Crop 50mm F/4 = Full frame 80mm F/6.4 (I think)

[whom]

Now I am sure there are equations and what not that prove or disprove this but my understanding is a lens on a crop body gives the same look as a lens full frame on a full frame 1.6 times longer and 1.6 times the aperture. I'm sure this has been done before....

So I tested this by taking a picture at 50mm F/4 and cropped it. Then I compared it to a shot at 80mm f/6.3. Now in order to maintain a fair test the ISO of the full frame shot was bumped up a stop (should have been a bit more really).

So in this hypothetical situation where I have a 80mm F/6.3 I would require 1.6x1.6 the ISO setting (2.56 times eg 2000 vs 800) of my friend with his 50mm F/4. With the same megapixel count and ISO performance matched at 800 vs 2000 then we should get identical pictures.

Here are my two shots with the main aim being to have an out of focus spot in the background for comparison.

Am I right? Does it really take a crop 50mm F/1.8 to match a FF 80mm F/2.8?

 

[woof woof]

Because the APS-C sensor is smaller than FF you'd have use a wider lens to maintain the same FoV so you get more DoF.

Therefore to get the same DoF on APS-C as you'd get with a longer lens on FF you'd need to use a wider aperture because you're using a wider lens

It's easier with MFT as that's a x2 crop so there's less brain ache.

PS. But for exposure f1.8 is f1.8 no mater what you mount it on. The "crop factor" is only for FL and DoF at the same FoV.

 

[HoppyUK]

<snip>

Am I right? Does it really take a crop 50mm F/1.8 to match a FF 80mm F/2.8?

The maths works

 

[heidfirst]

It does on a Canon - on Nikon, Pentax & Sony APS-C it would be a 75mm (1.5x crop instead of Canon's 1.6x).

 

[whom]

Hmmm so this means full frame has a false advantage with ISO.... it has to be just over one stop cleaner because for a matching picture it requires ISO2000 when the crop can use ISO800 (one a Canon crop 1.6)

So lets say we were taking a picture of a group, and the front and rearmost people had to be in focus and the left and rightmost had to be in frame and we were the same distance away both shots and the shutter speed must also be the same. For the same depth of field, our full frame needs to be F/6.3, 80mm and ISO2000 when our crop can be F/4, 50mm and ISO800.

So for the same megapixels, each full frame pixel sees 1.6x1.6 (in each direction) more light. So for at same ISO sensitivity it actually sees MORE photons making the signal to noise ratio lower, I.E. less noise.

So assuming everything scales linearly (which I bet it doesn't but anyway) ISO2000 on a full frame should match the noise of ISO800 on a crop of the same megapixels...... If noise is proportional to ISO and signal is proportional to the pixel area then the signal to noise ratio matches (ISO2000 vs ISO800 crop)

Am I still right... it is all sounding a little far fetched but seems to make sense.

 

[SamHH]

What you wrote is correct, but you wouldn't always need the same depth-of-field, and therefore wouldn't always need to stop down more with a full-frame camera.

 

[HoppyUK]

Hmmm so this means full frame has a false advantage with ISO.... it has to be just over one stop cleaner because for a matching picture it requires ISO2000 when the crop can use ISO800 (one a Canon crop 1.6)

So lets say we were taking a picture of a group, and the front and rearmost people had to be in focus and the left and rightmost had to be in frame and we were the same distance away both shots and the shutter speed must also be the same. For the same depth of field, our full frame needs to be F/6.3, 80mm and ISO2000 when our crop can be F/4, 50mm and ISO800.

So for the same megapixels, each full frame pixel sees 1.6x1.6 (in each direction) more light. So for at same ISO sensitivity it actually sees MORE photons making the signal to noise ratio lower, I.E. less noise.

So assuming everything scales linearly (which I bet it doesn't but anyway) ISO2000 on a full frame should match the noise of ISO800 on a crop of the same megapixels...... If noise is proportional to ISO and signal is proportional to the pixel area then the signal to noise ratio matches (ISO2000 vs ISO800 crop)

Am I still right... it is all sounding a little far fetched but seems to make sense.

