A Filmmaker's Guide to Sensor Sizes and Formats

January 15, 2018
Tips and Techniques
Thank you for joining! Be on the look out for new content every week.
Oops! Something went wrong while submitting the form

As a cinematographer and owner of a rental house, I often run into questions regarding sensor sizes and formats. Questions like: "will this lens work with my camera?" Or, "will this lens cover 8K?" Or, "Will this lens cover full frame." Or, "what will this full frame lens look like on my camera's Super 35 sensor." My hope is that this article will help answer some of these questions and more.


MORE FORMATS THAN EVER

With the release of more high-end video cameras with larger sensors like the ARRI Alexa LF, Panavision DXL2, RED MONSTRO, and the Sony Venice, we have more choices than ever when it comes to formats and lens options. However, it's important to know the differences as well as what results should be expected before selecting your project’s sensor size and lenses.

The new ARRI Alexa LF full-frame camera.

There are many formats to choose from. Digital or film? If film, which format? If digital, what camera and sensor size? There is no single "best" format that is right for every project or situation. Motion picture film formats are fairly straight forward: 8mm, 16mm, 35mm, 65mm, 70mm. These names are based on measurements of the physical size of film used to capture the images. Today with digital cameras, we have more formats than ever: 1/2”, 2/3", Micro Four Thirds, Super 35, DX, APS-C, Full Frame, Vista Vision, etc. All these formats need lenses, and the lens market is bigger and more confusing than ever. Also, not all lenses work with all formats, and not all formats might be the right choice for every project.


SENSOR SIZE

Sensor size is the physical size (area, not number of photosites or pixels) of a camera’s image sensor, usually measured in mm width x height. A Full Frame digital sensor like the ones found in a Canon 5D, Sony a7S II  or the ARRI Alexa LF, as well as traditional 35mm still photography film, all have areas that measure roughly 36x24mm. For the remainder of this article this format will be referred to as “Full Frame.” In the chart below the red rectangles represent the relative physical size of many common cameras and formats as well as the number of pixels represented in “HD” or “Ks” (the “HD” or “K” number of course does not apply to film).

RESOLUTION & PHOTOSITE COUNT

For film, resolution is the term used to describe how much detail can be resolved usually measured in line pairs (lines per mm or lines per inch). For digital sensors, we use the term Number of Recorded Photosites or Number of Effective Photosites. This is the number of individual photosites (often referred to as pixels) on a given sensor that contribute to the final image. This number can be measured in horizontal x vertical. For instance 1920 x 1080 is the standard pixel count for HD cameras. We are now abbreviating these numbers by using "Ks" (2K, 4K, 6K, 8K etc.), each K standing for 1000 photosites in the horizontal axis. For instance 4K represents roughly 4,000 photosites in the horizontal axis of a 4K image sensor. I say “roughly” since a common “4K” pixel count is actually 3840 x 2160 photosites (UHD), and there are multiple photosite counts that are accepted as “4K.” In still photography digital cameras, photosites count is often measured in “megapixels.” One megapixel = 1 million “pixels.” The Canon 5D Mark IV is 30.4 megapixels, which is about 30,400,000 photosites, which is 6720 x 4480 photosites, and in theory could also be called 6.7K.


FIELD OF VIEW 

Field of view or more accurately stated, "horizontal angle of view" is a measurement in degrees along the horizontal axis to determine just how much of the world you will see when looking through a lens. For example an ultra-wide-angle lens like a 16mm fisheye lens (one made to cover the Full Frame format) will have a 180° horizontal field of view on a camera with a Full Frame sensor. If you placed your camera and 16mm fisheye lens in a small square room, standing close to one wall, in the center of that wall, your camera would see the wall across from you as well as the walls to your left and to your right. A super telephoto lens like a 500mm will have a very narrow field of view of only 5° on a Full Frame camera, and you would only see a small piece of the wall directly across from you.

