Cinema glass has always been way more expensive than still camera lenses. Combined with the cinema camera sensors these high end lenses provide an image that's far superior to DSLRs capable of recording video. But yet, there are decent films created with DSLRs with still camera lenses. I'm not going to compare the glass quality here. I'm about to talk only about this peculiar T-stop measure on the cinema lenses while still camera lenses have an f-stop. Why should they differ?
Light Meter Reading vs. Desired Exposure
Both measurements relate to the theoretical amount of light that enters the sensor through the lens. How many of you have noticed your light meter tells a reading that's slightly off by a third or two-thirds of a stop on your reference monitor. Most of the time we either ignore it and fix it in post, or we simply change the aperture on the lens to compensate for that. This mostly happens with still camera lenses; even with high end ones. For example, you measure f/4.0, but on your monitor that's too light, so you change it to f/4.5, and it's just fine.
This also happens with the camera sensors. Not all camera sensors are the same and ISO 200 is not the same for every camera. You may have ISO 200 based on the light meter reading accurate on one camera and slightly off as exposure on another.
Why Accuracy Is Critical In Cinema
Knowing that the combination between lenses and camera sensors can lead to light meter readings inaccuracy, makes the manufacturers of cinema lenses to be more careful about their lens measurements on the barrel. When shooting video footage it's far more complex, as you may know. You have lots more people involved and for every scene you roll cameras and sound. You need more light in terms of power consumption and more expensive camera sensors; the whole industry is a lot more expensive. There are scenes you shoot on different days and different places, and you need them to be matched as exposure. Fixing the exposure in post costs more, and for a feature film you may find it's cheaper to use cinema lenses instead of paying for the post work. You may do that because cinema lenses are more accurate with their aperture measurements. The T-stop is exactly for that purpose.
T comes from "transmission". Every cinema lens is individually tested and T-stops are marked specifically for it. When your light meter tells you 4.0 it's a T 4.0 exactly on that lens. That's one of the reasons the production of a cinema lens is more expensive. When using different cinema lenses for the same or related scenes, you will can get the same exposure every time trusting your light meter readings.
As funny as may sound to you, cinema lenses help to lower the production cost on big budget projects.
Can't We Use Still Camera Lenses?
We can absolutely use them! Although we can rent expensive cinema glass for some low budget projects, it may not be necessary. Yes, we will not have the quality of that glass using still camera lenses, but compensating for the exposure readings is much easier today with digital cameras. We can always do this by checking our histograms, wave monitors, using our internal camera meters, and calibrating our light meters to the lenses we use. The difference between an f-stop and a T-stop is, most of the time, up to a third of stop. If the image quality of the still camera lens is good enough for the project we can surely go with it for our video production.
If you want to see more videos like the one above, check their YouTube channel.
This is not a carefully written article, and there are better explanations elsewhere.
ISO 200 on one camera is ISO 200 on another camera -- the claim that "ISO 200 is not the same for every camera" is simply untrue for any product of respectable quality. It is easy for the manufacturer to calibrate that aspect.
The f-stop is not some theoretical value. An f/2 aperture is created when the diameter of the aperture is exactly half the focal length (e.g. 25mm on a 50mm lens). There can be minor variations in practice because on many lenses the aperture is closed rather mechanically and swiftly as the photo is captured, and this mechanism might not be calibrated perfectly or work consistently. But that's a quality/manufacturing issue rather than a fundamental problem with f-stops.
The relationship between f-stop and the amount of light hitting the sensor is theoretical, as it assumes the optics are 100% perfect at transmitting light. The T-stop is about taking into account the transmittance of the lens optics, which always is less than 100% perfect. Unlike the example in this article, it won't be "too light" but will always be dimmer.
Aside from the difference between f-stop and T-stop, there is a difference in the precision of manufacturing for cinema lenses. The video/article claims the markings are calibrated for each lens. I very much doubt someone at Zeiss paints the dashes in different places for each lens as they test it. Instead they will ensure the aperture mechanism is aligned accurately with the dashes and adjust if necessary.
Experience shows otherwise both for lenses and camera sensors.
The reason most cameras tell you they have a "native ISO" is a particular sensitivity value gives you the most dynamic range for a certain lighting condition. Many times higher ot lower ISO's, depending on if the scene has enough light or doesn't, or is quite bright, won't give you the same picture. Yes, theoretically they all will meter internally ISO 200 (for example) but the picture will be different because of the different sensor characteristics. Some sensors require more light, others work well in the darkness. The metering will show you a result but the reference monitor will tell you the real truth.
A simple example is to get 2 cameras and shoot a picture with them with the same lens. To make sure the test is accurate, switch the lens from one body to another. Have a dark room with a lamp in the middle. Make a photograph of the room with the same settings. You will see different parts of the image lit differently on both cameras. This is because of the sensor sensitivity to light when there's not enough light or something is overexposed, or the light is just about right. If you try to shoot the dark sides of the room with both cameras, you will probably have different results.
