Without a doubt, your focal length choice affects the impact of your image, whether you are shooting video or stills. With the rise of digital imaging, there has been an uptick of available formats in which you can shoot. To help clarify the relationship between format and focal length, some have coined the phrase ¡°35mm equivalent.¡± This term has value when you are planning a shoot, your lens package, and as an aid for anyone who began their careers using 35mm film and has transitioned to shoot in a new digital format. However, the final determining factor of your image is the perspective that your lens will impart¡ªwe’ve all heard that ¡°a 25mm is a 25mm, is a 25mm¡± and that if you put a 25mm on a different format it will behave the same, no matter what. Or will it?
Above photograph by Steven Gladstone: The same 50mm lens on four different sensor sizes yielding the same image?
?
I have heard the arguments from all perspectives and decided to put it to the test for myself. I felt it important not to merely crop the image from a full-frame sensor, but rather to compare different formats with a lens native to each format. That is, correct native lenses to match the sensor size.
As it turned out, testing with different-sized sensors was very revealing, although not for the reason I set out to explore!
I enlisted the help of my friend, Justin Dise, a videographer who has written a few articles such as Selecting a Lens Package, in which he discusses what makes a normal lens normal. He brought along his Sony a7R II and a series of vintage Nikon AI-S lenses and a Voigtlander Nokton 40mm. I supplied a Panasonic G7 and? the Blackmagic Pocket Cinema Camera (BMPCC from now on), plus 20mm and 14mm Panasonic prime lenses in MFT mount. We used a Nikon to Sony E, and Nikon to MFT Adapter to round out our camera-lens packages. For support, we used an E-image EJ60AAM Tripod package and a Magnus VT300.
?
Just some of the gear, not enough extra handsSteven Gladstone
?
?
The tests
We made a few tests utilizing a frame with 16:9 aspect ratio, shooting the highest resolution of each camera in that format, unless otherwise noted. With the a7R II and the G7 we shot stills, and with the BMPCC we shot HD video. I used a combination of DaVinci Resolve and Photoshop to compare images.? Test 1 was shooting the 50mm Nikon in each format. For Test 2, we shot the 40mm on a full-frame sensor of the a7R II and the 20mm on the G7 with the cameras in the same position.
?
?
Cameras lined up next to each other; pay no attention to the man behind the Panasonic Lumix
?
Test 1
This is pretty straightforward, with no surprises. As you record on a smaller format, your field of view gets smaller, which is exactly what you would expect, but important to establish as a ground rule. If you want to compare cropping your sensor to the images shot natively, please feel free. There were some small variations since the lens height was different from camera setup to camera setup.
50mm on BMPCC Super 16mm
50mm on Panasonic G7 Four Thirds
50mm on Sony a7S MarkII APS-C?
50mm on a7S MarkII Full Frame
?
Test 2
This is where things became more interesting, because it is all well and good to crop your image in post or use a smaller sensor, which only captures part of what your lens is delivering. The question is: how does that affect the image your lens produces? We set up the a7R II with the 40mm Voigtlander next to the G7 with the Lumix 20mm and matched lens center height and frame, and then had a subject (me) walk toward the cameras. While at first it appeared that there was a difference in perspective between the two setups, I was concerned that this may have been due to the cameras not being on the same axis. This led to repeating the test, only this time we used a single camera position, and were able to configure the two cameras so that their lens center was very nearly the same.
40mm on Sony a7S MarkII Full frame
20mm on Panasonic G7 Four Thirds
We took great care not to adjust the position of the tripod after setting it. Based on the 2x crop factor, a 20mm lens on a Four Thirds/MFT sensor would yield the same angle of view as a 40mm lens on a full-frame sensor, and it truly did. What I was expecting, however, was that the shots would look different, that the perspective would change and objects farther away from the lens would be significantly smaller as the wider lens would exaggerate the distance instead of compressing it. What makes this framing so appropriate to test the theory of focal length affecting perspective is that it includes a receding fence with lampposts that can be used to judge height and also recede to the vanishing point. Additionally, off in the distance, just off center is the Statue of Liberty. If I was correct, then with the 20mm on the G7, the fence and lamp posts would become noticeably smaller the farther they receded from the camera, compared to the 40mm on the a7R II. Likewise, the Statue of Liberty would be significantly smaller in the shot with the 20mm on the G7. As you can see below, that did not happen¡ªMUCH to my surprise.
?
40mm on Full Frame
?
20mm on Four Thirds (MFT)
?
?
?
Okay, but what is really important is how the lens perspective affects the background behind a person. Enough with matching the field of view, I¡¯m surprised the perspective is the same in this example, but what about if I¡¯m specifically matching a human face between the two formats. How will lens perspective change then?
