As you probably know, the AF100 has a 2x crop factor when compared to the "standard" 35mm film frame (still photography, not 35mm cinema). Thus, the common logic is that the 50mm lens on an AF100 would perform exactly as a 100mm lens on a 35mm still camera. That's both right - and WRONG at the same time.

The crop factor *only* applies Field of View and not magnification. Here's a video explaining crop vs. magnification:

FOV Crop - YouTube (http://youtu.be/SCJOeRh349s)

While the video talks about still cameras the principles of sensor-cropping and lens characteristics are exactly the same for video cameras as well.

Field of view is just ONE of the components of any lens' characteristics, the other important factor is Compression. Again, here's another video to explain the phenomenon:

Telephoto Compression - YouTube (http://youtu.be/fX7iBhvYUqM)

To put into perspective:

An 85mm lens has always been deemed a "portrait" focal length or short telephoto (greater than 50mm = telephoto. Less than 50mm = wide angle), but on the AF100 an 85mm lens would have the *field of view* of a 170mm lens! To most that's considered too long for that kind of work as it would force the camera to be pulled further away from the subject to have the same framing.

Thusly, most would then consider that for a 2x crop factor that an "85mm equivalent" should be used instead, which would make it about a 42mm lens. That's fine, but 42mm is actually wide-angle, so depending on your distance to subject you're actually going to get some wide-distortion just as you'd expect to see from a wide-angle lens, not a mild telephoto.

Yet another video explaining this focal-length to subject relationship:

Lenses and Subject Perspective - YouTube (http://youtu.be/O_qe8mt9Cfs)

It may take a while to fully understand how all this plugs into the AF100 lens selection, but basically you can think of it this way: If you like the way any image looks using a particular focal length, then use that SAME focal length for the AF100 and simply plan on setting up your shot to accommodate having 50% percent of the field of view being cropped out.

Just remember, any crop factor relates solely to field of view (FOV), not magnification.

Thanks Robert.

One thing worth mentioning is that image compression has more to do with distance to subject than it does with the lens used. If you stand on the same spot and take a photo at 85mm with a full frame camera, and 42mm with a m4/3rds camera, the FOV and compression will be the same. This is because the photo is taken from exactly the same spot so each object within the image will appear with exactly the same proportions.

Where you will see a difference is when you change the distance to subject - for example if put the 85mm on the m4/3rds and back up to achieve the same framing as the Full Frame photo, then you will give the image a more compressed look that with the 85mm. Likewise, putting the 42mm lens on the full frame camera and moving forward to achieve the same framing as the 85mm lens will likely intoduce some WA distortion and a significantly different image compression.

It's kind of similar to the sensor-size/bokeh debate. It is a generalisation that larger sensors give shallower DOF, but this is not quite correct. DOF is determined entirely by the lens, but a sensor size determines how we might use a particular lens. For example, with an 85mm lens on a m4/3rds camera we might move further away to achieve the same framing as a full frame camera - and in doing so we increase the DOF. Or we might stay in the same spot and use a wider lens instead, which will also have the effect of wider DOF. The lens determines the DOF, but the sensor dictates how we can use that generated DOF and FOV.

Likewise with image compression - the perspective is entirely a product of distance to subject, but we can use different lenses and sensor size combinations to manipulate the look we want.

If you stand on the same spot and take a photo at 85mm with a full frame camera, and 42mm with a m4/3rds camera, the FOV and compression will be the same. This is because the photo is taken from exactly the same spot so each object within the image will appear with exactly the same proportions.

That's actually not true, and why in Barry's book he uses the line, "A 50mm lens is a 50mm lens is a 50mm lens....". The compression/magnification of any lens doesn't change based on what mount it's connected to, ever.

To use your example, a 42mm lens doesn't attain an 85mm magnification when it's on an MFT mount, it's still a wide-angle lens with the same magnification level. So to use your example above, that 42mm lens would actually show background elements further away because it's still a wide-angle lens and doesn't have the same magnification level of an 85mm - it simply has about the same FOV as an 85mm - when on an MFT mount.

If you watch the videos referenced above it's clearly explained, this relationship between crop sensors, full-frame FOV/DOF and compression.

Alternatively, see Barry's AF100 book from pages 98 to 107 - but the videos make it easier to understand. It's a total brain-tease and has fooled many people, myself included until I studied it.

It's a total brain-tease and has fooled many people, myself included until I studied it.

I just LITERALLY had my AH HA moment last week. All makes sense now.

The crop factor *only* applies Field of View and not magnification.


That's true, and it's a good point that is worth mentioning. However, it's important to clarify which definition of "magnification" you are referring to. In your case, your statements apply only to subject space to image space magnification. Your statements do not apply to reproduction magnification (AKA enlargement ratio) or total magnification (from original object space to the final display), because the size of the imager does come into play for both of those. For example:


A person with a head that is 312mm tall in subject space (i.e. the real world).
135mm lens
On a 36x24mm camera in landscape orientation
Framed so that the head just touches the bottom and top of the sensor.
Displayed on a 30" LCD with a viewable height of 10" (254mm).


