What is the visual difference between Full Frame, APS-C and Micro Four Thirds pictures? [duplicate]

Question

Possible Duplicate:
When do the differences between APS-C and full frame sensors matter, and why?

I am interested in a visual comparison of same scene but with different sensor technologies: One scene with bokeh, landscape, etc.

I found the following link: http://www.four-thirds.org/en/microftmerit/merit2.html

But I am a bit suspicious of the four-thirds omitting some of the more "juicy" details/errors/etc.

Since I have seen many people mentioning the loss of image quality when going to a smaller sensor, for example Micro Four Thirds.

Explanation of the scenes, on what to look for since I am a beginner, is appreciated too.

Something along the lines of what Nikon has: http://imaging.nikon.com/lineup/lens/simulator/

thanks in advance :)

PS. New around here but gonna say it anyway: Let's be civil and not start a flame war. I am just interested purely on the grounds of comparison for future purchases.

Answer 1

The same lens and settings on a full frame will be sharper as you can see in lens comparisons:

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=115&Camera=736&Sample=0&FLI=0&API=3&LensComp=115&CameraComp=9&SampleComp=0&FLIComp=0&APIComp=2

You can play around with lenses, settings and cameras on that page and you will notice that FF is (almost) always sharper, despite that the corner is further away from the center than the aps-c shots.

4:3 cameras are slightly smaller than aps-c but they apparently removed the low pass filter making them sharper than aps-c, more similar to fullframe, but full frame still has the noise/DR advantage:

http://admiringlight.com/blog/micro-43-vs-a-full-frame-legend/

Note that in this test the full frame lens (sigma 50mm) is stopped down past the diffraction limit , so it is not rendering as much details as the camera could with a better lens / setting.

Compare lens/setting here for sharpness comparison (mrk III as they dont have mark II for the sigma):

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=473&Camera=453&Sample=0&FLI=0&API=8&LensComp=115&CameraComp=453&SampleComp=0&FLIComp=0&APIComp=6

Finally there is the FOV and bokeh differences as illustrated here:

When do the differences between APS-C and full frame sensors matter, and why?

The compression/relative magnification factor is another advantage of full frame. To take good pictures of people you need a real 50-135mm (e.g. primes 50mm, 85mm, or 135mm) lens, not a "focal converted 50mm" 28-35mm lens on apsc or 4:3 camera. On the crop sensors even the 50mm lens comes across a bit narrow in your living room, so people tend to go for 28 or 35mm primes on crop sensors, which does give a fov like close to 50mm, but it warps the faces.

Answer 2

Ignoring for the moment aspect ratio, from a theoretical viewpoint there is will be no visual difference, provided you maintain

1. the same subject & camera position / orientation
2. the same angle of view
3. the same resolution (number of megapixels)
4. the same size entrance pupil (focal length divided by f-number)
5. the same lens characteristics

The first two are easy to achieve in practice. The third depends on the precise cameras you are comparing, but you can usually get close, or get the same output resolution by resampling. Given the same eventual resolution and sensor characteristics noise will be similar since according to point 4 each pixel receives exactly the same amount of light, and hence the same photon noise. In reality sensor characteristics differ, e.g. the signal doesn't have to travel as far to the ACD, but the pixel control circuitry takes up a larger proportion of the pixel area.

Point 4 is harder to maintain in practice. It is the size of the entrance pupil, not the aperture designation that determines depth of field and thus amount of background blur. A 100mm f/2.0 lens has an entrance pupil that is 100/2 = 50mm wide. To match the field of view of a 100mm lens using a sensor half as wide requires a 50mm lens. To achieve an entrance pupil of 50mm this lens must be f/1.0. Given these conditions the depth of field will be identical. If the number of pixels is the same then the amount of light hitting each pixel will also be the same. As the f/1.0 lens is two stops faster, it lets through 4 times as much light per unit area onto the sensor. But the pixels in the smaller sensor are half the width and height, therefore each pixel gets the same amount of light.