Field of view is determined by the focal length of the lens you are using as well as the sensor size of your camera. A particular lens will give different fields of view when paired with different sized sensors. For instance, a 50mm lens on a Full Frame camera will give you a field of view of about 46°, but on the smaller sensor of an APS-C camera, the same 50mm lens will give you a 31° field of view, showing you a narrower view of the world. The difference in field of view is extreme, if you are using a camera with an even smaller Super 16 size sensor like the one in the original Blackmagic Pocket Cinema Camera or a Digital Bolex D16, versus the larger Full Frame sensor of a camera like the Sony a7S II. In the images below, you can see that using a 50mm lens on a Super-16 sized sensor gives you an extreme-close-up, but with a Full Frame sensor, it's a medium-close-up, even though the distance from camera to subject stays exactly the same. This effect of a single focal length producing different fields of view with different sensor sizes is sometimes referred to as "crop factor.”


CROP FACTOR

Crop factor is a ratio of one camera’s sensor size related to another’s camera’s sensor of a different size.  Crop factor is most commonly used to compare the sensors of Full Frame cameras to smaller formats like APS-H, APS-C, Nikon’s DX format, etc. Crop factor is often used when shooting on these smaller formats to find the appropriate focal length to give you the same field of view you can achieve with a particular lens in Full Frame.

For example, photographers and videographers accustomed to the field of view they see when looking through a 50mm lens mounted on a Full Frame Nikon D810 ("FX" format in Nikon speak) might wonder what equivalent focal length will give them the same field of view on a Nikon D500 which has a smaller "DX" format sensor (which is similar in size to APS-C and Super 35 sized sensors). There is a formula to figure it out. To find the equivalent focal length to give you the same field of view when using the smaller sensor of Nikon DX cameras, you use a crop factor of 1.5x. To find right lens, you divide 50mm by 1.5, which gives you about 33.3mm. Since you probably don't have a 33.3mm lens, using a 35mm prime lens on a DX format Nikon camera like the D500 will give you roughly the same field of view as a 50mm lens on your Full Frame Nikon D810. If you want to go the other direction, and see what focal length you need on your Nikon D810 to give you the same field of when you use your Nikon 24mm f1.4 on your Nikon D500, you need to now multiply 24mm x 1.5, which gives you 36mm. Since you probably don’t have a 36mm lens, to get roughly the same field of view on your Nikon D810, use a 35mm lens. "Crop Factor" is really just another way to help understand and work with the changes in "Field of View" caused by using cameras with different sensor sizes.

IMAGE CIRCLE

A lens' image circle refers to the light (the image) projected out of the rear side of a lens, for our purposes onto a camera’s sensor or film plane. A lens projects a circular image, not rectangular. We simply crop the circular image into rectangular shapes with various aspect ratios. The diameter of the image circle is measured in mm. It is very helpful to know a lens’ image circle, because if you know the image circle, you know how large of a sensor a lens can cover. For instance, in order for a lens to cover the entire sensor of a Full Frame camera like a Canon 5D or the Sony Venice, it would need an image circle with a diameter of 43mm. Not every manufacturer makes this information easily accessible, especially with older lenses, so research or testing is needed to verify if a lens can completely illuminate or “cover” certain sensors.

In the images below it's easy to observe this lens' complete image circle. This lens was originally designed to cover Super-16mm film. When used on a camera with a Full Frame sensor we are able to see the lens' entire image circle as well as the rectangular cropped portion, which will end up being the final image.


LENS COVERAGE

A lens' image circle determines what sensors it can cover. With all the different formats and lenses to choose from, it can get confusing. A lens can "cover" the sensor size and film format it was designed to cover, as well as any formats smaller than that. However most lenses can't cover formats larger than the one they were engineered to cover (there are a few exceptions including some high-speed lenses and often prime lenses 50mm and longer as a result of their specific optical designs). If a lens was designed for Super-35 and APS-C (which are close in size), it will successfully cover the sensors of Super-35, APS-C as well as smaller sensors like Micro Four Thirds and Super-16. However, it more than likely will not cover larger sensors like the ones in Full Frame cameras, and even larger sensors like the one in the ARRI Alexa 65.

To add to the confusion it has become popular to “measure” a lens' image circle in "Ks." To say a lens “covers” 4K, 6K, 8K, is flawed. Since Ks refer to the number of photosites on a sensor not the physical size of an image sensor, using Ks in this way is problematic. When RED went from their original 4K sensor, to a 5K sensor and then a 6K sensor, there was a consistent relationship between photosite count and sensor size. So as the number of photosites grew, so did the relative size of the sensor. Therefore this method of discussing a lens’ coverage could in theory make sense for RED cameras of that time, but it couldn’t be applied to other cameras from other manufacturers. After RED released their Helium sensor, it couldn't even be applied to RED cameras any more! If you look at the comparisons below, you can see how there is no relationship at all between "Ks" (photosite count) and sensor size (physical area).

The reason for this lack of a constant in the relationship between photosite count and sensor size, is that the physical size of the photosites varies from sensor to sensor. For example the 8K Sensor in a RED Helium is much smaller than the 8K Sensor in a RED Monstro despite having the same number of photosites.  The physical size of the individual photosites in the Helium are smaller than the individual photosites in the Monstro.

"A 50...is a 50...is a 50..."

All lenses of the same focal length, regardless of what format they were designed to cover, "see" the world in the same way, in regards to magnification and compression, however the size of their image circles can vary. By that I mean a 50mm lens designed to cover a camera that shoots 16mm film, "sees" the world in the same way as a 50mm lens designed to cover a Full Frame digital camera. The difference between the two is how much of the world they are able to see and project onto a camera's sensor. In the images below, the same scene was shot 3 times with the same camera, with the same Full Frame sensor. The subject stayed in the same place. The camera never moved (always 4' 6" from her eyes to the sensor) and the aperture stayed at T4 for all three lenses.

As you can see, the magnification of the scene is the same. The size of the woman's face, the geometry of the room, and the depth of field are the same with all three 50mm lenses. The only difference is each lens' image circle and therefore how much of the room you are able to see.

"I NEED 4K AND FULL FRAME..."

The increasing number of sensor sizes and lens options, has made things more complicated than ever especially with all the lens adapters available and cameras with interchangeable lens mounts. There are also a lot of companies rehousing lenses, giving filmmakers the opportunity to use lenses that were not originally designed for motion picture use. With so many options it can be overwhelming when gearing up for a project. A director might say, “I really want to shoot this project in 4K and Full Frame.” So the DP lines up a rental for a Sony a7S II, which shoots Full Frame at 4K. Then the producer says, “I was able to get a great deal on ARRI/Zeiss Ultra Primes!” That’s great news for their budget, but if the DP doesn’t know the right information about those lenses, or doesn’t do research or doesn’t test them out before the shoot, he or she will find out the hard way that those lenses were not designed to cover Full Frame, and the DP will have to use the "center crop" feature on the a7S II, and now that project has to be recorded at Super 35, 1920 x 1080, which is not what the director wanted.

Canon K35 full frame lenses

In that scenario, they needed to track down lenses designed to cover Full Frame or larger sensors, and there are more lens options than ever to cover these big sensors. For example Leitz has their Thalia line of primes based on Leica medium format still photography lenses originally designed to cover a huge 60mm x 60mm film format. These lenses can be used on the giant image sensor found in the ARRI Alexa 65. That means these lenses will also cover any sensor or film stock that is smaller than the Alexa 65 (which is just about every current motion picture camera and format). If you happen to own Super 16 prime lenses you will be able to use them for Super 16mm film and similarly sized digital sensors, and any format smaller than that ( 8mm film, 2/3”, or on many of the “cropped sensor” formats available on cameras like the ARRI Amira, Canon C300 MKII, Sony F5, F55, FS7, Panasonic Varicam LT and pretty much any RED camera). The ARRI/Zeiss Ultra Primes mentioned earlier in the hypothetical scenario are in the middle of the pack as far as sensor coverage goes. They were designed to cover Super-35mm film, so they can’t cover larger formats like Full Frame or Alexa 65, but they can be used on any smaller formats, like Micro Four Thirds, Super-16, 2/3” etc.

