There’s an idea in photography that cameras
with different sized sensors have what we call ‘crop factors’. A large format Alexa 65 has
a crop factor of 0,56x compared to a Super 35 camera. A 90mm lens multiplied by 0,56 is 50mm.
Therefore, many people say that using a 50mm lens on this camera is going to look exactly
the same as using a 90mm lens on this camera. The truth is that this isn’t exactly 100%
correct - for quite an important reason. So, let’s go over a more detailed
explanation
on what ‘crop factor’ is, how it works and the big misconception about it. This video is brought to you by MUBI, a
streaming service that screens beautiful cinema from all around the world. Get
a free month at mubi.com/indepthcine. As photochemical film photography emerged
and cinema cameras were created, there was a push to create film with a standardised
size - that could be used across different cameras from different manufacturers and
be developed by different laboratories around the wo
rld. That film had a total width of
approximately 35mm and therefore was called 35mm. When digital cinema cameras
started getting manufactured, they replaced film with photosensitive sensors that stuck to the approximate size
of film’s 35mm 4-perf capture area. However, along the way some
other more niche formats emerged: from smaller 16mm film that was
a cheaper alternative, to large format 65mm which maximised the resolution
and quality of movies at a higher cost, to tiny ⅔” video chips
from early camcorders to
smaller micro four thirds photography sensors. The issue is that when you put the same lens on
two cameras with different sensor sizes they will have a field of view that is different, where
one image looks wider and one looks tighter. So, for prospective camera buyers or renters to
get a sense of the field of view each camera would have, many manufacturers started to publish what
they called a ‘crop factor’ to determine this. This means you take your lens’ focal le
ngth
- for example a 35mm lens - and multiply it by the crop factor of the camera - such as
2x - to arrive at a new focal length number, 70mm. This means that on this smaller sensor
your 35mm lens will have approximately the same field of view or magnification
as a 70mm lens on a Super35 sensor. Since Super 35 sensors are
considered the standard size, this has a crop factor of 1x. Camera sensors
larger than 35mm would have a crop factor of less than 1x and sensors that are smaller than
35
mm would have a crop factor of more than 1x. So where does the myth part come in? Well,
the issue is that many people interpret crop factors as saying that shooting with
a 70mm lens on a Super 35 sensor is exactly the same as shooting with a 35mm
lens on a smaller sensor with a 2x crop. What’s important to note is that while the level
of magnification of the image may be the same, there are still a bunch of
other characteristics that lenses have that will make images different
depending on
what focal length is chosen. So what we should say is that a 70mm lens on
a Super 35 sensor has approximately the same field of view of a 35mm lens on a smaller
sensor. We shouldn’t say that a 70mm lens on this camera is exactly the same as a
35mm lens on this camera in every way - as different focal lengths come with other secondary
characteristics beyond just their field of view. Rather than different sensors magnifying
or zooming out on the width of what we see it’s better to think abou
t it in
different terms. If you put the same lens on two different cameras with a
larger sensor and a smaller sensor, the way that the light enters the lens
and creates an image will be the same. The only difference is that the camera with the
smaller sensor has less surface area to capture the image with. This makes it feel like the image
is ‘cropped in’ in comparison to the larger sensor which can capture more of the surface area and
therefore produce an image which feels wider. Calculat
ing crop factor and then changing
the lens on the camera to a more telephoto lens - may make the width of the images match,
but will also change the very nature of the image by altering the depth of field, the
compression and the distortion in the image. The smaller the sensor is, the more cropped
in the image will be and therefore the wider focal lengths you will need to use. Whereas
the larger the sensor is, the wider the shot will appear which means cinematographers will
often choose lo
nger, more telephoto lenses. One of the secondary effects from using
longer focal lengths is that it will create a shallower depth of field. This means
that the area that is in focus will be much narrower on a telephoto lens, which means
the background will be softer with more bokeh. This is why movies shot on cameras with
large format sensors bigger than Super 35, like the Alexa 65, which cinematographers
pair with longer focal length lenses will have a much shallower depth of field,
with
soft, out of focus backgrounds. It is a misconception that larger
sensors create this effect. In fact, it is the longer focal length lenses that do this. Another effect that focal lengths have
is on how compressed the image is. Wider focal lengths expand the background and make
objects behind characters appear further away. Telephoto lenses compress the background and have the effect of bringing different
planes closer to the character. For this reason, cameras with smaller
sensors that ne
ed to use wide lenses, may produce images that appear a
bit ‘flatter’ without much depth, especially in wide shots. While large
format cameras, with their longer lenses, compress the background to create a bit more
of a layered perception of dimensionality. Wider lenses also have a tendency