Yes, that all pans out equally, in theory. Except that that as you say it doesn't 'scale linearly' as you suspect and in recent years we have seen smaller formats like APS-C and M4/3rds benefiting from more development resource and closing the gap on full frame.

The gap there looks like it is being widened again with new cameras like the Nikon D800 and Canon 5D3 - the first all-new full framers for quite a while. The area that seems to be suffering most is medium format digital whose CCD sensors are now way behind the times, despite the cost.

The bit that's missing is lens performance. It is much harder for a lens to deliver the same standard of sharpness at the higher resolution levels demanded by smaller formats (with the same pixel count). Laws of physics dictate that as resolution goes up, contrast goes down, and it is contrast that contributes most to perceived 'sharpness' (basic lens MTF theory). It's a bit like expecting a car that can do 0-60mph in six seconds, to take only another six seconds to go from 60-120mph - it cannot happen.

Lens performance is much less 'linearly scaled' and issues of size/weight and cost have a disproportionate impact according to format. For example, very low f/numbers are much easier on smaller formats, and wide zoom ranges, very short focal lengths too.

At the end of the day, it's a question of which particular format offers the optimum mix of compromises for a particular use.

 

[woof woof]

Hmmm so this means full frame has a false advantage with ISO....

I sold my APS-C camera and now have MFT and FF. FF still offers the best image quality (of FF, APS-C and MFT) but shooting at lower ISO's helps to minimise the gap a little and shooting with faster shutter speeds for the same FoV and DoF can have advantages too.

 

[danbroad]

The f stop is a lens dependent constant, and is unaffected by sensor size. F/4 is always f/4.

However, the view you get on the APS-C final image is equivalent to a 75/80mm lens on full frame, with the depth of field increase proportional to the APS-C crop.

 

[whom]

lots of interesting stuff

Thanks for the info Hoppy, it makes sense. Another thing that doesn't scale linearly is the size of the photodiode in each pixel (because the circuitry would not get bigger) I assume..

The f stop is a lens dependent constant, and is unaffected by sensor size. F/4 is always f/4.

However, the view you get on the APS-C final image is equivalent to a 75/80mm lens on full frame, with the depth of field increase proportional to the APS-C crop.

I think that is wrong. The focal length divided by the f stop gives aperture diameter (as seen when looking at the front of the lens, see wikipedia "entrance pupil") and that is the only constant here. So increasing the focal length causes the f stop to increase by the same amount.

So a 17-55mm F/2.8 on a crop is the same as 27-88mm F/4.5.

To put it another way, if you want the same look of as 24-70mm F/2.8 gets you on a full frame you'd need a 15-44mm F/1.8 on your 1.6 crop....!

 

[HoppyUK]

The f stop is a lens dependent constant, and is unaffected by sensor size. F/4 is always f/4.

However, the view you get on the APS-C final image is equivalent to a 75/80mm lens on full frame, with the depth of field increase proportional to the APS-C crop.

As Steve says, while that's right in terms of exposure per pixel, it's a ratio to focal length and not the same in terms of total light delivery. Eg, f/4 on say M4/3rds (crop factor x2) is the same as f/4 on FF for exposure, but as the the FF sensor has four times the area, so the size/area of the actual hole must also be four times larger.

 

[woof woof]

I think that is wrong. The focal length divided by the f stop gives aperture diameter (as seen when looking at the front of the lens, see wikipedia "entrance pupil") and that is the only constant here. So increasing the focal length causes the f stop to increase by the same amount.

So a 17-55mm F/2.8 on a crop is the same as 27-88mm F/4.5.

To put it another way, if you want the same look of as 24-70mm F/2.8 gets you on a full frame you'd need a 15-44mm F/1.8 on your 1.6 crop....!

