Andre's Blog
Perfection is when there is nothing left to take away
DOF: 35mm vs. APS-C

I was always curious to see how different would be a scene captured on 35mm film and with a digital camera, so I dusted off my good old Canon EOS SII, bought a fresh battery and a few rolls of film. The tests described in this post were captured on FujiFilm ISO 160, color-balanced for daylight, which is the only low-sensitivity film the local photo store had in stock.

I set up the 7D on a tripod 50cm away from my test subject, measured between the koala nose and the back of the camera, and used the built-in electronic level to make sure the camera is horizontal. The shot was taken at ISO 100 and f5.6 through a 18-250mm Sigma lens at 50mm. Then I switched cameras and took a shot with the SII at f5.6 though a 35-80mm lens at 50mm. I did not use the same lens because lens parameters, except the focal length, should not have any effect on the field of view and the subject size.

I had the film developed and scanned it at 3556 dpi, which is 140 px/mm or 70 lp/mm, and seems to be considered as good resolution for film. The SII shot was slightly sharpened with a high-pass filter. Hover your mouse over the captions to see the SII and 7D shots. The one labeled 7D is the entire 7D frame and the one labeled 7D x1.6 is the same shot scaled down 1.6 times.

50mm, f5.6 -- [SII 50cm | 7D x1.6 50cm | 7D 50cm]

The koala in the SII image is about 87px in height, while the one in the 7D image is about 140px, which yields 140px/87px = 1.6, which is expected. The second image shows the 7D image scaled down 1.6 times, so it's easy to see that the subject in focus is the same size.

For the second test I moved the tripod further away to 50cm * 1.6 = 80cm and took a shot with the 7D using same settings - 50mm and f5.6. The expectation was that the subject will now be the same size and the field of view will also be comparable between the shots.

50mm, f5.6 -- [7D 80cm | SII 50cm]

The subject did remain approximately the same size (82px vs. 86px), the field of view is not even close in these two shots.

While the SII was still on the tripod, I took a shot through its viewfinder with my compact and later moved the 7D until I saw the same field of view, which was 133cm away.

50mm, f5.6 -- [7D 133cm | SII 50cm]

Later, when I looked at the developed film, I noticed that even though I marked the location of the tripod legs on the floor and tried to move it parallel to the initial line of sight, I still missed about 10% of the field of view on the right because the viewfinder image captured by the compact was too dark on the right side and I missed the correct frame boundary. I also knew that the SII viewfinder coverage was 90%, so the 133cm I used was approximately 20% off and the actual distance should be somewhere in the 133cm * 1.2 = 160cm ballpark.

Even though the last number is approximate, it still tells me that I have to move an APS-C camera roughly 3.2 times further away from the subject in order to capture the same field of view as a 35mm camera would capture at the same focal length.

Conclusion

  1. A 50mm lens remains a 50mm lens on any camera and the depth of field and the background blur will be the same for a full frame and APS-C sensors, given the same focal length, aperture and focusing distance.
  2. An APS-C camera has to be 1.6 times further away from the subject to capture the subject of the same size as a full-frame camera would capture. The field of view will be different between these cameras.
  3. Given the same size of a subject in the frame, depth of field of an APS-C shot will be greater than in a full-frame shot because the APS-C camera has to be further away from the subject and greater focusing distance means greater depth of field.
  4. In order to capture the same field of view, an APS-C camera has to be further than 1.6 times away than a full-frame camera. My tests seem to hint 3.2 times, but this is not conclusive due to lack of data.

 

Comments:
Posted Sat Feb 5 20:21:07 EST 2011 by Mistake

Koalas aren't bears... 

Name:

Comment: