645-1Ds
Canon EOS 1Ds vs. Canon EOS 10D vs. Bronica ETRSi / Provia 100F
I've read as many reviews on the Canon EOS 1Ds as I could. It sounded like the camera that could replace my medium format film system, but I wasn't sure I would get the same quality. I finally decided the only way to find out would be to buy one and do the test myself. In July 2003 I was able to convince my wife that a 1Ds was a necessary part of my arsenal. (Actually, she is fantastic in regards to my photography and everything else.) The camera arrived on Friday afternoon. Unfortunately, I was working that weekend. I was able to only find time for a few shots from my front porch on that Saturday and Sunday. The subject matter may not be the most interesting, but it should answer my question about quality from the 1Ds. The basic format was to try to take the same shots with each of the three cameras. I thought this would be fairly easy, but with only the 1Ds showing the actual frame in the viewfinder and the 10D's focal length crop it was harder than I thought. All shots were tripod mounted and shot with a remote release. All shots were made at f/8 to try to give the best image quality for each camera. I also threw in a comparison of upsizing with Genuine Fractals vs Extensis pxl SmartScale. The 645 images were all taken with the Zenzanon PE 45-90mm f/4.0-5.6 on Provia 100F and scanned at 4000 dpi on a Polaroid Sprint Scan 120 dedicated film scanner. Each image was resized to the print dimension at 360 ppi with the exception of the third group of images which were 300ppi and then all were sharpened with Nik Sharpener Pro or Focal Blade. All images in each group were sharpened with the same settings and the same method. In several instances, I tried extra sharpening of the 645 images, but it did not improve the quality of the print. All images are a 2" square cropped from a 20x30 or 12x18 print. I then scanned them with my Epson flatbed and applied the same amount of USM to each crop. The resulting images on the monitor should approximate looking at the print with a 2-3x loupe depending on the resolution of your monitor.
 645 Center 20x30 |
 1Ds Center 20x30 |
 645 Corner 20x30 |
 1Ds Corner 20x30 |
The brick images above show much better edge definition in the 1Ds images. There is information present in the 645 image, but it seems almost blurred.
 645 Center 20x30 Genuine Fractals |
 645 Edge 20x30 Genuine Fractals |
 1Ds Center 20x30 Genuine Fractals |
 1Ds Edge 20x30 Genuine Fractals |
 1Ds Center 20x30 SmartScale |
 1Ds Edge 20x30 SmartScale |
Again, the 1Ds images appear sharper. The 645 information is there, but it just doesn't look as good.
 645 Center 12x18 |
 645 Edge 12x18 |
 1Ds Center 12x18 |
 1Ds Edge 12x18 |
 10D Center 12x18 |
 10D Edge 12x18 |
I tried 12x18 to give the 10D a chance, but when you put them next to a 1Ds print of this size, they just don't hold up. Again, the 645 seems to suffer from lack of crispness.
 10D Center 20x30 |
 10D Edge 20x30 |
 1Ds Center 20x30 |
 1Ds Edge 20x30 |
 645 Center 20x30 |
 645 Edge 20x30 |
This is a real stretch for the 10D. Viewed from a distance I think it would be fine by itself, but next to the other two it is lacking in information. The same tradeoff between crispness in the 1Ds and possibly some extra detail in the 645 is seen. I suppose which you prefer is a matter of taste. It's hard to represent these on the screen, but on a print from my Epson 2200 at 2800 dpi I prefer the 1Ds prints by a significant margin. In regards to the two resizing programs; on a print I really don't see a difference. On the magnified views above some very subtle differences are seen.
In summary, I tried to evaluate the quality of the prints made at various resolutions with 10D, 1Ds, and the 645. The 10D is a good camera, but at larger print sizes it can't hold up to the two other cameras. The 1Ds has an advantage in 20x30 inch sized prints over the 645 film. Genuine Fractals and Extensis pxl SmartScale seem to do an equally good job of upsizing the images.