The Fascination with Full Frame
Last updated on May 19, 2014 (published May 1, 2014)
- Full Frame Sensors
- APS-H/C Sensors
- The "Crop Factor"
- Depth of Field
- Image Quality
- Pixel Size
- Resolution and Megapixels
- Leica's Conundrum
Ask pretty much any digital photographer what "full frame" is and they'll tell you it's sensor based on a full 35mm frame's dimensions, which measures 24x36mm. Back in the heyday of film, there were fewer options... Especially when the 35mm format came about 101+ years ago in 1913 - though it wasn't really popularized until a dozen years later, in 1925 when Oskar Barnack used this format for his Ur Leica. But in today's digital cameras it seems like there are countless variations of sensors sizes (e.g. APS-C, APS-H, m4/3, etc.).
So what is it about full frame that fascinates so many shooters? Does the sensor size really matter? Who cares if the sensor is smaller, if it has more megapixels in resolution? Surely that means it's better. Well, yes and no. In this article we'll take a look at some of the considerations of sensor size and why full frame seems to be the Holy Grail of digital photography. It's not meant to be a lengthy, in-depth explanation. Quite frankly, it's just not that exciting.
Full Frame Sensors
It's important to point out that full frame has, until recently, been solely the realm of DSLRs. Why is that? For the most part, camera body size and the fact that most lenses on the market at the beginning of digital photography were designed for 35mm film cameras. You needed a camera that was large enough to hold such a sensor (which is physically larger) and more importantly, work with the existing lenses. Lenses designed for 35mm SLRs have particular dimensions that demand a larger body - because of the flange focal distance (the depth from the lens mount to the film/sensor plane). SLRs required this extra depth because of the mirror, and lenses had to stay clear of it. As a point of contrast, the Leica M rangefinder lacks a mirror (relying on the optical rangefinder instead) which makes the overall camera depth much thinner. Of course, this presented a unique conundrum for Leica which we'll get into below.
These days, there are several technologies and developments that have eliminated the need for a traditional SLR-based design; namely the lack of a mirror. So-called "EVIL" (Electronic Viewfinder with Interchangeable lenses) or "mirrorless" cameras, which as the name suggests, use an electronic viewfinder instead of the mirror. Thus the flange focal distance can be reduced... Which of course also mandated that new lenses be designed, as they had to project their image circle on a closer plane. While this allows for smaller camera bodies and lenses, it also means that you needed to invest in all new lenses. On top of this, there are other issues that come about as a result of varying sensor sizes. The geometry of everything is inter-related.
Regardless, full frame cameras are still in the minority out of all the digital cameras on the market. This adds a bit to the "Holy Grail" mystique to the notion of having a full frame camera. Fact is, there just aren't that many when compared to the myriad of other cameras now available - though they have been increasing in number lately with the new demand for full frame among enthusiasts.
The most common sensor size therefore isn't actually full frame, but rather a bit smaller. One popular choice is known as "APS-C" (or its close and slightly larger cousin, APS-H). These sensors, as far as Leica cameras go, measure 16x24mm (and 18x27mm, respectively) and are often referred to as "half frame." Why use a smaller sensor? Early on the reason was primarily technology. Full frame sensors just weren't possible. Once they were, the cost was still prohibitively expensive. Since they also demanded a larger camera in the early days, manufacturers couldn't put them in compact P&S cameras - not without redesigned or custom lenses, which came about later. For example, Canon's line of EF-S lenses.
For their initial digital M, the M8 and M8.2, Leica utilized a sensor with an APS-H size. In the newer X and T line of cameras, Leica went with an APS-C sized sensor.
The "Crop Factor"
Using smaller sensors with the industry standard 35mm lenses created what was to become known as the "crop factor." The smaller the sensor, the higher the crop factor. The reason for this is quite simple. A 35mm lens is designed to project an image circle at a specific flange focal distance, to cover a complete 35mm frame of film. A smaller sensor simply can't capture all of it, thus cropping out the image circle beyond the size of the sensor itself. Thankfully, there was a loose (but still widely varying) standard among manufacturers. Canon, for example, had a crop factor of 1.62x with APS-C and 1.3x with APS-H whereas Nikon had 1.52x and 1.57x, while other manufacturers had yet different factors. The Leica M8 and M8.2, with their APS-H sensors, had a crop factor of 1.33x.
So what does this crop factor mean to the photographer? In a nutshell, it means that your 50mm lens isn't really a 50mm lens when it comes to the angle of view it captures. Sure, it's still a 50mm lens in optical length - but when you apply the crop factor by multiplying the focal length by its value, you invariably end up with a longer effective focal length. For example, on an M8 with a crop factor of 1.33x with a 50mm lens, what you ended up with was essentially an 67mm lens (50 x 1.33 = 66.50mm).