The whole point of the f system is to make the exposure the same. f2.8 is always f2.8. DoF is another matter and for the same FoV you have to use wider lenses and thus wider apertures as your format gets smaller but the exposure is always the same.

It's important to remember that "crop factors" are nothing to do with exposure as such, other than exposure being affected by any decisions you make relating to DoF / aperture selection. f2.8 is the same for exposure no matter what your sensor size but on a smaller system you could use f1.4 and get the same DoF because you're using a wider lens and be able to use a higher shutter speed or lower ISO.

This is quite easy as long as no one mentions CoC

 

[HoppyUK]

Yes, that all pans out equally, in theory. Except that that as you say it doesn't 'scale linearly' as you suspect and in recent years we have seen smaller formats like APS-C and M4/3rds benefiting from more development resource and closing the gap on full frame.

The gap there looks like it is being widened again with new cameras like the Nikon D800 and Canon 5D3 - the first all-new full framers for quite a while. The area that seems to be suffering most is medium format digital whose CCD sensors are now way behind the times, despite the cost.

The bit that's missing is lens performance. It is much harder for a lens to deliver the same standard of sharpness at the higher resolution levels demanded by smaller formats (with the same pixel count). Laws of physics dictate that as resolution goes up, contrast goes down, and it is contrast that contributes most to perceived 'sharpness' (basic lens MTF theory). It's a bit like expecting a car that can do 0-60mph in six seconds, to take only another six seconds to go from 60-120mph - it cannot happen.

Lens performance is much less 'linearly scaled' and issues of size/weight and cost have a disproportionate impact according to format. For example, very low f/numbers are much easier on smaller formats, and wide zoom ranges, very short focal lengths too.

At the end of the day, it's a question of which particular format offers the optimum mix of compromises for a particular use.

Further to this on the lens sharpness effect and sensor size, here's a link to the brand new Leica 50mm f/2 Summicron that is claimed to be the sharpest lens of its type. It looks damn good BTW, though if you want to see the best MTFs I've ever seen, check the Canon site and see what their super-tele 300/400/500/600mm L lenses can do Click on Technical Data on the right of this link, then scroll down to the MTF graphs on page 3 http://en.leica-camera.com/photography/m_system/lenses/8884.html

It's a good example as all three graphs clearly show four pairs of lines, at 5, 10, 20 and 40 lines-per-mm. Note how as resolution demands are doubled, % MTF on the vertical axis drops in a fairly consistent doubling relationship too.

Also bear in mind that the 5/10/20/40 measures are a historical thing relating to film use (most other manufacturfers quote MTFs at 10 and 30lpmm for the same reason). Digital sensors have far greater resolution potential, and while comparisons between lpmm and pixels are difficult, if you simply take the Nikon D800 with 36mp, that translates to 204 pixels per mm.

 

[whom]

The whole point of the f system is to make the exposure the same. f2.8 is always f2.8. DoF is another matter and for the same FoV you have to use wider lenses and thus wider apertures as your format gets smaller but the exposure is always the same.

It's important to remember that "crop factors" are nothing to do with exposure as such, other than exposure being affected by any decisions you make relating to DoF / aperture selection. f2.8 is the same for exposure no matter what your sensor size but on a smaller system you could use f1.4 and get the same DoF because you're using a wider lens and be able to use a higher shutter speed or lower ISO.

This is quite easy as long as no one mentions CoC

No I don't think you understand. For the same photo, on a full frame you need 1.6 times the focal length, 1.6 times the aperture and 1.6*1.6 the ISO. This gives the same shutter speed for the same picture with the same signal to noise ratio. In theory.