Here are the three types of magnification for this full-frame example:


The image space magnification is 24:312, or 1:13, or 0.08X.
When the 24mm-height image on the imager is blown up for display on the LCD, the reproduction magnification is 254:24, or about 11:1, or 10.58X.
The total magnification (from size in object space to size on the display) is 254:312, 0.8:1, or 0.81X.


Now consider how format size factors into the scenario, using the same lens (which is not how someone would use the camera in real life, but for illustration only):


A person with a head that is 312mm tall in subject space (i.e. the real world).
135mm lens
On a 18x12mm camera in landscape orientation
With the camera in the same position as before (framed so now only half the face is visible: 156mm)
Displayed on a 30" LCD with a viewable height of 10" (254mm).



The image space magnification is 12:156, or 1:13, or 0.08X. Same as above.
The reproduction magnification is 254:12, or about 21:1, or 21.16X. This is twice as high as before.
The total magnification (from size in object space to size on the display) is 254:156, 1.6:1, or 1.63X. This is also twice as high


As you can see, switching format without adjusting the focal length to get the same angle of view (as any photographer would actually do in real life) causes a significant difference in both reproduction magnification and total magnification. Only image space magnification remains unchanged between format sizes.

That said, it is useful to point out, as you did, that changing format sizes does not affect the size of objects in the image circle.


That's fine, but 42mm is actually wide-angle, so depending on your distance to subject you're actually going to get some wide-distortion just as you'd expect to see from a wide-angle lens, not a mild telephoto.


I agree with everything you said right up until this sentence. There are several important mistakes:


So-called "wide-distortion" has nothing to do with the focal length. It is wholly and completely explained by perspective, i.e. the position of the camera. (Or more technically, by the center of perspective, which is the center of the lens entrance pupil.)

That can be easily proven. Take photos of the same subject from a variety of focal lengths while the camera and scene remain unchanged. You will find the exact same amount of wide-distortion (or lack thereof) at every focal length. (Assuming a fixed entrance pupil.)

If you meant to say another type of distortion, such as barrell, pincushion, etc., then it would still be incorrect, since there is no law that a 42mm lens will have more or less distortion than a 85mm lens.

There is nothing about a 42mm lens being used on Four Thirds that is "actually wide-angle".
First, the 42mm lens may have been designed specifically for Four Thirds, such that when mounted on a full frame camera the image circle does not even cover the format, making it physically impossible to achieve a wide angle image.

Second, even if the 42mm lens was designed with an image circle large enough to cover full frame, and a lens mount that would work on a full frame camera, it could be that it was still not *designed* to be used on full frame. It may have been designed so that the Four Thirds part of the image circle is very sharp and high contast, while the outer full-frame part turns in very poor contrast and resolution.

Third, if the 42mm lens was designed as a wide angle lens for full frame, it will still provide the exact same angle of view and perspective as a 42mm lens designed for a smaller format. So rather than say "It's actually a wide-angle lens," I think it's better to say "It was designed to provdie a ide angle of view on full frame, but on Four Thirds it provides a narrow angle of view."


By the way, personally, I consider 40 - 60mm focal lengths as "normal" on full frame, with "wide" starting more around the 35mm territory. But you may be correct in that anything under 50mm is technically "wide", I don't know.


If you like the way any image looks using a particular focal length, then use that SAME focal length for the AF100 and simply plan on setting up your shot to accommodate having 50% percent of the field of view being cropped out.


That is wrong. The correct method is to choose the perspective and angle of view you want, then scale your focal length with format size. Longer focal lengths on larger formats, and shorter focal lenghts on smaller formats.

Say I like the way my 8mm lens looks on my 1/3" format for a certain half-length portrait. The same focal length on full frame would be an extreme 180-degree circular fisheye. If I use a 56mm lens, however, I get the exact same angle of view, field of view, and perspective as the 8mm lens (and vice-versa).

One thing worth mentioning is that image compression has more to do with distance to subject than it does with the lens used.


Great post, John. I see you put things a little more mildly here than you did in your conclusion. I prefer the stronger wording of your conclusion ("perspective is entirely a product of distance to subject"). Emphasis mine. :)


DOF is determined entirely by the lens, but a sensor size determines how we might use a particular lens.


That's a good way to put it. Another way to frame it is in terms of lens availability. For example, format size determines lens availability for a given angle of view. Larger formats only tend to be capable of thinner DOF because they tend to have lenses with wider iris diameters available. But not always (e.g. Medium Format is not capable of the same thin DOF as 35mm due to lens availability).

So to use your example above, that 42mm lens would actually show background elements further away


That is not right. Compression and perspective will remain the exact same as long as as the camera position (or technically, entrance pupil) remains the same. The subject and background will both shrink or enlarge by the exact same degree, no matter how much you change the format size and/or focal length.