to distort the image more. So, shooting a close up of an actor
on a Super 35 camera with a wider focal length will expand their face and
make their features unnaturally larger, while using a longer foca
l length on a large
format camera with the same equivalent field of view will compress the faces of actors a bit
more which many say is a bit more flattering. Although modern digital cinema camera
sensors come in many shapes and sizes, in general they conform to a
few approximate dimensions. Some cameras come with the option to shoot a very
small section of the sensor that is equivalent to 16mm film. This has an approximate
crop factor of 2x compared to Super35. This little format will usu
ally be paired
with wider lenses designed for 16mm - such as the Ultra 16 Primes which range
from 6mm up to 50mm focal lengths, which with a crop factor applied produces
a field of view of around 12mm-100mm when adjusted for Super35. As we discussed,
this 6mm will produce an image with extremely limited bokeh and a deep depth of
field that feels quite dimensionally flat. Next we have Super35 sensors which are usually
considered standard, such as we find on an Alexa 35 or Red Helium. Each m
anufacturer produces
sensors with subtly different dimensions - but most will be the approximate size of 4-perf 35mm
film and produce the standardised field of view, where a 18-24mm focal length feels
wide, a 35-50mm lens is about a medium, and anything longer at around 85mm starts
to have a compressed, telephoto feel. Anything bigger than Super35 size is usually
considered to be ‘large format’. This includes ‘full frame’ sensors modelled on still
cameras that are approximately 36x24mm. So
me examples are the Arri Alexa Mini
LF, the Sony Venice 6K or the Sony FX9. These cameras will have a crop factor of
somewhere around 0.67x, which bumps a wider perspective up to around 32mm, a medium feel to
around 65mm and a telephoto lens to about 110mm. 65 cameras like the Alexa 65 push this even more
with their approximate 0,56x crop factor that makes a 45mm lens a wide, a 90mm lens a medium
and a 150mm a telephoto. As we discussed, shooting a wide field of view with a 45mm will produc
e
much more compression, bokeh and dimensionality than using a 12mm lens on a 16mm camera - even
though they’ll produce a similar field of view. It’s important to note that these crop factor
numbers are all relative to what sensor size is considered the ‘standard’. For example,
in still photography a full frame sensor is usually considered normal with a 1x crop
factor, which means smaller APSC sensors which are roughly close to Super 35 will
have around a 1.5x or 1.6x crop factor. What is
much more important than getting super
technical about these crop factor numbers is understanding how larger or smaller sensor sizes
affect the field of view and understanding all the secondary effects that using different
focal lengths will have on the image. The trick with cinematography is
taking all this technical camera geekery and actually applying it to
making movies. One of my favourite streaming services to watch films on
is MUBI - the sponsor of this video. Since we’re talking ab
out crop factor, let's
look at two different films streaming on MUBI shot in two different formats. First, we have
Christopher Nolan’s extremely low budget first film Following - which he shot himself on
the smaller gauge 16mm. Watching this is a great insight and introduction to a director
with a clear vision and voice whose stylistic trademarks from this film would continue to build
into some of his later monster blockbusters. On the other end of the crop
factor spectrum we have This Muc
h I Know To Be True - a musical
performance documentary shot by DP Robbie Ryan on a large format camera that
enhances its super wide field of view. To check out their full catalogue of films you can
try MUBI free for 30 days at MUBI.com/InDepthCine, that’s M-U-B-I dot com slash In Depth Cine
for a whole month of amazing cinema for free. Hope you enjoyed the video. As always thanks to
all of you for supporting the channel by watching, commenting and especially to those
that get the videos e
arly on Patreon. Otherwise, until next time,
thanks for watching and goodbye.
Comments
Bit of a swing and a miss on this one. I had to stop when you stated that the lens creates compression. Aside from the unique characteristics of each lens (sharpness, aberrations, etc), you will indeed get the exact same image if you use the crop factor principle. What's commonly forgotten though is that the crop factor applies not just to the focal length, but to the f-stop too. Framing up a head-shot on a Super 35 sensor with a 100mm lens at f/4 will give you the exact same frame, compression, DOF and distortion as staying in the same position and switching to a Super 16 sensor, with a 50mm lens at f/2. The only difference will be the exposure change due to the new f-stop, but this is also equaled out by the difference in sensitivity between the two sensors. All else being equal, a Super 35 sensor will show a similar amount of noise at, for example, ISO 3200, as a Super 16 will show at ISO 800, due to the smaller sensor having a worse signal to noise ratio given the smaller area on which it is registering light. I'm a fan of your channel, so don't take this as an attack, just a correction.
There's one major mistake, the compression is not due to the focal lenght but is due to the relative distance of the subject and the background from the sensor. So if you take a S35 camera with a 50mm lens and shoot a scene, and then, without moving anything, you shoot the same scene with a 16mm camera with a 25mm lens (same field of view) you will have the exact same compression of the background.