Lost horizons
Setting up an actor, we proceeded to see how the format affected the image produced by the lens. The results, from a testing standpoint, were disappointing and, in the end, revealed certain weaknesses in the testing procedure. I’ve included those here, for full disclosure.
The first part involved shooting the 50mm in all four formats¡ªfull frame, APS-C, Four Thirds, and Super 16. The goal was to frame our actor so that he was the same size in all four frames. This would involve moving the camera farther away from the actor as we changed formats.
As you can also see from the images below, the amount of visible environment in the frame around the actor clearly changes.
50mm Super 16
50mm Four thirds
50mm APS-C
50mm Full Frame
?
For the next pass, I chose to compare a 28mm lens on a full-frame sensor to a 14mm lens on a Four Thirds sensor, because this would prove that there is a difference between the perspectives that each lens delivers. So, we mounted the Nikon 28mm on the a7R II and the Lumix 14mm on the G7. The results are below, and check out the chart for the distance between the camera and the subject for each lens/sensor combination.
?
14mm on Four Thirds (MFT)
28mm on Full Frame
The image from the a7R II is a frame grab from a full-frame video in HD. With the G7, I shot a still so as to maintain the 2:1 crop factor for comparison (the G7 crops its Four Thirds-sized sensor in 4K video). Both were shot in 16:9 mode. Look at the composite of the two below, where the 14mm is overlaid at 45% opacity on top of the 28mm full frame. Note the lamp post and the peak of the building.
?
?
Composite of above two images
?
There¡¯s something rotten in Denmark
After careful scrutiny, I determined that something was not right¡ªthat the actor didn’t maintain a constant position and the ground was too uneven, so tilt was introduced. It was obvious that far too many unanticipated variables affected the test, but I was determined to try again.
New premise: depth of field¡ªaye, there’s the rub
I wanted to simplify the test to compare a lens on a full-frame camera to a lens of half that focal length on a Four Thirds (MFT) camera when photographing a person, to see how the image changes when you have matching frames. To remove as many variables as possible, I used a wig head on a stand so it wouldn¡¯t move, and I placed various registration stickers on it so I could compensate for faulty registration (when the camera’s sensors and lens axis do not match). I shot indoors and didn¡¯t tweak the camera position for the test.
I wanted to push the test as much as possible, so I used a 28mm on the full-frame camera and a 14mm on the Four Thirds camera, hoping to see a difference in the amount of distortion each image delivered. For full frame, I shot with a Sony a7S in full-frame mode and in 3:2 aspect ratio. The 12MP sensor would better match the Panasonic G7’s sensor also shot in 3:2 mode. I took a few test shots, and saw that the images were basically the same¡ªexcept they weren’t. With everything matching between the two setups (distance, f-stop), the perspective of the lenses matched, but there was a noticeable difference in depth of field.
14mm Four Thirds f/2.8
28mm Full Frame f/2.8
So I shot with each camera, racking through the f-stops while changing shutter speed and ISO for consistent exposure. By now I’d completely accepted that ¡°35mm equivalent¡± is a valid term, and that perspective is really only affected by distance from the sensor plane to the subject. Two cameras with different-sized sensors will produce images with a matching perspective, as long as the lenses used produce the same angle of view in their respective formats. This is where equivalent focal lengths come into play: a 28mm on a full-frame camera delivers the same image as a 14mm on a Four Thirds camera. Granted, there may be slight differences due to different optical designs, distortion, and any lens corrections performed by the camera.
However, the real difference is evident in the depth of field. To get roughly the same depth of field with the full-frame camera as with the Four thirds, I had to narrow the aperture 2 to 3 stops more¡ªf/5.6 or f/8 on the a7S, compared to f/2.8 on the G7. See the images below, so you can reach your own conclusions.
28mm Full Frame f/8
14mm Four Thirds f/2.8
28mm Full Frame f/5.6
?
All’s well that ends well
The camera¡¯s sensor format has a decided effect on the field of view of the lens, causing the need to move the camera closer or farther away from your subject to get the desired framing. Also, the closer the camera is to the subject, the more exaggerated the distance between subject and background will appear to be; the farther away the camera is, the more compressed distance will appear. This isn¡¯t a focal-length issue¡ªit¡¯s a distance-to-subject issue.
At the end of the day, I feel satisfied that the phrase ¡°a 25 is a 25 is a 25¡± isn’t accurate once you start changing formats. Put that theory to bed, I’m convinced. The format affects not only the perspective of your lens, but the depth of field of your image, and now I¡¯ve got the images to prove it.
