2

I have a Canon 100-400L telezoom lens. It gives up to 400mm of focal length, but when used on a crop sensor camera, the equivalent focal length would be 640mm.

Because the equivalent focal length would be 1.6x larger on a crop sensor camera, I should according to the theory get 1.62 = 2.56 times more pixels if using the lens on a crop sensor camera with equivalent megapixel count, when compared to a full frame camera.

So, is there any truth to this? Is there a large benefit in using a telezoom lens designed for full frame on a crop sensor camera to get more reach?

  • 1
    What does "equivalent megapixel count" mean? Are you referring to absolute numbers or sensel density? – xiota Jun 29 at 14:37
  • Absolute numbers. Count refers to absolute numbers, not to density. – juhist Jun 29 at 16:23
  • @juhist Absolute numbers of the total uncropped frame? Or absolute number of pixels which show the subject? – Michael C Jun 30 at 15:30
5

By using an APS-C camera with higher pixel density, you will get "more pixels on the subject" than with a FF camera with lower pixel density.

Sometimes that can be a good thing.

There are many wildlife photographers who use APS-C cameras with telephoto lenses for this reason. It allows one to get more "reach" with shorter (and less expensive) lenses without taking the light loss penalty of a 1.4X or 2X extender.

Take, for example, the Canon EOS 7D Mark II. It has 20.2 MP in a 22.4 x 15.0 mm APS-C sensor. A 4.09 µm pixel pitch equates to 51.7 MP on a 36 x 24 mm full frame sensor. Interestingly enough, that's just slightly more than the Canon EOS 5Ds/5Ds R with 50.6 MP.

So if one were comparing the 20.2 MP 7D Mark II to the 50.6 MP 5Ds, there would be very little difference in terms of how many pixels the same subject at the same distance seen through the same lens would occupy on either sensor.

What would be different:

  • Cost: The 20.2 MP 7D Mark II sells for well less than half of the cost of a 5Ds
  • Speed: The 7D Mark II can shoot 10 frames per second for up to 31 raw images or as many jpegs as the memory card can hold. The 5Ds/5Ds R can only shoot at 5 fps for 14 raw or 510 jpegs before bogging down. This is mainly because of the increased processing load of 50.6 MP images versus 20.2 MP images.

On the other hand, if one is comparing the 20.2 MP 7D Mark II to something like the 22.3 MP 5D Mark III, then there would be a noticeable difference when viewing images at 100%. There would be a 53% (1.53X) advantage linearly, and a 2.33X advantage in terms of area.

In the end it comes down to the pixel pitch of each camera, regardless of the total size of the sensor.

As for lens performance of the same lens on an APS-C camera versus on a FF camera cropped to the same dimensions of an APS-C camera: There would be no difference whatsoever. Both images would be using the same exact part of the image circle cast by the same lens. Assuming the resolution of the lens is not the limiting factor, then the sensor with more pixel density (smaller pixel pitch) should give better performance.

The original EF 100-400L IS is a pretty good lens but might be the limiting factor with very dense sensors such as the 7D Mark II or the 5Ds. The EF 100-400L II IS is enough sharper that it should not be the limiting factor with any of Canon's current sensors in EOS DSLRs or EOS R mirrorless cameras.

  • 1
    I somewhat disagree with the "light loss penalty". A 1.6x sized sensor (in linear units) has 2.56x more area (in areal units), so one can set the ISO about 1 + 1/3 stops higher on the full frame camera. A 1.4X extender "loses" 1 stop, but when comparing FF vs crop, the native 1 + 1/3 times sensitivity benefit needs to be taken into account, so there is still 1/3 stop benefit for the FF + 1.4X extender combination. The only thing where the 1.4X extender light loss should matter is DSLR autofocus, but then again mirrorless has solved that. Crop camera is effectively a 1.6X extender with good AF! – juhist Jun 29 at 16:39
  • Take a high quality but still affordable lens, such as an EF 70-200mm f/2.8 L IS II/III, and compare the results of the bare lens on the Canon 7D Mark II vs. lens + EF 1.4X III on a Canon 5D Mark III when both still need to be cropped further to get the framing one desires of a distant animal. Many have done so. – Michael C Jun 30 at 15:23
0

Absolute numbers. Count refers to absolute numbers, not to density

If the pixel count is the same, regardless of sensor size, you will get the same number of pixels, regardless of which one you use.

I should according to the theory get 1.62 = 2.56 times more pixels if using the lens on a crop sensor camera with equivalent megapixel count...

A 1.6x sized sensor (in linear units) has 2.56x more area (in areal units), so one can set the ISO about 1 + 1/3 stops higher on the full frame camera.

A crop sensor is smaller than a full frame sensor. It has less area than a full frame camera. If pixel density is kept constant, a crop sensor will have fewer pixels.

