Interesting look at sensor size, apertures & focal lengths

[Bryce Comer]

Hi all,
In my search for information on the GH4 i came across this very informative (at lest to me) youtube video about sensor size, apertures, & focal length.
https://www.youtube.com/watch?v=DtDotqLx6nA#t=19
Certainly interesting to see how the lenses being marketed by Panasonic & others are marketed in a way that is very misleading. A long video, but well worth watching especially if you are not exactly up to speed on all the technicalities of this stuff.
Regards,
Bryce

[Dave Baker]

Hi Bryce,

Thanks for the link, most informative.

Having been into photography for well over half a century, I thought I knew all this stuff but boy, was I wrong where ISO and crop factor are concerned! Like most (I imagine), I assumed that if I buy a lens or camera stated to be (say) f2.0, that's what I will get, whatever size the camera sensor is. Naughty marketing tactics by some? It would be interesting to know how manufacturers determine f-stop values and if they all use the same method.

I would like to know how this relates to fixed lens camcorders.

It's thought provoking stuff, I will certainly watch it a few more times until I memorise the maths.

Dave

[Gary Huff]

Originally Posted by Dave Baker

Like most (I imagine), I assumed that if I buy a lens or camera stated to be (say) f2.0, that's what I will get, whatever size the camera sensor is.

That's exactly what you get regardless of the sensor size.

[Bruce Dempsey]

+Gary ....not entirely
excellent find Bryce
in watching the vid, one is exposed to a good deal of light on the subject :)
I'm hoping that the info in this very informative piece applies to video in a much less dramatic fashion given the slow shutter speed 1/60 necessary for smooth video and in particular to the lumix fz1000 which I'll be using as a full time camcorder rather than primarily a stills cam, 1/60th of a second allows a generous exposure incident for lighting the sensor.
The next 6 weeks of waiting for my purchased Lumix fz1000 (if you use paypal they charge it right away) will be a new experience. Usually getting a new cam produces gratification within the week

[Chris Medico]

That video is more of a political statement than real information.

A couple of reasons for my opinion:

He doesn't talk about the real reason why DOF is different for large versus small sensors - circle of confusion. As the imagers get smaller the cross sectional area of the pixels get smaller. This changes how the light from the lens interacts with the pixels and with that changes the appearance of the DOF. The actual DOF for a 200mm f/2.8 lens doesn't change regardless of the imager it is resolving onto. What changes is how small the pixels are and in that the APPARENT DOF changes due to COF.

He also ignores a real issue with digital photography sensors which is that the manufacturers do not state signal-to-noise ratio in any relatable terms. That is what he should really be pointing out in the section about noise and ISO.

ISO makes sense when you understand that ISO is a numeric representation of the sensitivity of film. Films' sensitivity is determined by how much light is needed to expose the emulsion to a specific density of grey in the developed film. What makes this more or less irrelevant to digital photography is manufacturers don't tell you the S/N ratio for a specific illumination of a pixel to get some mid level output signal. That would be the equivalent way to think about it as compared to film. THAT is what he should be bringing to the conscienceless of the viewers instead of what he presented. How much light does it take to reach a specific S/N ratio with a specific electronic signal level (this is the equivalent of film density) would be the real way to calibrate ISO in a camera and would give you more comparable noise results between sensors and manufacturers. Don't hold your breath for that though.

Anyway, the information in the video has some crude relevance if you are looking to get similar looking images from different sensor sizes assuming a similar absolute resolution for each sensor. What is lacking is the true underlying reasons for the differences and disguised in its place is a rant about corporate dishonesty.

[Bryce Comer]

Yes for me though, it was very interesting to see just how much light i loose with a smaller sensor. Being able to put that into stops of light was very handy. My Nikon 80-400 f4.5-5.6 on my EX-3 for example has the field of view on a full frame sensor of a whopping 432-2160mm, BUT, it is far less sensitive on the 1/2" sensor of the EX-3 & is like using an f 24.3-30! Certainly that is how i understood it. Just made me think of how much harder you have to work to get the light you need onto the sensor to make the images we want.
Certainly sounds like we could use an overhaul of the system used to measure sensitivity, & it sounds like you are on the right track Chris. Something that represents the S/N ratio would be ideal. Apertures won't change, for a given lens, but how that relates to a specific sensor size does. This video for me at least, highlighted this.
Bryce

[Tim Polster]

I think the entire video is serving a good purpose as there is a lot of misunderstanding as well as liberties being taking in the industry.

