digital camera series

Choosing an APS-C camera: 26MP or 40MP?

/ spqr / Leave a comment

The most obvious choice when it comes to APS-C cameras is usually the number of megapixels. Not that there is really that much to choose from. Usually it is a case of 24/26MP or 40MP. Does the jump to 40MP really make all that much difference? Well, yes and no. To illustrate this we will compare two Fujifilm cameras: (i) the 26MP X-M5 with 6240×4160 photosites, and (ii) the 40MP X-T50 with 7728×5152 photosites.

Firstly, an increase in megapixels just means that more photosites have been crammed onto the sensor, and as a result they have been reduced in size (sensor photosites have dimensions, whereas image pixels are dimensionless). The size of the photosites in the X-M5 is 3.76µm, versus 3.03µm for the X-T50. This is a 35% reduction in the area of a photosite on the X-T50 relative to the X-M5, which might or might not be important (it is hard to truly compare photosites given the underlying technologies and number of variables involved).

Fig.1: Comparing various physical aspects of the 26MP and 40MP APS-C sensors (based on the example Fuji cameras).

Secondly, from an image perspective, a 40MP sensor will produce an image with more aggregate pixels in it than a 26MP image, 1.5 times more in fact. But aggregate pixels only relate to the total amount of pixels in the resulting image. The other thing to consider is the linear dimensions of an image, which relates to its width and height. Increasing the amount of pixels in an image by 50% does not increase the linear dimensions by 50%. For example doubling the photosites on a sensor will double the aggregate pixels in an image. However to double the linear dimensions of an image, the number of photosites on the sensor need to be quadrupled. So 26MP needs to ramp up to 104MP in order to double the linear dimensions. So the X-T50 will produce an image with 39,814,656 pixels in it, versus 25,958,400 pixels for the X-M5. This relates to 1.53 times as many aggregate pixels. However the linear dimensions only increase 1.24 times, as illustrated in Fig.1.

So is the 40MP camera better than the 26MP camera? It does produce images with slightly more resolution, because there are more photosites on the sensor. But the linear dimensions may not warrant the extra cost in going from 26MP to 40MP (US$800 versus US$1400 for the sample Fuji cameras, body only). The 40MP sensor does allow for a better ability to crop, and marginally more detail. It also allows for the ability to print larger posters. Conversely the images are larger, and take more computational resources to process.

At the end of the day, it’s not about how many image megapixels a camera can produce, it’s more about clarity, composition, and of course the subject matter. Higher megapixels might be important for professional photographers or for people who focus on landscapes, but as amateurs, most of us should be more concerned with capturing the moment rather than getting going down the rabbit hole of pixel count.

Further reading:

The APS-C dilemma

/ spqr / Leave a comment

Should you buy a camera with an APS-C sensor, or a full-frame?

This argument has been going on for a number of years now, and still divides the photographic community. Is APS-C better than full-frame, or is it sub-optimal? Well, I think it’s all about perspective. APS-C, along with Micro-Four-Thirds are frequently viewed as mere crop-sensors, a designation that only exists because we perpetuate the falsehood that full-frame offers the “standard” sensor size. This stems from the fact that 36×24mm was the standard film size before digital cameras came along. As digital cameras evolved, “full-frame” became the name for the sensor size that matched a 35mm negative.

However we are at the point in time where each sensor size should be considered on its own merits, (and pitfalls) without unnecessary inference that it is a mere “stepping-stone” to a full-frame. Identifying an APS-C sensor, which has a size of 23.6×15.7mm, as “just a crop” sensor does not give the camera the kudos it deserves. The problem lies in every aspect of how these cameras relate to one another, but manifests itself best in lenses.

APS-C versus full-frame camera
The physical differences between APS-C and full-frame. The full-frame Leica SL3 is nearly twice the weight of the APS-C Fujifilm X-T50, and has a much bigger form factor.

Most APS-C sensors have a crop-factor of 1.5 (except Canon which is 1.6). This means lenses function a little differently than on full-frame lenses. Now a 50mm lens is always 50mm, regardless of the system it is associated with − it’s how that 50mm is interpreted in relation to the sensor that is important. For example a 50mm lens is a “normal” lens for a full-frame camera, while in APS-C land a normal is going to be a lens with a focal length of 33-35mm. A 50mm lens on an APS-C sensor will give a smaller picture than a full-frame, because well obviously the sensor is smaller. So an APS-C 50mm has the same effect as a 75mm lens on a full-frame camera in terms of what is in the picture.

Some basic visual comparisons of APS-C versus full-frame

There are obviously things that full-frame sensors do better, and things that APS-C format cameras do better. Image size is the first, which is purely the result of full-frame cameras having more photosites on their sensors. With the evolution of pixel-shifting technology this may be a mute-point as super-resolution images are already available on some systems. Full-frame cameras also tend to have better dynamic range and low-light performance. This is because photosites are often bigger on full-frame cameras, so they can collect more light and better differentiate between light intensities. This means they work better in low-light situations introducing less noise. But digital cameras rely on software to turn the data from photosites into the pixels in an image, and so as software improves, so too will things like noise suppression algorithms in APS-C.

How lenses function on APS-C and full-frame lenses

But not every full-frame has larger photosites. For example a Fuji X-H1 camera with a 24MP sensor has 6000×4000 photosites, with a photosite pitch of 3.88μm. The Sony a7CR has a 61MP sensor (9504×6336) with a pixel pitch of 3.73μm, which is actually smaller than that of the APS-C sensor. So more pixels, but perhaps a low-light performance that isn’t that much better. And what is anyone going to do with images 60MP in size? Post them on the web? I think not.

featureAPS-Cfull-frame
low-light performancegoodexcellent
depth of fielddeepermore shallow
lens availabilitylarge selectiongood selection, fewer third-party lenses
lens costaffordablemore expensive
portabilitylight, easy to carryheavy, bulky
dynamic rangeslightly reducedwider
applicationsstreet photography, sport, wildlife, travellow-light, studio, landscapes, portrait
camera body costtypically affordableusually expensive
wide angle lenses18-23mm28-35mm
normal lenses26-38mm40-58mm
A comparison of some of the characteristics of APS-C versus full-frame

Full-frame cameras, just like medium-format cameras are for people who need the things they provide – high resolution, low-light abilities, etc. Many people tend to correlate a full-frame camera with high quality because of its sensor size, but quality isn’t necessarily associated with high-resolution images. Yes, more data captured by a camera means more detail in an image, but that doesn’t automatically mean that APS-C sensors (or even MFT) are inferior.

Most non-professional photographers don’t need huge image sizes, just like they don’t need a Leica. APS-C cameras are considerably lighter, and more compact than their full-frame brethren. APS-C lens are also cheaper to purchase, because they are easier to build, and require less glass. In all likelihood there is also a broader ecosystem of third-party lenses for non-full-frame cameras as well, as they are cheaper to manufacture. Over time as newer sensors evolve, APS-C may be well positioned to take a more prominent role in the camera world.

Further reading: