Do the math - focal length and Format demystified

[Guy Mancuso]

gotta give stefan his due for keeping cool with all the angst flowing

I Agree but everyone needs to keep there cool please. Its a discussion lets treat it like that and not personal. PLEASE

Thanks as I really don't like the insults and such. If you think someone is wrong or correct than post accordingly.

I honestly don't like being the peace keeper here folks. Im just pretty laid back and we are not going to die over this one.


The key here is figuring out what type of system may work best for your style or on a Pro level your clients needs. For the hobbyist and Pros alike its about getting the best bang for the buck and something you enjoy working with. Really end of the day all these systems are crazy good than when I started digital back in 1990. I can't tell you how much better off we are today.

 

[Shashin]

Now if they did this if you wanted a 74 degree angle of view in any format than you buy that lens that is in that format system. Easy Peasy.

I think this is actually a 24mm in FF. But if they created that system a lot of confusion would have been eliminated. Not all but some

Im still having a issue with loss of light on format change, not sure I am buying into that one yet. I may have to put the A6000 24mpx APS sensor up against the A7 24mpx FF sensor with exactly the same lens and than change they magnification on the A6000 by backing up the camera to the same angle of view. Im just not thinking there is any loss of light with this. The same amount of light is hitting the same area of view, that does not change. Why would my effective aperture number change is my real question. The answer given is bigger photo sites , correct. Well is that not compensated in the design of the smaller formats???????????????????????????????????

I need a better understanding of this because i am not sure i have really seen it in action. Now I am somewhat of APS newbie but if memory serves me well which it does sometimes the same dang sunny 16 rule applied to both MF and 35mm equally. ISO 100 is 1/125th at f16 regardless of system

This sensor area/light gathering thing I think is coming from a badly formed idea of information theory. The idea is that a larger sensor is gathering more "information" in terms of total signal. The problems comes that information needs context (not Contax). It does not mean a great deal in terms of what most photographers do--making pleasing images. Images only need, in terms of information, a limited set. Optimizing information will not lead to "better" results as the consumer of the information, the human visual system, does not need it. That is gobbly gook for a nice image actually looks nice and it does not matter what camera it is from.

And yes, the sunny sixteen rule applied to all formats. I have even used it with 4x5.

 

[stephengilbert]

"This sensor area/light gathering thing I think is coming from a badly formed idea of information theory."

So what? Is it really so hard to ignore someone you think is wrong?

 

[Shashin]

"This sensor area/light gathering thing I think is coming from a badly formed idea of information theory."

So what? Is it really so hard to ignore someone you think is wrong?

:wtf:

Guy had a question about this. I was just trying to answer it and trying to understand where this idea comes from.

Why are you even interested in this thread? You have not made any positive contribution and are just sneering at people.

 

[Godfrey]

One of the biggest problems that was created years ago was using the term focal length and putting a number on it like 24mm lens or 50mm. Back than everything was full frame in photography given the format. Today with APS , m3/4 and so on we are stuck with these multiplication factors to deal with to find the equivalent to FF. What they should have done back in yesteryear was called lenses by there angle of view. Than any lens made in any format change the consumer would go by angle of view. Than all this conversion stuff would have been eliminated. Okay that ship passed a long time ago

"Everything" certainly wasn't all 35mm. Long before a digital camera existed, I had Minox 8x11mm, Instamatic (26x26mm), half-frame (18x24mm), 35mm film in 24x32/24x36/24x65 mm variants, medium format in 6x4.5/6x6/6x9/6x12 cm variants, large format in 4x5/5x7/8x10/11x14 inch format.

I never bothered to think of lenses as "focal length==field of view" except when the context of the format was well known. I always thought of lenses as normal, wide, ultra-wide, portrait tele, tele, long tele instead. I never compared DoF scales either ... I looked at the lens scale, or a calculator table, to get a notion of what a given format/lens opening/distance setting would produce in DoF and learned to remember it.

Of course, I have regularly ignored marketing pap most of my life. And droning video bloggers are surely as bad as marketing pap ... ;-)

G

 

[Godfrey]

One of the biggest problems that was created years ago was using the term focal length and putting a number on it like 24mm lens or 50mm. Back than everything was full frame in photography given the format. Today with APS , m3/4 and so on we are stuck with these multiplication factors to deal with to find the equivalent to FF. What they should have done back in yesteryear was called lenses by there angle of view. Than any lens made in any format change the consumer would go by angle of view. Than all this conversion stuff would have been eliminated. Okay that ship passed a long time ago

Field of view angles are hard to remember. I bet no-one here can pull out of their head what the horizontal, vertical or diagonal angle of view of a 50mm lens on 35mm format is.

I would have preferred an even simpler system based on magnification. Define normal to be a focal length 1.15 times the format diagonal (to place the normal at the useful convention of a 50mm lens on 24x36mm format). Then:

ultrawide = less than .55x
wide = .73x to .62x
normal = 1x
portrait tele = 1.5x to 2.5x
tele = 3.5x to 5x
long tele = greater than 5x

That scales, can be applied to any format proportions, and whether it is horizontal, vertical, or diagonal angle of view is irrelevant. You can use either the name mnemonics or the magnification numbers to determine what a lens is giving you.

It'll never happen, of course.

G

 

[Stefan Steib]

Useful idea ! I think .... something like this was used back in the 50ies and 60ies and that was certainly not bad !
By this factor also zooms could be defined very easily.

:thumbup:

Greetings from Germany
Stefan

 

[Lonnie Utah]

WOW. You guys must REALLY care about this...

 

[iiiNelson]

Im still having a issue with loss of light on format change, not sure I am buying into that one yet. I may have to put the A6000 24mpx APS sensor up against the A7 24mpx FF sensor with exactly the same lens and than change they magnification on the A6000 by backing up the camera to the same angle of view. Im just not thinking there is any loss of light with this. The same amount of light is hitting the same area of view, that does not change. Why would my effective aperture number change is my real question. The answer given is bigger photo sites , correct. Well is that not compensated in the design of the smaller formats???????????????????????????????????

I need a better understanding of this because i am not sure i have really seen it in action. Now I am somewhat of APS newbie but if memory serves me well which it does sometimes the same dang sunny 16 rule applied to both MF and 35mm equally. ISO 100 is 1/125th at f16 regardless of system

The way I understand it from a "real world" perspective is the aperture value (read: ratio) doesn't change but rather the FOV due to the crop sensor only really projecting the inner 2/3 of the lens. I believe you can still use the lens in the same lighting conditions at the same ISO personally to create a similar look to the image. Case in point my 35/1.2 is able to be used in the same lighting conditions at similar apertures on FF or APS-C. On APS-C is obviously gives a "53mm" FOV but the DOF would still be the same DOF a 35mm lens would have on FF or APS-C.

With that said there's no conspiracy. Larger sensors are capable of more shallow DOF but that's not because of the lenses. That's because of sensor size. Aperture values like stated many times before are a ratio based purely on the lens and is sensor independent. Lens FOV equivalents are just that. The same way people can convert MF lenses by approximately 0.6 for FFMF or 0.8 for cropped MF to get a 35mm "equivalents." They will have more shallow DOF but I guarantee a MF lens with a aperture value of 4.5 can't be used in the same lighting conditions as a Noctilux wide open and handheld just because of a larger sensor.

Moral of the story aperture values (read: ratios) are constant and independent of sensor size. DOF is not and is dependent on sensor size... Oh and sorry I refuse to Google articles or math equations. I just did an experiment the old fashion way. Manual mode and constant set values.

 

[Leigh]

A lens projects a cone of light to the rear.
The distance from the rear lens node to that image when focused at infinity equals the lens focal length.

The intensity of the light at the image plane is a function of the lens aperture.

The lens does not know or care what if any film or sensor is at the image plane.

These are just basic laws of physics, unrelated to photography but applicable thereto.

- Leigh

 

[mazor]

hmm, based on the video all sensors are approx equal, and hence why the "light gathering" is less on micro four thirds, than it is APS-C, than full frame. But no sensor is made equal, and so it is possible that sensor manufacturers like Sony are able to make smaller sensors perform better in terms of signal to noise, and hence able to have a greater light gathering potential for a particular ISO.

 

[Guy Mancuso]

hmm, based on the video all sensors are approx equal, and hence why the "light gathering" is less on micro four thirds, than it is APS-C, than full frame. But no sensor is made equal, and so it is possible that sensor manufacturers like Sony are able to make smaller sensors perform better in terms of signal to noise, and hence able to have a greater light gathering potential for a particular ISO.

Exactly my thinking as well

 

[Guy Mancuso]

My a6000 and A7 where almost identical settings tonight. Was not scientific as I was shooting a gig and not testing but if I was to guess felt like 1/3 of a stop. Now the A6000 LCD is a touch dimmer too. So I would have to go by numbers to be sure

 

[ohnri]

Im still having a issue with loss of light on format change, not sure I am buying into that one yet. I may have to put the A6000 24mpx APS sensor up against the A7 24mpx FF sensor with exactly the same lens and than change they magnification on the A6000 by backing up the camera to the same angle of view. Im just not thinking there is any loss of light with this. The same amount of light is hitting the same area of view, that does not change. Why would my effective aperture number change is my real question.

