Interesting look at sensor size, apertures & focal lengths

[Frank Grygier]

Chris is the voice of reason here. A lens has no idea what sensor it is attached to. A lens with a F2.8 maximum aperture transmits more light than a F5.6 lens on any given sensor. If I am shooting M43 I want to have a 12 to 35mm F2,8. If I am shooting a 5D MK3 want a 24 to 70 2.8. However you want to twist it around these are the choices for a given format,

[Tim Polster]

I think you guys are missing the video's point. Yes you want an f2.8 lens on a 5D and a GH4 but the end result will NOT be the same from an exposure/ISO point of view. The GH4 due to the laws of physics will not be able to match the 5D in ISO settings because the 12-35mm lens does not let the same amount of light through as the 24-70. Even though they are both operating at f2.8. Duct tape the larger element 24-70 to the GH4 and focus all of the light to hit the sensor and it will be on par with the 5D. It is the lens size which is matched to the sensor size that is holding you back.

That is pretty much what the guys was after. f2.8 is not "the same" across different sensor sized cameras from a results point of view.

I think it is valuable for everybody to know this info. My compact camera shooting at f2.8 will need a ton more light than my 5D will...

[Chris Medico]

Originally Posted by Tim Polster

I think you guys are missing the video's point. Yes you want an f2.8 lens on a 5D and a GH4 but the end result will NOT be the same from an exposure/ISO point of view. The GH4 due to the laws of physics will not be able to match the 5D in ISO settings because the 12-35mm lens does not let the same amount of light through as the 24-70. Even though they are both operating at f2.8.

Please don't take this as argumentative. I do quite well get the point. You are right that the reason is due to physics. Quantum Physics to be more specific.

I'm not disagreeing with the relative observation regarding differing noise and sensitives between a FF35 and a M4T imager. Where I do not agree with the video is with the conclusions as to the source of the differences. The source of the differences is plain and simple pixel size and not lens sensitivity. If the pixels of a GH4 (3.7 micron) were the same size as the pixels of a 5d miii (6.25 micron) the GH4 would exhibit similar noise and sensitivity results with the same efficiency lens. The difference would then become resolution. You would have to give up about half of the resolution of the GH4's imager to bring its sensitivity and noise characteristics up to a 5d miii. I am not condemning the GH4 and praising the 5d. I am only trying to explain why you are seeing these differences with actual information instead of blaming the lenses which is completely misdirected.

I also disagree with this specific statement in the video. - Small pixels are not more noisy than large pixels (if fed enough light).

This is completely false. Smaller pixels have less dynamic range and higher noise coefficients. This is unavoidable due to quantum effects of making the P/N junction material thinner and thinner to accommodate the smaller pixels. Adding more light does not fix this problem. Electronically reducing the sensitivity of the sensor as you have to do when adding more light will only further reduce dynamic range. He shot the noise test and may have gotten things to look similar at that one light density but I suspect if he had done a full battery of tests for dynamic range with those same settings it would reveal compromised DR.

As sensor technology improves this situation will also improve. I will not say it will always be this way. Using the current CMOS technology it is a real limitation and the true source of the problem. Not the lens.

[Chris Medico]

Originally Posted by 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.

You are absolutely right that moving a magnifying lens closer and further from an object is NOT the same thing as changing the image circle of a compound lens. What I was attempting to do is describe the principal of light concentration in a way that would be more relatable. I should have made that more clear in my explanation.

[Dave Baker]

Once again, thanks to everyone for your replies.

I found this on Mirrorlessrumors.com (I hope I'm not breaking forum rules by pasting it here!). It, together with your comments, tells me that what I have always believed to be true, is.

"UPDATE: This is an important update by Ale (Mirrorlessrumors administrator)

There are a couple of logical errors made by Tony Northrup.