Is there a large benefit in using a telezoom lens designed for full frame on a crop sensor camera to get more reach?

If you need the additional "reach" such that you would be cropping anyway on full frame, you save some time from not having to crop later. Otherwise, a lens that is "sharp enough" for full frame may not be so for crop sensor because greater magnification is needed to produce an image of the same size. But other aspects of the image, such as corner sharpness and vignetting, may be improved because only the central area of the imaging circle is used.

  • I think you misunderstand the question. Wildlife photographers often speak in terms of "pixels on subject". In other words, they are not interested in how many pixels the full uncropped image contains, they are concerned with how many of those pixels are being used to represent a specific subject at great distance. – Michael C Jun 30 at 15:28
-1

The answer is: perhaps, but not as much as you'd think. Full frame lenses are at their best in full frame cameras. While they can be used for crop sensor cameras too, they are not optimized for this use. So, the benefit may be there but it's not 2.56 times more pixels.

Let's take a look at the lens on DxOmark.

DxOmark for 100-400L IS USM lens gives these perceptual megapixels (P-Mpix) counts:

  • 760D: 7 P-Mpix (24 Mpix sensor)
  • 5DIII: 13 P-Mpix (22 Mpix sensor)
  • 5DIV: 14 P-Mpix (30 Mpix sensor)

Considering the smaller sensor size of 760D, one needs to do a 1.6x crop (in length units), or 2.56x crop (in areal units) on the full frame 5D figures to obtain the same area. Thus, the equivalent figures for 5DIII and 5DIV are 5.1 P-Mpix and 5.5 P-Mpix. Normalizing for sensor megapixel count (24 Mpix is 75% of 22 Mpix + 25% of 30 Mpix), we get 5.2 P-Mpix on full frame with a 24 Mpix sensor.

So, yes, a crop sensor camera has some advantage here but not as much as you'd expect. 7 P-Mpix / 5.2 P-Mpix = 1.35 (approximately), so instead of getting 2.56 times more perceptual pixels, you get only 1.35 times more perceptual pixels.

Furthermore, a practical detail: when trying to photograph flying birds, it's hard to get the bird to the viewfinder with a 250mm lens on a crop sensor camera because of the long focal length. A 400mm focal length would be even more insane, making it hard to aim the camera. So, a full frame camera would probably be a better choice to use with this lens, when photographing flying birds: aiming the camera is somewhat easier, and the final photo can be cropped at a minor cost in perceptual pixel count.

However, when taking photos of the moon, a tripod can be used. With a 250mm lens on a crop sensor camera, one can clearly see the rotation of the Earth, but this is mainly a minor annoyance rather than a major problem when taking multiple photos of the moon. A 400mm lens on a crop sensor camera would not be a problem in this case, and the 1.35 times more perceptual pixels would create a slightly sharper image.

If the full frame camera is a mirrorless one (such as EOS RP), there's always the option of using a teleconverter. Autofocus doesn't work with DSLR + teleconverter unless the aperture number is at most 5.6 (there may be some exception cameras to this rule allowing f/8 autofocus), but on a mirrorless camera, f/11 should work with autofocus.

Moreover, one more thing to consider: the 100-400L lens, like most lenses, is sharpest in the center of the picture. I didn't take this into account when calculating the 1.35 times benefit. It may be the case the 1.35 times benefit diminishes to equality when cropping at the center of the frame, which is the sharpest location.

An interesting observation: the cheap 55-250mm has on the 760D 8 P-Mpix, but the focal length does not go up to 400mm.

  • 3
    "Perceptual Megapixels" is a bunch of mumbo jumbo that has nothing to do with actual pixels. It's just the way DxO attempts to reduce a large number of variables down to a single number. – Michael C Jun 29 at 15:26
  • 2
    Re: better performance at the center. Since the APS-C camera is already cropping the image circle to an APS-C sized rectangle, that should give the APS-C an advantage, not a penalty. – Michael C Jun 29 at 15:28
  • There are multiple Canon bodies, including the APS-C 7D Mark II that have at least some AF points that function to f/8. – Michael C Jun 29 at 15:29
  • @MichaelC Is there a better source for lens accuracy than DxOmark? The only other one I'm aware of is the-digital-picture, and it also shows a difference between full frame and crop camera: the-digital-picture.com/Reviews/… so for the crop camera, the lens is definitely the limiting factor. – juhist Jun 29 at 16:32
  • DxO's actual measurements, rather than their "single number" composite scores (for which they do not reveal their methodology and weighting), are more useful. Both DxO and TDP suffer from only testing a single copy of each lens. The DxO scores for the EF 70-200mm f/2.8 L IS II, for instance, are much less impressive than everyone else's. Roger Cicala at his lensrentals.com blog only publishes averages of 10 copy samples when he posts measurements. – Michael C Jun 30 at 15:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.