In the examples below, you can see how three 25mm lenses designed for different formats, shot on a camera with a Super 16 size sensor, at the same T stop will produce images that have the same field of view, the same magnification, same compression, and the same depth of field. Therefore a 25...is a 25...is a 25. These three 25mm lenses were designed to cover different sensor sizes, but because they were all designed to at least cover a Super 16 size sensor, they give us virtually identical shots.

ARE LARGER-FORMAT LENSES THE ONLY LENSES YOU'LL EVER NEED?

One could come to the conclusion that the only lenses you will ever need are some rehoused medium format lenses because they will work on any current camera format, right? It’s not that simple. For instance the 24mm in the Thalia lens set mentioned earlier has a maximum aperture of T3.6, which compared to prime lenses designed for Super-35 or even Full Frame (which can be as fast as T1.5, T1.4, or even T1.3), is a somewhat slow lens requiring more light, and delivering images with more depth of field compared to a faster lens of the same focal length. What if you have a project that will be shot on an Ursa Mini Pro, shot mostly at night that relies upon practical lighting or available light? Since the Ursa Mini Pro performs better at lower ISOs, you decide that you need lenses that can open up to T2 at least, but T1.4 would be even better. If that’s the case, larger format lenses like the Thalia primes might not be the best option, and it might be better to get something like Zeiss Super Speeds, which have a maximum T stop of T1.3. Both the Thalia 24mm and the Zeiss Super Speed 25mm will give you almost identical fields of view, but the Zeiss Super Speed requires less light, and has the ability to give you quite shallow depth of field should you choose to open up to wider apertures.

Leica Thalia lenses

It’s also important to note, the widest lens in the Thalia set is 24mm. When shooting on Super 35 or smaller formats, it’s very common to shoot focal lengths wider than 24mm to achieve a wider field of view. So if you only had that set of Thalias, you couldn’t shoot any ultra-wide shots on cameras with smaller sensors. Leitz didn’t need to design any focal lengths wider than 24mm for the Thalia lens line because on the formats they were designed for, a 24mm lens has an extremely wide field of view.

A 24mm T3.6, shot at T3.6 on an Alexa 65 can provide a wide field of view that still has shallow depth of field. When shooting on Super 35, to get a shot with the same field of view and depth of field, you’d need a lens that was roughly 13.5mm and T1.8. So depending on the look you’re going for and the size of the sensor you are using, the Thalia 24mm T3.6 may leave you looking for wider focal lengths with faster T stops. So as amazing and beautiful as the Leitz Thalia lenses are on an Alexa 65, they might not be the best choice for projects using smaller sensors.


THE ULTIMATE LENS SET...

Then why doesn't someone just design large format lenses that cover the huge Alexa 65 sensor, have a maximum aperture of T1.3, and make focal lengths wider than 24mm? Then you could buy one set of primes that you could use on every job! Physics is one of the big reasons why it’s difficult to make lenses like that. For example an 18mm lens that could cover the Alexa 65's sensor and was T1.3 would be very large, and very heavy. It would also be very difficult to actually design and build. The individual glass elements would be enormous and would be difficult to match the tolerances achievable with smaller format high-speed lenses.


Another challenge in designing these “ultimate lenses” would be cost. The cost to design and manufacture such lenses would make the resulting lenses so expensive, no buyer outside of NASA could afford to own them. There is also practicality to think about. If this hypothetical lens set was somehow built, it would be difficult in certain situations to shoot at T1.3, especially with medium and longer focal lengths. If you were using a theoretical 85mm T1.3 (designed to cover the Alexa 65 sensor) on the Alexa 65 at T1.3, the depth of field would be so shallow, you would only be able to keep a thin sliver of an object in focus. This effect could be interesting for specific styles and scenes, but for many shooting situations, such extremely shallow depth of field could be distracting, pulling focus would be very difficult for your 1st AC. The DP might end up stopping the lens down to T2.8 or T4 anyway.