I'm usually a big fan of your videos but i think you should delete and re-do this one. You got a few things wrong in the background compression section. The only real difference in image between sensor sizes is the amount of background blur, but this has nothing to do with the sensor; it's just that a 30mm lens at f1.4 generally has less background blur than a 50mm at 1.4 if used from the same distance, and you need a larger focal length with a bigger sensor to achieve the same FOV as with a smaller sensor. In theory, if you had a sensor with infinite resolution, you could shoot everything with say a 16mm lens and then crop in later to get every other focal length. The only thing that dictates background compression is the position of the camera.
Sorry to pile on but it's not called a "micro four thirds sensor" (2:12). "Micro four thirds" refers to the camera system that uses a 4/3 sensor but doesn't have a mirror. Thus the cameras can be smaller (micro).
The lens does not "compress" the image. So-called compression is an effect of perspective. As you move away from something, it becomes relatively closer to the background. Making them seem closer. A 25mm with a 2x crop will have the same "compression" as a 50mm with no crop. It's the same perspective change that makes stars feel like they're inches apart, despite being lightyears apart.
Compression does not work like this. That's just a matter of distance between the camera and subject. You wouldn't use a 50mm as a "wide angle" close up on a s35mm being right in front of the subject, instead you might use a 35mm to get a wide close up. But if you put a 35mm on a 65mm you need to be much closer to the subject to get the same framing and therefore you get a more distorted wide effect. If you take a 50mm on a large format camera, match up the aperture a few stops for the same bokeh as a 35mm on a s35 sensor, frame it the same, they will look identical. The only two things that differs is that the resolution and glass distortions become much more visible the further out of the image circle you get; which leads to more lens artifacts visible on larger sensors, alternatively if the image goes into the extreme wide, the s35mm requires much wider lenses for the same effect and extreme wide angles usually are harder to get without heavy glass artifacts. The second part is covered in the video, but it's not as simple as this. You can match up much of the look between s35mm and large format, to a point. It's when you get into wider images that larger sensors show their prime. If you use a 24mm T1.4 for a really wide shot on s35mm with a shallow DOF, you get the same on a larger sensor with increased focal length around T2.8, but if you try to go wider on the s35mm you cannot retain the shallow DOF as wider lenses cannot be that fast. The larger sensor however can still go lower and retain a fast T-stop, resulting in a shallow DOF at lower focal lengths. So, you can get the same image on a s35mm sensor using a 35mm at T1.4 as a large format sensor with a 50mm lens stopped down. But as soon as you open that 50mm up you reach a DOF that isn't possible on the s35 sensor. The only way for that sensor to get a similar image is to go below T1.0 and that's a hell hole for glass clarity. This is the sole reason why larger sensors produce a shallower DOF, but the lenses do not compress or change any other aspects other than imperfections at the edge of the image circle.
All someone needs to do is shoot with 2 different focal lengths from the same spot with matching depth of field (ie. 35mm @ f/2 vs 50mm @ f/4), and then crop to match. The compression, bokeh, and distortion should look pretty much identical if the optics are of similar quality, but the image fidelity is higher on the larger sensor. Optical distortion varies from one lens to another - if someone compared a highly-corrected wide-angle lens to a pancake lens of a slightly tighter focal length with a ton of barrel distortion and then cropped to match, then you might think the tighter lens has more of the wide-angle distortion look.
I want to clarify things up. Its a hard thing to say, but the compresion effect you were referring to as the result of the focal lenght of the lens used does, in fact, hypothetically, depends solely and only by the distance between the camera and the subject not precisely the focal lenght itself. the focal leght of the lense does only one job -- cropping the image in front of your eye into preferred sizes, in other words, control the field of view. let's say we are to shoot a shot opted for medium shot size, on a full frame camera (namely large format for cinimatography). you would be needing 50 mm focusing at 1.0m to get that shot size on this sensor size -- this shall be our control set. Then, we are to shoot this same shot on a camera w different sensor size, -- lets say super 35 having a crop factor of 1.5 compared to Full frame(LF) -- if we use this same 50mm lens focusing at the same 1.0 meter the image will come out to be 1.5 time croped, maybe it will become medium close up, so you will either need to move the camera backwards to, lets say, 2 meters to get the same opted medium shot, which creates a compression effect on the background and ,hence, has more feeling of dimensionality blabla , or -- more normal usual way -- change the lens in front of this super 35 camera into a 35mm which will give 50mm "FULL FRAME FEILD OF VIEW EQUIVALENT" and be able to shot at the same 1.0 meter focusing distance. Apart from this, talking about depth of feild (aka background seperation aka bokeh) is a different subject and is already mentioned correctly in the video. I am currently teaching cinematography in an art school using mainly the pocket 4k having micro 4/3, and was a photography using a normal full frame dslr. honestly, there is no actual difference form this crop factor "compression " and "dimensionality" kinds of things (apart from the size of the bokeh which no one who actually shoot cinematograph cares about much once its already shallow enough). You just need to know how to choose what lenses to used. The only concrete enough differece ive heard so far is about the difference in lens designs for each "focal lenght" which tends to make the wider (lower number) prone to more distorsions and spherical abberation and conversly, the longer(higher number) prone to chromatic abberation, but, honestly, in this era, should be all taken for granted.