This increase in effective focal length has an interesting side effect. The speed of the lens is not affected. So if you have a Noctilux-M 50mm f/0.95 ASPH lens, it becomes a Noctilux-M 67mm f/0.95 ASPH! This effect is more exciting at longer focal lengths... A Tele-Elmar 135mm f/4 becomes quite a telephoto lens at 180mm.
This effect also creates an interesting problem at the opposite end. Wide angles aren't quite as wide as they used to be with a crop factor. In order to have an effective focal length of 21mm (a common street shooting focal length) you'd have to use a 15mm or 18mm lens. Early on in digital Leica photography, this presented quite a dilemma as the lens options that we enjoy today didn't exist back then... Such as the Voigtländer Super-Wide Heliar 15mm f/4.5 or Zeiss Biogon 4/18 ZM. But these days, this isn't much of an issue anymore. Especially when it comes to lenses designed specifically for smaller format sensors.
Full frame sensors don't have a crop factor, simply because there's no cropping. If you expressed it as a value, it would just be 1x.
So? Well, for starters it meant that lenses you were used to working with were entirely different. You had to think differently. In order to get the view of a 50mm lens that you used to enjoy meant now shooting with a 35mm lens - and even then, the math doesn't quite work out to exactly 50mm. One easy way to wrap your head around this was to use the "next" focal length wider than what you needed. For a 50mm you use a 35mm, for a 35mm you use a 28mm, for a 28mm you use a 21mm, etc. But the fun doesn't end there.
On the M8 and M8.2, Leica conveniently adjusted the framelines in the viewfinder to correspond to these "adjusted" focal lengths. So the 50/75mm frame lines for example, show what you'd actually get with the crop factor applied.
Depth of Field
After you get used to the little bit of mental arithmetic required for determining focal length, it boils down to the angle of view for a given lens. Since a normal 50mm lens is now more like an 67mm, that means you now have to step back from your subject to frame it the same way as you would with a "true" 50mm. You might ask, "so what's the problem with that?" We'll give you that. However, because the distance between the camera and your subject has now increased, this effectively increase the depth of field (DoF) in your image. You would have to open up your lens to maintain a correspondingly thinner DoF, and eventually you run out of aperture range. One way around this is to actually use a wider lens, such as a 35mm, and get closer to your subject. Either way, it's all fun and games until the light levels drop. Essentially, you can only open up so much to reduce DoF and in the end you typically have less than with a full frame sensor. The smaller the sensor, the more this effect is exaggerated. Take a cell phone photo for example, which uses a positively tiny sensor - notice how pretty much everything is in focus? So much so that Google recently updated its Camera App to create "fokeh" or, fake bokeh.
Related to apertures is the diffraction limit. As you may know, apertures smaller than the given limit will decrease image quality because of diffraction. Smaller sensors effectively lower the limit, meaning that with such cameras you might be diffraction limited by f11, whereas on full frame you can go up to f/13, for example. This further decreases your DoF control. Less out of focus on one end, less in focus on the other.
Speaking of image quality, another point in favor of smaller (than full frame) sensors is when it comes to lenses. If you've ever read an MTF chart (which graphs technical image qualities over the dimensions of the frame) you've no doubt noticed that technical image quality degrades as you approach the frame borders and especially corners. Qualities such as field flatness (and thus sharpness) as well as vignetting. If your lenses aren't up to modern standards, full frame sensors will expose these shortcomings in all their glory. However, with smaller sensors, these trouble areas are essentially cropped out. They make your lenses "better." Or the corollary, with full frame cameras you need the best lenses to extract the last bit of ultimate image quality. Thankfully, the newest and current line of Leica M lenses are frightfully good.
As we alluded to earlier, just because a smaller sensor has more megapixels than a larger one - does not mean it is "better." From a resolution standpoint, sure. But the pixels (individual photosites on the sensor) get smaller the higher the resolution for a given sensor size. This has a negative effect on two key aspects; light sensitivity (ISO) and as a result, noise. As we know, higher ISOs increase noise levels. Inversely, the larger the pixel, the more sensitive it is to light and the cleaner the output. They also tend to have nicer, more saturated colors. Some call these larger pixels "fat pixels." So from an IQ standpoint rather than resolution, larger sensors with less megapixels are actually better than smaller sensors with higher megapixels.
Pixel size also affects dynamic range. With larger pixels, you can capture a wider range of values from black to white without again, increasing noise.