I know what you are saying but it isn't the point. I am not talking about exposure. I am talking about the same picture. My two hypothetical lenses, 50mm F4 and 80mm F6.3 have the same sized entrance pupil! Entrance pupil = focal length divided by f number. 50/4 = 80/6.4 = 12.5mm. It is the entrance pupil that dictates the depth of field. Nothing else I believe.
http://en.wikipedia.org/wiki/Entrance_pupil

 

[whom]

Further to this on the lens sharpness effect and sensor size, here's a link to the brand new Leica 50mm f/2 Summicron that is claimed to be the sharpest lens of its type. It looks damn good BTW, though if you want to see the best MTFs I've ever seen, check the Canon site and see what their super-tele 300/400/500/600mm L lenses can do Click on Technical Data on the right of this link, then scroll down to the MTF graphs on page 3 http://en.leica-camera.com/photography/m_system/lenses/8884.html

It's a good example as all three graphs clearly show four pairs of lines, at 5, 10, 20 and 40 lines-per-mm. Note how as resolution demands are doubled, % MTF on the vertical axis drops in a fairly consistent doubling relationship too.

Also bear in mind that the 5/10/20/40 measures are a historical thing relating to film use (most other manufacturfers quote MTFs at 10 and 30lpmm for the same reason). Digital sensors have far greater resolution potential, and while comparisons between lpmm and pixels are difficult, if you simply take the Nikon D800 with 36mp, that translates to 204 pixels per mm.

Thanks for all this info Hoppy, makes for interesting reading. I love how Leica use a full frame sensor. As they say, "no compromises". That lens looks good but would look a little silly on my 5D I fear.

From your understanding, do you agree with my calculations etc? For example, up-scaling a crop sensor to full frame (back to linear scaling here) sized would give the same "noise" (or signal to noise ratio) at 1.6^2 the ISO?

 

[arad85]

It is the entrance pupil that dictates the depth of field. Nothing else I believe.

Nope.

Depth of field is ONLY perceived when you render the image.

Everything that is off the focal plane (and that itself may not be planar) is out of focus. This is by definition. The further away you get from the focal plane, the more out of focus an object is.

Whether you can see it as out of focus depends on how big you render it, how far away from the render you are, how good your eyesight is and how much bigger than the sensor the render is (which is why sensor size comes into it). For example, if you looked at your pictures posted above from a far enough distance, they would look to be "in focus" as far as you could tell even when they clearly aren't when looked at on a monitor.

The entrance pupil just dictates how quickly the focus changes over distance. Whether you can see that as out of focus (and hence perceive a DoF in the first place) depends on the factors mentioned above.

See.. I didn't mention CoC once

 

[whom]

Nope.

Depth of field is ONLY perceived when you render the image.

Everything that is off the focal plane (and that itself may not be planar) is out of focus. This is by definition. The further away you get from the focal plane, the more out of focus an object is.

Whether you can see it as out of focus depends on how big you render it, how far away from the render you are, how good your eyesight is and how much bigger than the sensor the render is (which is why sensor size comes into it). For example, if you looked at your pictures posted above from a far enough distance, they would look to be "in focus" as far as you could tell even when they clearly aren't when looked at on a monitor.

The entrance pupil just dictates how quickly the focus changes over distance. Whether you can see that as out of focus (and hence perceive a DoF in the first place) depends on the factors mentioned above.

See.. I didn't mention CoC once

I understand what you mean, I should not have used the term depth of field. I suppose what I should have said is the rate of change of focus or something. For a subject at a given distance the entrance pupil dictates the maximum distance apart two light rays can be that come from the same place.

As in, as long as the entrance pupil is the same then it doesn't matter what crop or sensor size you use, the stuff out of focus will be the same out of focussness (assuming you blow it up to the same size on a screen)

 

[HoppyUK]

Thanks for all this info Hoppy, makes for interesting reading. I love how Leica use a full frame sensor. As they say, "no compromises". That lens looks good but would look a little silly on my 5D I fear.

From your understanding, do you agree with my calculations etc? For example, up-scaling a crop sensor to full frame (back to linear scaling here) sized would give the same "noise" (or signal to noise ratio) at 1.6^2 the ISO?