Maintaining entrance pupil also means the amount diffraction is the same across format sizes. Smaller pixels feel the effects of diffraction sooner, however for the same entrance pupil the small format lens has a lower f-stop, thus experiences less diffraction. In summary there is no respite from diffraction when moving to a larger format as you have to stop down more to get the same depth of field.

Everything is fine so far, both 100mm f/2.0 lenses and 50mm f/1.0 lenses exist (if you look hard enough). But what happens when you use a 24mm f/1.4 lens with a full frame sensor? To get the same entrance pupil (and hence the same background blur) you would need a 12mm f/0.7 lens. Which doesn't exist in practice, except possibly in some scientific instruments. In fact the closest you can usually get at this focal length is f/2.8 or perhaps f/2.0, a whopping three stops short.

At the other end of the focal length spectrum things are a bit better, if you use an 800mm f/5.6 on a full frame sensor, you need a 400mm f/2.8 on a half frame sensor, a lens that certainly exists in practice. So in summary at the telephoto end you can maintain the size of the entrance pupil, but at the wide end it's almost impossible, meaning that wide shots with a larger format may have a degree of background blur impossible to attain using a smaller sensor.

Point 5 is also tricky to obtain in practice. Going back to our first example, Canon make a 100mm f/2.0 lens that is relatively cheap and quite sharp at f/2.0. Canon used to make a 50 f/1.0 lens, but it was far from cheap and very soft wide open. In general it is easier to make a sharp lens with higher f number. It is also easier to make sharp contrasty lens for a larger format since the number of line pairs per millimetre it must resolve is lower (as a bigger sensor has more millimetres over which to project an image). For these reasons shots with a larger sensor may have a level of sharpness and contrast that are impossible to replicate with a smaller sensor.

Answer 3

The background blur will look significantly different. Unless I am mixing things up, a smaller sensor will require a larger aperture for the same background bluring effect, even when you take the variation in focal length into consideration. I.e. if you want nice subject isolation in for example a portrait, it is easier to achieve with a full frame SLR than with a micro four thirds camera.

The smaller sensor, especially at high resolutions will also lead to significant noise at higher ISO settings. On the other hand, in good bright light (without background blur) most cameras, no matter whether SLR, micro four thirds or compact can deliver a good high quality image.

Answer 4

If you stand in one spot, have a person stand in one spot, and take a picture of them with some background (say a lighthouse with the ocean and all that) with all three kinds of cameras, otherwise the same focal lengths and everything, then what you'll see is less background which gives the appearance of a larger subject. If you adjust yourself to have the same subject size in the picture then it will still have a different background (slightly).

On Nikon, if you take a lens intended for a cropped sensor and put it on a full frame camera you get black borders because the image circle is smaller. This is the reverse effect we normally experience with cropped sensors.

Noise-wise... generally, full frame cameras have larger pixels (on the sensor) so they tend to have less noise. However, a newer cropped sensor will likely outperform an older full frame sensor and that doesn't even get into the brand-wars. For noise comparisons you really need to look at them each individually.

Edit: I mention Nikon in the second paragraph. The same holds true for Canon (and probably others brands that have multiple sensor sizes), but Canon doesn't let you put a APS-C lens on a 35mm sensor while Nikon will which is what you would need to do to see this effect.

Answer 5

Do you have access to a full frame camera?

If so, compose and take a photo.

Then, re-compose so that the same exact framing is present on a "cropped" portion of the viewfinder. How much you'll crop will depend on whether you're looking for APS-C or or M43. Take a photo. compare.

Don't pay attention to noise or dynamic range. Those are dependent on the sensors involved and not the size of the capture.

Answer 6

The only differences are only about width/height of the frame: in four-thirds system this fraction is equal to 4/3 in APS-C system is 3/2. Because so the diagonal of 4/3 is less then APS-C with the same height and the lens circle coverage is a little bit less in 4/3. About image quality it doesn't depend about proportions.