For a flat lit color chart you may not have exposure diffeences but who shoots images and footage of charts for clients? We all shoot things that actually tell stories, not color charts. In lab environment all cameras should follow certain standards, but in extreme lighting conditions every camera and lens has its own good and bad sides, including exposure deviations.
I even talk about some sub $10K cinema camera sensors. The higher the price, the more accurate they all are (sensors and lenses).
As for the Zeiss lenses, the video in the article shows an example with 2 of their lenses to show the difference between labeled aperture and actual exposure.
This is the problem with the article. It isn't careful with language. It isn't that we disagree that T-stops are important or what the difference is, but if you are going to write an article saying why, then it needs to be correct about that. And quality-control / precision-design is a separate issue from the f/T difference.
You agree that for test charts there will be little exposure difference. Yes: that's exactly what ISO is.The clue is in the name: an international standard. Actually, there are few ways manufacturers can measure/calibrate their ISO, but the point is that they are all attempting to achieve a consistent result. But this is for a very simple standard, such as an 18% grey card and a very simple curve from dark to light. It won't be the case that "some sensors require more light": that's too vague a statement and misleading. Within the limitations of their measurement technique and manufacturing consistency, ISO 200 is ISO 200. Like f/2 is f/2 no matter what lens you pick. The variations we see are caused by things other than the ISO setting or the diameter of the aperture.
Of course a real picture taken with different cameras at the same settings will look a bit different. Every manufacturer has their own technique for producing a JPG, and every raw conversion software also has their own techniques. Look at the various camera calibration options offered by Lightroom to see the possibilities. Some manufacturers have great dynamic range with superb shadow detail, and others are poor in that regard. Some are noisier than others. But this isn't because ISO200 is not ISO200 on another camera, or because f/2 is not f/2 on another lens.
Wrt your Zeiss lens comment: yes even Zeiss lenses do not have 100% transmittance or the same transmittance from model to model. That isn't my point. My point is the sentence " Every cinema lens is individually tested and T-stops are marked specifically for it" implies that the location of the painted T-stops on the barrel differs from individual lens to individual lens. The video says much the same. I very very much doubt that is the case, especially as these are not just painted-on but etched by machine and filled with white paint. But I would be most interested to be proved wrong. What I think is far more likely is that Zeiss cinema quality-control will ensure the markings on the barrel are correctly aligned, and that their aperture mechanism has better accuracy and consistency than consumer optics.
With that clarification, I agree.
Looking at files from two different cameras at ISO 200 will give different results.
I'm a little confused on the f-stop explanation; he called it theoretical light transmission. I was under the impression that f-stop was a measure of the ratio of the lens-focal-length to aperture diameter (100mm lens with 50mm aperture is f2.0). His other video on the logic behind aperture discussed the math behind stops based on √2. The math makes sense, but he never explained the physical method by which the numbers are derived. If you close down the aperture to allow 1/2 the light in then of course the √2 math will follow because the aperture number is based on a 2D ratio.
What I think confuses most people (myself included) is that the same math works for both physical ratios (f-stop) and measure of light transmission (T-stop). Meaning, the T-stopt takes into account the amount of light that is transmitted the lens ONTO a 2D plane (the sensor) and is measured in stops based on √2 and f-stops measures the amount of light transmitted THROUGH a 2 dimensional plane (the aperture) which is also necessarily measured in stops based on the √2. Basically The f-stop doesn't take into account the light loss due to glass elements because it is only concerned with aperture size.
TL;DR Both f-stop and T-stop measure light transmission based on 2D planes and thus use the same measurement scale because math.
Hopefully that makes sense. I'd love input if you understand it differently.
Here is the original video: https://www.youtube.com/watch?v=LPlVBLN4Pp0
Yes, roughly speaking, T-stop = f-stop - <imperfections>
The confusion comes from the F-stop (f-ratio) and EV(exposure value) that are using F-numbers as a unit.
EV is a derivative of f-ratio because f-ratio is a theoretical value based on theoretically ideal lens. All basic theories and calculations in optics and physics are based on ideal, theoretical scenarios.
Much disagreement about the math. As I read this article, it's about light and the way cinema lenses better deal with exposure issues. F-stop expresses a geometric relationship. T-stop references light transmission. The way I understand it f-stops really breaks down when you are dealing with lenses of different focal lengths. Say your shooting a scene with two cameras. One camera shooting the wide shot with a 35mm lens and the other shooting the close-up with a 135mm lens. To accurately match exposure between cameras you need T-stop cinema lens.
Cinema is different and it's refreshing to think about things differently some times.
That's correct. The good news for us, who don't have piles of money around us, is f-stop is not that far from the truth. We pay for that in post.