Most of what has said is under the real world use umbrella. Pick up this camera you get this, pick up that camera you get that. All will not be equal even though it is implied it can be. You have to know your stuff and be an informed consumer.

I do agree with you Chris, noise or the signal to noise ratio is really at the heart of all of this. The S/N ratio is the area where camera manufacturers can cheat the science so to speak. If they can make a super efficient sensor it will break though the usual yardsticks and make up for the lower levels of light being let in with smaller lenses.

One example would be 1/3" camcorders. The most recent cameras have huge resolution and almost zero noise compared to the earlier generations. The lenses have stayed to the same size but they have made up the difference through sensor development.

[Dave Baker]

I have just re-watched the first part of this video and it's giving me more questions than answers.

Taking the analogy of a sensor surface area being like a bucket in the rain, then surely the lens is like a funnel. If I put a funnel with the same size inlet as the large bucket into a smaller bucket, the smaller bucket must collect the same amount of water. So if the analogy works, a lens designed for a crop sensor should put as much light on it as the equivalent maximum f-stop lens designed for full frame does on the full frame sensor.

If the sensor can't handle all the light, maybe due to it having smaller photo receptors (the analogy being the small bucket overflowing), then it's simply the equivalent of having a slower film in the camera?

Please humour me, I am trying to understand what the man meant.

Dave

[Tim Polster]

You have to think of the front element of the lens as the size of your bucket. What he is focusing on is just the amount of light allowed to hit the sensor. A lens with a larger opening will just let more light in than a lens with a smaller opening. Very simple concept.

Cameras with smaller sensors usually have lenses with smaller front elements. Even though these smaller lenses will say f2.8, they will not let the same amount of light pass as a full frame f2.8 lens. So this naturally inhibits the smaller sensor cameras due to light starvation, not sensor weakness.

If you put a Canon EF 24-70 f2,8 on a GH4 with an adapter that focused all of the light passing through (like a Metabones adapter does) the GH4 will then have sensitivity more like a larger sensor camera. Solely because more light is getting to the sensor.

Does this help?

[Chris Medico]

One thing we need to make sure stays straight in the discussion.

A f/2 lens can be large or small. Don't think the f2 lens on your pocket camera isn't f2 because it is small. It is f/2 and small because the image circle is also small. The image on a pocket sized f/2 will be equally as bright as a huge f/2. What will be different is the size of the image circle. The bigger lens will transmit more TOTAL light but since it will be spread over a LARGER image circle resulting in a brightness in any one spot will be equal to the small lens. You need a large image circle for a large sensor and you can get away with a small image circle with a small imager.

What the SpeedBooster adapter does is change the size of the image circle for the lens. It takes the larger image circle of a full frame lens and compresses it into a smaller space. With that the same number of photons are concentrated over a smaller area and the spot brightness is increased. You didn't change the f-number of the lens. You only changed the size of its image circle.

Lets look at it this way - Remember back when we were kids and used a magnifying glass to burn things? You figure out pretty quickly that as you move the lens back and forth from the object the size of the light spot grew and contracted. Once it contracted to a tiny dot we concentrated the energy enough to get things really hot. This is a practical demonstration of image circle. As you shrink the size of the image circle the light is more and more concentrated. You didn't change the f-number of the magnifying glass. You only changed the size of the image circle.

So, we need to keep in mind that unless you are stating the size of the lens pupil, the focal length of the lens, and the size of the image circle you aren't able to directly compare lens efficiencies. Or we can just use the f-number. ;)

[Dave Baker]

Thank you both for your replies.

My confusion arises from having known most of this stuff for years, but not being able to grasp the concept of modern lens labelling, or should I say mis-labelling?

I am picking up conflicting information about what appears to be the same thing and that is what does not make sense to me. Let me try to explain........... If I buy an f2.0 lens, it will be an f2.0 lens whatever the sensor size, as confirmed by Gary, right? Now, this is the bit that sounds wrong to me, if I buy a pocket camera with a lens that says it's f2.8 but it does not let as much light pass as a full frame f2.8 lens, then how can it be an f2.8 lens?

Mmmm........you're right Chris, I think I will just stick with familiar f-numbers and forget all this high fallutin' stuff, once I have found the answer to this. I have to because it will bug me until I do!

