|
Canon 1Ds. 50mm f16 1/2sec. RAW file.
|
35mm Provia 100F. 50mm f16 1/2sec.
|
Approx. in the focus plane:
|
Canon 1Ds. 50mm f16 1/2sec. RAW file.
|
35mm Provia 100F. 50mm f16 1/2sec.
|
Very slightly behind the focus plane. (White point set manually in Photoshop):
|
Canon 1Ds. 50mm f16 1/2sec. RAW file.
|
35mm Provia 100F. 50mm f16 1/2sec.
|
All the way through, we clearly see more details in the 1Ds pictures.
Conclusion
When using a prime lens (50mm, F1.4) on the film camera and a zoom lens on the digital, I might have given the film camera a slight advantage(?) Still, the 35mm pictures does not in my test here exhibit more details than the digital ones - on the contrary.
We see the grains clearly on the 35mm scans, so I doubt that there are much more details to be extracted - if any at all - using an even better scanner. Except maybe for the shadow areas in which a better scanner would be able to extract even more - perhaps.
The 1Ds pictures are free of grain and noise and they therefore looks very clean and, because of this, they can be sharpened a lot more than the scanned film.
Stay tuned, to see the next article in which I compare the 1Ds to scanned medium format transparencies.
If you subscribe to my newsletter, you'll be kept informed!
Links
[Update 11.07.2003]:
• Canon 1Ds vs. film - Prints - the third article is now ready.
[Update 17.06.2003]:
• Canon 1Ds vs. film - medium format - the second article is now ready.
• Photografica Visit Photografica in Copenhagen
• Luminous-Landscape The first of a number of articles about the Canon 1Ds
• DPReview Read a review of the Canon 1Ds by Phil Askey
• Phase One Visit Phase One - the company behind the Capture One software.
• Norman Koren Digital cameras vs. film - don't miss this very informative site with tests and comparisons.
|
|
|