First: No one cheats. All company aperture lens info are correct! The Panasonic 12-35mm f/2.8 has a f/2.8 aperture and not a f/5.6 aperture. You don't have to make the equivalence he says has to be done! Use the Sony A7r and Sony A6000 for the same shot. To get same result the camera automatically sets on both the same ISO and same lens aperture despite the different sensor size. Just to say that the aperture remains CONSTANT and is not relative! There is no equivalence to make on that!

Please read that article why you cannot apply the aperture equivalence you mention on no other than the "depth of field" and "field of view" only:
"Full Frame Equivalence" and Why It Doesn't Matter - Admiring Light

The focal length and aperture do remain constant!!!!

Second: He says Olympus, Panasonic and Sony do cheat. Nope. All companies use the same kind of measuring aperture for all lens formats (medium format, MFT format, APS-C and so on).

---

UPDATE of the UPDATE: After re-watching the video I am sure Tony confused "Focal length" with "Field of view" ! It's the field of view where you can make the equivalence and not the focal length. Tony has been misleaded by the companies because they do that (small) error too. That's why when he did the math in the video to explain how aperture gets calculated he made the mistake to change the focal length variable. But actually that variable doesn't change at all! What changes is the "field of view" which has no influence on aperture. Hope you got the message :)

LIke Admiringlight says:

"I’ve heard many times “Yeah, your 75mm f/1.8 is crap – it’s like a 150mm f/3.6.” No, it’s not, it’s a 75mm lens with an f/1.8 aperture and a field of view that is the same as a 150mm lens on full frame."

I know Tony had good intentions and I wrote him and hope to will remove the video soon. But please guys, spread the word that what he tells about the equivalence is plain wrong. Don't worry your f/2.8 MFT or APS-C lens is really a f/2.8 lens!"

I shall stick with good old f-numbers and only worry about the ISO setting if it is likely to cause noise.

Dave

[Chris Medico]

If he had opened up the video stating something more like:

In my xxx years experience as a photographer I've used a lot of cameras with different sensors. It can be frustrating to figure out how to get the image you want from the array of different cameras and sensor sizes. In this video I'm going to share what I've learned about making images that have the same look and feel with DOF and noise regardless of the size imager in your camera.

From there if he had stuck to his experience as the angle for what he was presenting and left out the undertone of dishonesty from manufacturer's I would have given it two thumbs up.

I did find the setting information for how to get similar looking shots with different imager sizes interesting. He could have covered that info in 5 min and in the end had a much better video.

[Tim Polster]

Obviously this has a bunch of moving parts. I am not defending Tony the video maker as being all knowing. I just think he hit upon an area that is misunderstood.

Dave, I respectfully disagree with what you just posted. The Panasonic 12-35 f2.8 lens is an f2.8 lens but from a noise characteristic, on a M4/3rds camera, it will perform like an f5.6 lens on a FF camera. Yes, you will get the same exposure, but if you care about noise they are going to be different.

The ISO/ F-stop system is designed to have a constant photographers can rely upon. And it is achieved through varying the noise/sensitivity through cameras. So a 5D set at f4 ISO 100 will get you the same exposure as a GH4 at f4 ISO 100. But that does not mean the images will be "the same". The 5D will have less DOF and less noise.

This may not even be an issue given sensors are very low noise today, but it does exist.

After further thought, I almost think this new world needs a baseline ISO system. A universal constant that all cameras could be compared to. That way a compact camera and a FF camera would not be classified in the same category.

[Dave Baker]

I couldn't agree more Chris.

What threw me was, I learned what I know about photography the hard way, magazines, clubs and much trial and error. I also started a very long time ago, so I am quite prepared to believe that I have been happily snapping and, lately, shooting away and getting good results, in total ignorance of some fact or other.

I am still learning. For example, I only recently learned enough about diffraction limiting to understand its effects, but this video turned what I have learned over the years on its head. When I started to think about it, I realised something was wrong and started asking questions.