For the same exposure the light hitting both sensors does have the same intensity.

Since the big sensor has more surface area but is hit by the same intensity of light it is hit by more total light.

For FF vs m4/3's it is about 4 times more total light since it has about 4 times more surface area.

So if you fill your image with the same face and shoot at the same exposure the FF image will have about 4 times the total light to create the face.

Four times the light equals two stops.

So, at the identical exposure settings, a FF image has two stops more total light to work with which means two stops less noise.

That is exactly what I found when I compared my EM5 to my D4.

You can also test this by shooting a scene then backing up and shooting it again with the same camera and same settings.

Crop the more distant image and blow it up to equal the size of the close up image. Presto, more noise. Also, more DOF.

Great photography depends on many things, format least of all.

-Bill

 

[Georg Baumann]

I was wondering, in the 4/3 and MFT world, if I remember correctly, they use so called telecentric lenses, right?

The f number is only considering objects at infinite distance!

The "Working f number" calculates objects differently (think macro).

Now this is what a significant manufacturer has to say about it:

Note that numerical aperture (and F-number) refer to both image and object space, as they can define both the cone angle of incoming and outgoing rays. Usually F-number refers to image space and numerical aperture is more commonly used in object space (incoming rays).
In macro lenses, like Telecentric Lenses, the F-number parameter loses its meaning as the object is not located at infinity; the working F-number should be used instead. Those two parameters come together in the formula:

source: Frequently Asked Questions | Opto Engineering 

 

[f64]

Nobody seems to mention the circle of confusion, but does not most of DoF discussion boil down to this? And it is not limited to sensors but applies to film as well or first.
http://http://en.wikipedia.org/wiki/Circle_of_confusion

 

[Stefan Steib]

Guy and all

the usable base sensitivity of the ready made camera will probably be kept the same (according to the given ISO´s) some can do 50, or 100 or 200 Iso.
Where the so called Photon count of the sensor- which is one of the real measurements of the quality of this sensor- is somewhat hidden in an electronic amplification to reach that pronounced base level.
This amplification will be lower on a chip with larger pixels and higher with smaller pixels ("photosites"). Normally such small chips as with MFT or even a Nikon V1 do have smaller pixels thus higher amplification-thus more noise- thus less quality.
Of course manufacturers have made huge progress the last years, the visible results are quite pleasing - at least when there is enough light. Things change already at low light conditions and noise will be more visible.

This was also shown in the video. Maybe not with a scientific nomenclature, but with the results, which is what many demand as they do not want to deal with formula and math.....

I still think this video is quite good.

Proposal: make a better one and then I hope anyone understands this in a second .....

Greetings from Germany
Stefan

PS.: the ideal sensor would have a photon count of 1, indeed there are some cameras which have such sensors - see e.g. Andors sCMOS cameras for scientific usage.

http://www.andor.com/scientific-cameras/neo-and-zyla-scmos-cameras/neo-55-scmos

 

[Leigh]

hmm, based on the video all sensors are approx equal, and hence why the "light gathering" is less on micro four thirds, than it is APS-C, than full frame.

Then the video is wrong.

I use exactly the same lens aperture for an exposure on 35mm and on 8x10".

That's about an 8x range of "sensor" sizes. Absolutely no difference in exposure.

The same is true on 35mm and MF digital.

- Leigh

 

[Shashin]

The problem of noise being a product of sensor area, rather than a function of the performance of the pixel is that is hard to find proof of that. Here are four sensors: m4/3, APS, 35mm, 44x33 at ISO 200 and 1600. There is no linear relationship that can be observed. If you want to know what noise a sensor has, just look at the specifications of the pixels.

 

[Leigh]

If you want to know what noise a sensor has, just look at the specifications of the pixels.

Part of the problem is that "noise" is the combination of several different error sources.

Although these are produced by different physical effects at different locations within the device, they all look the same to someone looking at a photo.

As an example, the real "sensor", being the light-sensitive device, generates noise based on its temperature, assuming that is above absolute zero. This noise is extremely low amplitude.

The sensor is followed by an amplifier chain that brings the signal level up to a level that can be processed by the microprocessor. This amplifier chain also generates noise based on temperature, but with different characteristics than the noise from the sensor.

That amplifier chain is also susceptible to environmental electrical noise, and amplifies that just as though it was a signal from the optical sensor. A gross example of this would be a lightning strike, but there are millions of weaker sources in our modern environment.

The list goes on and on, but I won't bore you further.

The bottom line is that you're trying to evaluate a very complex set of effects and processes as a single function, which it definitely is not.

- Leigh