Let's say we have two systems:

1) full-frame camera, 80 mm lens at f/2.2

2) APS-C camera, 50 mm lens at f/1.4

The subject is the same, as well as the camera position, focus distance, shutter speed and ISO. And DOF is the same too (because of 80 mm lens and f/2.2 on full-frame). And let's say that we compare two photographs from these system, resized to the same size.

Can we see the difference in the noiseness in this case?

  • I want to say again: we compare two photographs, resized to the same size. Sep 10, 2013 at 16:08
  • In the case of resizing both images to a smaller resolution, you would also need to include the native resolutions of the two sensors to arrive at a fully meaningful answer. But as everyone has said, in the real world with today's cameras, there is virtually no discernible difference when the image is of a well lit scene that is properly exposed.
    – Michael C
    Sep 10, 2013 at 17:49
  • Thank you for the answer. What concerning the size, the point is to compare images that are brought to one size. That's the important moment, we should not care about the native resolution. Sep 10, 2013 at 18:02
  • The resolution comes into play because the light that falls on the 'gap', if any, between sensels is not collected by the sensor.
    – Michael C
    Sep 10, 2013 at 18:10
  • 1
    you aren't talking about the sensors, except their size, making the comparison useless. Different technology is far more important to the noise sensitivity of the sensor than the small difference in size.
    – jwenting
    Sep 11, 2013 at 5:26

3 Answers 3


From the same generation of sensors, it is reasonable to expect that the full frame images will out perform the crop, as the sensor sites will be larger. This gives the engineers more with which to work, e.g. more photons and more area on the die.

But (and this is the reason I wanted to add another answer), for the last couple of generations, today's sensors and cameras have become absolutely remarkable. Ignoring DxO, pixel peeping, and other counting of angels on pinheads and paying attention to real use cases, i.e. on screen display and prints, even discerning viewers are going to have difficulty distinguishing well exposed photos from one sensor to the next. Crop/FF, Canikon/Pentax/Olympus matter far less than good composition and light.

I have an engineering background and it is easy to fall into rather meaningless discussions about various differences, which in truth have very little impact on actual images. In part because these things are measurable, we can debate them objectively, but it is far too easy to lose sight of what matters (and is much harder to objectively discuss), the final images.

  • fwiw (not necessarioly much :-) ), the D700 pixels are about 70% as "noisy" as those on the D800, using DxO's criteria for "noisiness". ie DxO scale noise results down by sqrt (megapixels/12) and the adjusted score for the D700 is a few tads lower than the D800's. Scale D800 down by sqrt(36/12) and multiply by adjusted ISO noise test scores to get actual sensor cell performances. Sep 10, 2013 at 12:52
  • Please consider that I'm talking about the case when we resize both photographs to the same size (i. e. to 5 Mpx). Doesn't a full frame sensor give less noisy picture then? Sep 10, 2013 at 17:11
  • Almost ANY reduction in size from either camera will also reduce the impact of noise, as the values for multiple pixels are usually averaged when the image is re-mapped to the smaller size.
    – Michael C
    Sep 10, 2013 at 17:43

Not necessarily.

There are two factors here:

  1. Technology gets better, a newer crop sensor with better technology can have less noise then an older full frame sensor (this is especially true in high ISO where technology moves very fast)

  2. Noticeable noise is caused by not having enough light, if the lighting is good and there isn't that much noise in the picture to begin with than the difference between crop sensor and full frame will be negligible.


Theoretically - Yes. Practically - Not really.

Since the full frame camera is stopped down to f/2.2 compared to the APS-C camera at f/1.4, under the same lighting conditions the same amount of light is falling on both sensors. But the field density of that light is not the same. The same amount of light is being spread over a larger area with the FF camera, making the image dimmer. If we want the images to be the same brightness we need to either increase the exposure time or increase the ISO to compensate for the lower f-number. This would take away the most significant advantage that the full frame camera has: collecting more light when using the same f-number as used with the APS-C camera.

The entrance pupil of an 80mm lens at f/2.2 is 36.36mm. The entrance pupil of a 50mm lens at f/1.4 is slightly smaller at 35.7mm. The minor difference between the two is probably less than the difference between actual aperture and stated aperture of the two lenses in question. So to even be able to calculate the minute theoretical difference we would need to know the absolute measured apertures of each lens at the indicated settings, the absolute measured ISO sensitivities of each camera, and the absolute measured shutter speeds of each camera (compared to the set Tv).

We must then consider the relative resolutions of each sensor. We know the relationship in terms of overall size between each, but we don't know the pixel pitches of each, how much 'gap' there is between the pixels for each sensor, nor the overall resolution of each sensor. All of these factors would come into play, but just like the question of aperture, they are all probably so insignificant in their differences as to be negligible in terms of a detectable difference by an observer at standard viewing conditions.

How you choose to apply noise reduction will likely have a much greater effect than the small differences each system is natively capable of in terms of handling noise. And that is only in situations where the available light is low enough that the noise is an appreciable percentage of the signal generated by the limited light.

  • As I've already said, consider that I'm talking about the case when we resize both photographs to the same size. Doesn't it make sense? Sep 10, 2013 at 17:14
  • Yes, in theory you are correct that there will be a difference, but in practice the difference is usually not detectable under standard viewing conditions (8x10 inch print viewed at 10 inches by a person with 20/20 vision). The conditions at the time of exposure and the amount of noise reduction selected will have a far greater impact than the difference between the performance of the two sensors under identical conditions.
    – Michael C
    Sep 10, 2013 at 17:17
  • If there's no significant difference, why many people talk about the less noise level and better colors/dynamic range on full-frame sensors then? Sep 11, 2013 at 5:02
  • Because you can use the lens at f/1.4 on the full frame camera, you don't have to stop it down to f/2.2 to match the performance of the APS-C sensor. Opening up to f/1.4 from f/2.2 increases the signal (light reaching the sensor) by just over one stop, so the signal to noise ratio is improved by just over one stop.
    – Michael C
    Sep 11, 2013 at 5:23
  • 1
    ...You are correct that the APS-C sensor that is roughly 40% the size of a full frame sensor collects only 40% of the light cast by the same lens at the same aperture as the full frame sensor does.
    – Michael C
    Sep 15, 2013 at 6:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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