I shall stick with what I am happy with - shutter speed is shutter speed, f-stop is f-stop, ISO is another name for film speed (and therefore sensitivity) and crop factor doesn't matter until I mount a full frame lens on my APS-C DSLR. By the way, over here film was marked with both ASA and ISO for many years.

Thanks for helping.

Dave

[Dave Baker]

Originally Posted by Tim Polster

Dave, I respectfully disagree with what you just posted.

Tim, feel free to disagree as much as you like. I didn't say that, Ale did!

I appreciate your help.

Dave

[Bill Anderson]

The manufacturers are correct in their assertion that, for example, a MFT 12 to 35 f2.8 will equal a FF 24 to 70 f2.8, but only in two of three criteria - field of view and aperture speed (f2.8). What they omit to say, and perhaps because it's obvious, is that an f2.8 12 to 35mm lens will NOT give the same depth of field as a 24 to 70mm f2.8 lens at any given f stop. No surprise there. So Tony would be correct in applying math to the aperture when it comes to depth of field performance between the MFT lens and the Full Frame one: they are not the same, and when the manufacturer goes to great lengths to strongly imply that (by some miracle) a MFT 12 to 35 f2.8 equals a FF 24 to 70mm f2.8 in all regards, then that is just silly.

Edit: clarity

[Luc Spencer]

As a (now somewhat less proud) GH3 owner, I have done a DoF test a few months back. I mounted my camera on a tripod and took a photo of a close subject using the Panasonic Leica 25mm f/1.4, wide open. I then took the GH3 off the tripod and put a Canon 6D instead, that had a Canon 50mm f/1.4 lens attached to it.

I took the same picture.

There was an obvious difference in DoF between the 2 pictures. Meaning the picture taken with my GH3 had more depth of field, so noticeably more stuff was IN FOCUS.

I have also talked with a fellow videographer who films with a 5D, and when I told him that with my GH3 and that same Leica lens I can film at f/1.4, he raised an eyebrow and said that with his 50mm lens he needs to be at least at f/2.8 to avoid things getting out of focus fast.

It pains me to admit this, but this whole Panasonic GH-hype and the tiny "miracle" lenses for it are not actually what they advertise themselves to be. At almost every event I film, I always proudly show the photographers my Olympus 45mm f/1.8 and make fun of their excessively huge and heavy telephoto lenses. Not in a mean way of course, just to tease.

But the truth is, their Canon 100m f/2.8 has better looking bokeh than my f/1.8 AND has image stabilization which allows you to shoot handheld, something I can NEVER do with my Olympus and not even my Panasonic 14mm f/2.5, unless I am being very careful. Its Canon counterpart, the 28mm f/2.8 actually has stabilization and it works beautifully for video! I tried it 2 weeks ago. In addition, unless I'm being somewhat close to my subject, my 14mm lens even at said f/2.5 produces an almost camcorder-like DoF which kind of defeats the purpose of filming with a DSLR. I'm not surprised now, considering my f/2.5 is actually f/5.

I guess the only "take that!" argument that I actually have for the full frame videographers is, perhaps, that I paid somewhere around 1/3 - 1/4 of the price for almost all of my lenses.

A bit disappointing for us GHx users, to finally have to admit this.

PS: The GH4 is a bit of a different animal with quite a few extra features, chief among which is the 4k-capability, so I still consider that particular camera to be a very good buy and I would buy it myself if I could.

[Tim Polster]

DPReview has a comprehensive article on this issue. Basically saying the same info as in the YouTube video. What is equivalence and why should I care?: Digital Photography Review

[Frank Grygier]

I really dislike these discussions. They basically breakdown to my sensor is bigger than your sensor. We all know the pros and cons of FF vs MFT vs APSC. A lens with a diameter to cover a MFT sensor designed with a 2.8 aperture is a lens with a 2.8 aperture. If you want to talk about equivalency in framing a shot in the same lighting conditions with different size sensors accommodations will have to made in subject distance and exposure settings. So what.