ARRI Alexa 65 on the set of Rogue One: A Star Wars Story. Courtesy of StarWarsUnderworld.com

Another challenge in designing these “ultimate lenses” would be cost. The cost to design and manufacture such lenses would make the resulting lenses so expensive, no buyer outside of NASA could afford to own them. There is also practicality to think about. If this hypothetical lens set was somehow built, it would be difficult in certain situations to shoot at T1.3, especially with medium and longer focal lengths. If you were using a theoretical 85mm T1.3 (designed to cover the Alexa 65 sensor) on the Alexa 65 at T1.3, the depth of field would be so shallow, you would only be able to keep a thin sliver of an object in focus. This effect could be interesting for specific styles and scenes, but for many shooting situations, such extremely shallow depth of field could be distracting, pulling focus would be very difficult for your 1st AC. The DP might end up stopping the lens down to T2.8 or T4 anyway.

DOES SIZE MATTER?

As exaggerated as the above example is, it’s a good demonstration of why we benefit from having different sensor sizes and film formats. For ROGUE ONE: A STAR WARS STORY, shooting on Alexa 65 with big, heavy, vintage Panavision Ultra 70 anamorphic prime lenses was a great recipe for achieving epic, large-format images perfect for the big screen. That production also had the budget and crew to support using gear that is bigger, heavier, slower to set-up, and more expensive. However for a run-and-gun documentary, shooting on a smaller format like the S16 HD crop on an Alexa Mini or a 2K Center Crop on a Sony F55, shooting 3.5K on a RED Helium or using a Blackmagic Pocket Cinema Camera, can result in a smaller, lighter camera, smaller lenses, typically faster lenses that need less light, cheaper rentals, and possibly even a smaller crew, needing less set-up time.

Blackmagic Pocket Cinema Camera

Even smaller than Super 16, the 2/3” sensors in the cameras used for most live sports, make it practical to build a lens like the Fujinon XA101x8.9BESM/PF that can go as wide as 8.9mm and as long as 900mm (1800mm with its built-in doubler engaged!) and has a maximum aperture of f1.7…wow. To build a lens that could cover that massive 101x zoom range and cover a much larger sensor like the ones in a Canon C300 or Sony FS7 at f1.7 would be the size of a Civil War era cannon and probably weigh as much. The Fujinon mentioned above costs $233,490, so if you needed to build its Super 35 bigger brother what would it cost, half a million dollars? To build one that covered an even larger format like Full Frame or the Alexa 65, would probably be the size of a minivan and cost 5 million dollars!

AESTHETIC DIFFERENCES

The physical and financial reasons for having all these different sensor sizes might be more obvious than the aesthetic differences that go along with them. One of the simplest differences to understand is magnification. When we finally show our recorded images to an audience, they are magnified from their original size (the size of the sensor or of the film). Whether the images are being watched on a cell phone or an IMAX theater screen, the original recorded images will be magnified. With film, when projected onto let’s say a 60-foot-wide theater screen, a 16mm print has to be magnified more than a 35mm print to illuminate that 60-foot screen. Also, since the 35mm print is able to resolve more detail (resolution), the resulting projected image will appear sharper and cleaner. If both of these hypothetical films were shot using the same film stock, they would both have the same size film grain. If you then projected each film onto the same size screen, although the film grain would be the same size in both stocks, due to its smaller physical size, the 16mm print will need to be magnified more than the 35mm print for projection, and so the size of the film grain would also be magnified more. So the 16mm film would appear more grainy and textured, while the 35mm film would look cleaner due to its grain appearing smaller since it’s being magnified less.