An interesting thing to note as far as Leica is concerned, is that they kept the pixel size the same (6.8 µm) between the M8 and M9 series of cameras despite the higher megapixel count because the sensor size also increased. Therefore, while the resolution stayed the same, you got more pixels (really just a larger image). Because of this wizardry, a 10MP image from the M8 technically has no greater detail than an 18MP M9!
Resolution and Megapixels
But an APS-C sensor doesn't have 24MP or 36MP! No, not yet. Some day, very soon, they will. So? There are few of us that really need such a resolution. Do you make very large prints of your photos? Chances are, you don't - and if you did, you could probably count them with your fingers. Sure, it sounds nice on paper to have such a high-resolution megapixel monster. If nothing else, "just in case" you grab that once in a lifetime shot that you want to print mural size and make (actual) wallpaper out of. But for most shooters, 16-18MP represents the sweet spot in resolution. At this size, storage cards hold a lot of photos, demands on your hard drive storage are minimal and the files quick to process on even mediocre hardware. Print sizes, assuming good technique and equipment can be quite large, even with "only" 16MP. The "Megapixel War" essentially ended a couple of years ago - and yet today we're still seeing 16MP cameras being released, such as the Leica T or Fujifilm X series of cameras. There's a reason for that, and not just because they happen to feature an APS-C sensor. Don't get caught up in the megapixel game.
As mentioned in the previous section, the higher the resolution for a given sensor size, the smaller the pixels have to be. And that's not necessarily a good thing, as we know. There's a point of diminishing returns where more pixels just degrade your image quality. Full frame sensors, having more real estate, can go up to 24MP or 36MP before they see the negative effects that a smaller sensor will. In the end, the actual resolution of your image might be exactly the same (just larger)!
Food for thought: a double page spread in National Geographic is 14" x 10 1/4" which at 300dpi equates to 4200px by 3075px - a measly 12.915MP! With a 16MP camera you even have room to crop/straighten.
The Leica Conundrum
Finally, as if all these things weren't enough conspiring against us, Leica had one more issue to deal with in developing a digital rangefinder. And this was a big hurdle to overcome at the time, make no mistake. The problem with a digital M implementation is due to one of its biggest benefits; a short flange focal distance. What makes the camera (with lens) so compact - fights against it when implementing a digital sensor. Simply, because the rear element of the lens is so much closer to the sensor (again, because M cameras don't have mirrors like SLRs) means that the light rays exiting the lens are at a much more oblique angle in order to cover the same 35mm image circle. On film this wasn't much of an issue. A light sensitive crystal in the emulsion didn't much care about the angle of light rays. A digital sensor however, is much more demanding. Each photosite can be thought of as a "well" that photons must fall into to be captured. This is what they look like on the Leica M (Type 240) camera:
So Leica had to develop microlenses that not only angled to account for this - they had to do so increasingly towards the edges and corners of the frame! At the time the M8 was developed before its 2006 release, this was rather new territory. A full frame sensor at that time was out of the question and Leica themselves literally called it "impossible." The cool thing about technology is that it both improves and becomes cheaper as time progresses. What was impossible earlier is later possible, and begat the full frame sensored M9 and all the full frame cameras that followed.
What about cameras like the X2, X Vario and T that use a smaller APS-C sensor - isn't this a compromise? Well, yes. While this article seems to paint a grim picture of less than full frame sensors, everything in photography is a compromise of some sort. You have to decide what's most important to you. Do you want a small, portable system? Perhaps full frame isn't the answer, then. Is there anything really wrong and with an APS-C sensor? Absolutely not... Especially if you don't have so-called "legacy lenses" to deal with. These new cameras all have either integral lenses (in the case of the X cameras) or specially designed ones taking the APS-C sensor into account (in the case of the T camera). Like we said, technology gets better and better, and today's APS-C sensors, from an image quality (e.g. resolution, light sensitivity and noise) standpoint have also. Rest assured, the sensor found in these cameras (they're all the same) delivers outstanding results.
Perhaps now you can see why having a full frame sensor is desirable. Larger sensors allow for more megapixels but more importantly, larger pixels for increased light sensitivity, lower noise, more saturated colors and higher dynamic range. For a couple of reasons, they also allow for a greater range of DoF control. Finally, there's no mental math required when using 60 years worth of M lenses; a 50mm lens is a 50mm lens.
Is it worthwhile to chase after a full frame camera? Depends on your needs and wants more than anything else, rather than technical reasons. That is to say, a crop sensor camera can make just as beautiful an image as a full frame camera in most situations.