yikes... you got the section on "compression" completely wrong. it is NOT a function of the focal length. it is a function of DISTANCE FROM CAMERA. in all the cinema books where they're showing you the differences in focal length on image compression, they're actually MOVING THE CAMERA in order to keep the subject the same size in the frame. THAT is what causes the difference in compression. so for example, if you take the typical example of shooting telephone poles receding into the horizon vanishing point - if you shoot that with a wide angle lens, the poles CLOSER to camera will look farther apart. but in that exact same shot, shot with the wide angle lens, if you zoom in on the telephone poles in the distance with a magnifying glass, their distance from each other looks COMPRESSED. those poles, at that distance, shot with a telephoto lens to fill a screen will have the exact same compression that was visible when shot with the wide angle lens... it would just fill the screen.
What Mic do you use? On your YouTube videos?
If you have ever used a zoom lens you should immediately recognize that compression doesn't exist. Zooming in changes the focal length, yet zooming in expands the image the same as 'cropping in' on the sensor, same as zooming in in post. The one exception is that at a constant f-number, the zoom lens will make the background more out of focus as it zooms in. All lenses (outside of special machine vision lenses used in manufacturing) have the same perspective. Fish eye, 10mm, 1000mm, anamorphic - they just have different magnifications. Another way to convince yourself is to think of why objects don't look darker as they get further away. We know further away light sources dim by "one over the distance squared". Well all objects imaged by a lens also are rendered smaller by "one over the distance squared" - twice as far away and they appear half as tall, and half as wide, so one quarter as big, and so the light per area in the image never changes.
Take a full frame lens, change the back elements to narrow the image circle to match S35 (similar to how speed boosters work), you get the same FOV and DOF, but a higher exposure output, which in-turn, would help against the slightly worse sensitivity of the S35 sensor. This is why you tend to get faster APSC lenses relative equivalent Full Frame FOV. The relationship to the numbers change but thats about it. I think the differences between sensor size lay in the technicalities of the manufacturing and marketing departments.
Just to be clear: isn‘t the 1x crop factor referring to 35 mm wind up film in photographic stills cameras? Which is apparently larger compared to the S35 sensor area. I think that debating about crop factors and sensor sizes in digital cinema technology first started when people used full frame lenses on their (digital) APS-C or mFT camera bodies - which in the first place were DSLRs. The so called „full frame image circle“ was - as far as I know - never a significant measurement size for cinema lenses. Although lenses for larger format film areas like 65 mm negative stock existed since the mid 1900s. There are of course 0.x crop factors which refer to sensor/film areas larger than „digital“ full frame or 35 mm celluloid stills film but I think the 1x crop always referred to what is called „full frame“ (nowadays).
You went through this entire video without talking about aperture and how you DO get an identical image if you also apply the crop factor to the aperture. 27 mm f/1.4 on APS-C (photography term) / Super 35 (cinematography term) will look NEAR IDENTICAL to 41 mm f/2.1 Full-frame (photography term) / Large Format (cinematography term)
thank you for your content Sir, are you UK based? and Will you ever consider some live workshops?
One big question. Is s35 equivalent to full frame? Or in cine world the standard sensor size is s35 (aps-c) instead of full frame?
This is close but not quite right. Gerald Undone has a good video explaining the nitty gritty of crop factor, and Tony & Chelsea Northrup have a good video demonstrating an easy formula for applying crop factor to not only focal length but to F/T-stop (regarding its effect on DoF). Even a formula for ISO. But as a cinematographer shooting on a native ISO sensor, all you have to remember is to multiply both the focal length and the T-stop by the crop factor to get the equivalence (and then adjust your NDs or lights accordingly, obvi) if you're going to swap out lenses for the same FoV. In this case, a T/2.8 on Super 35 is more like an T/1.8 on Full Frame—not directly because of sensor size, but because if, when switching to a smaller sensor, you're changing to a wider lens to keep your FoV and distance from the subject the same, then your lens diameter at T/2.8 is going to be narrower than it was on the longer focal length when that one was set to T/2.8. So you'll have to open up to roughly T/1.8 to get the same DoF. If that isn't confusing lolol
I've been using MFT for almost 20 years now and finally understand what all my lenses do...
Bokeh and depth of field has always been a lens characteristic, never a sensor characteristic.
Super35 is crop factor 1.0 when we are in Cine Mode?