Dave

[Bryce Comer]

Dave,
The way i understood it, & i may be completely wrong here, but the f-stop of a lens is determined by the focal length divided by the diameter of the aperture. So for a 70mm full frame lens, if with its iris opened up fully, is 25mm in diameter, then 70 divided by 25 = 2.8 & that is the f-stop of the lens.
So what i think he is saying, is that if you have for example the Panasonic 12-35mm f2.8 lens, the lens is always going to be a 12-35mm f2.8 lens, but on the micro four thirds sensor, that same lens has a field of view half that of what it would be on a full frame sensor, so effectively like having a 24-70mm lens. Now this is where he is saying, that the manufacturers are being misleading, because in many cases they are saying that a given focal length lens, with a given f-stop gives a longer focal range in terms of 35mm equivalent, but also that the light gathering abilities of the lens is the same. He is saying no it's not. For a micro 4/3" sensor for example, the light would be half that of the f2.8 so, f5.6.
I think Chris has it nailed when he talks about the magnifying glass & moving it in & out to concentrate the light. It's the same amount of light hitting the magnifying glass's lens, but depending on where you have it, the image circle is increased or decreased. The smaller sensor size only gathers the light from the middle portion of the light coming through the lens, therefore, for a 1920x1080 image, we have half the light hitting the sensor than we do for a 1920x1080 image on a full 35mm sensor.
Metabones seems to have fixed this problem by using another lens element to focus all the light from the lens, onto the smaller sensor for any given camera, hence the need for all the different models for different cameras with different sensor sizes.
As for small pocketable cameras. If you look at the Sony RX10, the focal length on that is 8.8-73.3mm f 2.8 lens. Now with the 1" sensor size of this camera, if we think of focal length in full frame terns, that would be the equivalent of a 24-200mm lens. The iris diameter however, doesn't change. So if we divide the 73.3mm focal length of the lens by the f-stop of 2.8, we can see that the iris diameter when fully opened up, is 26.18mm. This sounds about right judging by the size of the lens on that camera. But if we say that the lens is equivalent to a 24-200mm f 2.8 lens, then the iris opening diameter would need to be a whopping 71.43mm! This i would have to agree is simply not possible.
Regards,
Bryce

[Tim Polster]

David, how about this approach: An f2.8 lens is always an f2.8 lens. If you have a beam of light that is 70mm wide passing through a lens it will provide a certain amount of illumination (photons) hitting the sensor. The issue with different camera/sensor sizes is how much of that beam will make it to the sensor.

So a FF lens with a 72mm front element would let all of the beam pass through to hit the sensor. A M4/3rds lens with a 35mm front element would only let half of the light beam hit the sensor. The lens is still a constant f2.8 on both cameras but the M4/3rds camera is receiving half of the light. So the M4/3rds camera would need to be set at a higher ISO to keep exposure constant. Which leads to more noise and less Dynamic Range.

So what the guy was trying to point out is the very question you are asking about. F-stops alone are not the only indicator in this scenario. ISO sensitivity has to play a part as well.

The advantage of FF cameras is being able to operate the camera at lower ISO levels which results in lower noise and possibly greater dynamic range images. The disadvantage of smaller sensors is being forced to use higher ISOs compared to larger sensor cameras. Image quality is not at issue given enough light in the scene.

[Chris Medico]

This is the kind of confusion I was concerned about with how the information was presented in that video.

The short answer is that video does not provide real information. It only gives the appearance of real information. It is more confusing that it is valuable. The only thing I saw of value was the charts that can help you get similar looks from cameras with differing sized imagers. Ignore the rest and you will be much better for it.

[Alastair Traill]

I think Chris’s short answer has nailed it – to me the video has more style than substance.

However I am not happy with Chris’s magnifying glass analogy, by moving the lens in and out you are focusing the image of the sun and the sharper focused it is the smaller and hotter is the image.

I thought ‘image circle’ referred to the size of sensor or film the image can usefully cover. For example can it cover a 1/3” sensor, the full 35mm frame or a full plate? The image circle is a function of the lens design.

My minor issue with ‘f’ numbers is that if you actually measure the iris diameter and divide it into the stated focal length you come up with a different ‘f’ number than stated on the aperture ring. I have always assumed that the discrepancy is related to the iris’s position in the lens and the properties of the various elements.