It’s the same in digital. If you shot two theoretical projects, one on an ARRI Alexa Mini shot at 3.2K, and one shot on an ARRI Alexa 65 in 6K, (which both camera’s sensors have the same size photosites), keeping the camera’s settings the same (aspect ratio, ISO, shutter, codec, etc), and projected the two films onto the same size screen, like with film, you will see the same aesthetic differences. Because the Alexa 65 sensor is so much larger than the Alexa Mini sensor, the Alexa 65’s images would need to be magnified less than the Alexa Mini’s images, therefore the appearance of grain (or “noise” in the case of digital) would be less in the Alexa 65. Also, since the Alexa 65 has a much higher number of photosites, the Alexa 65’s image should appear to be cleaner, and probably sharper, with less texture.

The other big difference between smaller formats and larger formats can be the depth of field you are able to achieve. I say “can be” because sensor size alone does not dictate depth of field.  Depth of field is the distance between the nearest and farthest objects that are in acceptably sharp focus in an image. Depth of field is affected by a lens’ f stop, its focal length, and the distance from the camera to the subject. The wider the f stop, the less depth of field. A 50mm lens at f1.4 has far less depth of field than a 50mm lens at f4. The longer the focal length of the lens, the less depth of field it will have compared to a wider lens. If the distance from the camera's sensor to the subject is the same in both shots, an image captured with an 85mm lens shot at f2.8 will have less depth of field than an image captured with a 25mm lens shot at f2.8.

If you compare images shot at the same f stop, with the same field of view on cameras with different sensor sizes, the smaller the sensor, the more depth of field will be present in the images, and therefore the larger the sensor, the less depth of field will be present in the captured images. Therefore in smaller formats like ⅔”, and Super 16, the background and foreground stay in sharper focus relative to the subject. In larger formats like Full Frame and Alexa 65, depth of field is much less (or shallower) compared to smaller formats. Foregrounds and backgrounds appear more out-of-focus relative to the subject, which separates the subject from the background, and gives the image an overall more three-dimensional feel. However, this effect can be compensated by changing your f stop. If you want to shoot on smaller formats, but also want shallow depth of field, you will need extremely fast lenses, which is why most prime lenses designed for smaller formats are fast, often as fast as T1.3. Conversely, if you want to shoot on a large format camera, but do not want the look of shallow depth of field, you can always stop the lens down to f5.6 or f8.  The problem with achieving extremely shallow depth of field on smaller formats, is that it’s difficult to design lenses that are fast enough to give you the same shallow depth of field you are able to achieve on larger sensors, especially on images captured with wide-angle lenses.

For example, the difference in depth of field between images captured by cameras with Full Frame sensors and Super 35 format sensors, using lenses that will give you the same field of view, is about 1 full stop. For instance, if you capture an image with an ARRI Alexa Mini using a 50mm lens at f1.4, to get an image with the same field of view and the same depth of field on an ARRI Alexa LF, you’d need to use a 70mm lens shot at f2. If you want to go the other direction, and capture a third image with the same field of view, and the same depth of field, but this time using the ARRI Alexa Mini’s S16 HD format (Super 16), you would need to use a lens that was 26mm f0.7. As you can see, it is theoretically possible to achieve extremely shallow depth of field in smaller formats, but it might be difficult or impossible to find lenses that have such huge maximum apertures. There aren’t too many f0.7 lenses out there.

CONCLUSION

There are more choices than ever for acquisition format and more sensor sizes are popping up every year. New lenses seem to be released monthly. With all these options at our fingertips, it means we have so many creative possibilities. However, more than ever we need to make sure to consider all the variables before deciding which format is best for a given project. What’s the style of the project? Do you need a physically small camera and lenses? Will you be in a low-light situation? Do you need primes, zooms, or both? Do you want more or less depth of field? Do you have limited set-up time? What’s your budget? You also need to know which lenses are compatible with which cameras and formats. The last thing you want is to choose the wrong equipment, which leaves you fighting with your gear instead of making amazing images. There are more tools than ever available to filmmakers, and it’s up to you to figure out which ones are the best for your project.


Thank you for joining! Be on the look out for new content every week.
Oops! Something went wrong while submitting the form
Mark LaFleur

Mark LaFleur is a Los Angeles based cinematographer and lens addict.

Related Posts