Yes, but the ISO thing depends on a lot of other factors relating to sensor design and performance. What is certainly true is that at a given f/number there is about 1.25 stops more light hitting a full frame sensor compared to APS-C. Though of course, amount of light falling on the sensor, and the amount captured, are not the same thing

However, it does translate more or less to a similar change in ISO, but only when comparing sensors of similar construction and technological era. For example, if you take a Nikon D800 and crop it down to 1.5x you end up with something extremely similar to the D7000 in every respect.

There are physical differences between sensors, particularly relating to the architecture between pixel sites, the efficiency of microlenses (none are truly gapless, or anything like 100% efficient) and then again back-lit sensors also have advantages. Last but certainly not least is the processing engine, which is how Nikon tends to get a bit more performance out of Sony sensors than even Sony itself sometimes, basically because their amplifiers etc are more expensive/better.

BTW, I think the debate about entrance pupil/aperture etc is a bit of a semantic distraction.

 

[whom]

Yes, but the ISO thing depends on a lot of other factors relating to sensor design and performance. What is certainly true is that at a given f/number there is about 1.25 stops more light hitting a full frame sensor compared to APS-C. Though of course, amount of light falling on the sensor, and the amount captured, are not the same thing

However, it does translate more or less to a similar change in ISO, but only when comparing sensors of similar construction and technological era. For example, if you take a Nikon D800 and crop it down to 1.5x you end up with something extremely similar to the D7000 in every respect.

There are physical differences between sensors, particularly relating to the architecture between pixel sites, the efficiency of microlenses (none are truly gapless, or anything like 100% efficient) and then again back-lit sensors also have advantages. Last but certainly not least is the processing engine, which is how Nikon tends to get a bit more performance out of Sony sensors than even Sony itself sometimes, basically because their amplifiers etc are more expensive/better.

BTW, I think the debate about entrance pupil/aperture etc is a bit of a semantic distraction.

I've wondered when backlit will hit DSLRs, has it arrived yet? I quite liked the idea of that sensor Sigma use. Let me find a link...
http://www.sigma-sd.com/SD1Merrill/system.html

Seems sensible but I know it's got its drawbacks.

Yeah I didn't start it... haha. what I was trying to say that from the same distance, it doesn't matter what your focal length and aperture are, as long as the entrance pupil is the same diameter you'll get the same results when cropped to the same size.

Something else, while I am here, is the idea of a "tele-compressor". I think it is a GREAT idea. But why don't they exist? Eg, take a 0.63x converter on a EF lens and suddenly your crop sensor will no longer have a crop factor.

And before someone says it can't be done, read how they made a 50mm F/0.7 (spoiler, 70mm F/1 and 0.7x converter)
http://photoog.com/2008/06/worlds-fastest-lens-zeiss-50mm-f07/

"The engineering looks eloquent to me. The key idea is to make a supersized 70mm f/1 that lights a much larger image circle, and then design a “condenser” that brute-forces your way to 50mm f/0.7 by shortening the focal length and condensing the light. Basically it’s adding a 0.7x teleconverter that gives 1 f-stop."

 

[HoppyUK]

I've wondered when backlit will hit DSLRs, has it arrived yet? I quite liked the idea of that sensor Sigma use. Let me find a link...
http://www.sigma-sd.com/SD1Merrill/system.html

Seems sensible but I know it's got its drawbacks.

Yeah I didn't start it... haha. what I was trying to say that from the same distance, it doesn't matter what your focal length and aperture are, as long as the entrance pupil is the same diameter you'll get the same results when cropped to the same size.

Something else, while I am here, is the idea of a "tele-compressor". I think it is a GREAT idea. But why don't they exist? Eg, take a 0.63x converter on a EF lens and suddenly your crop sensor will no longer have a crop factor.

And before someone says it can't be done, read how they made a 50mm F/0.7 (spoiler, 70mm F/1 and 0.7x converter)
http://photoog.com/2008/06/worlds-fastest-lens-zeiss-50mm-f07/

"The engineering looks eloquent to me. The key idea is to make a supersized 70mm f/1 that lights a much larger image circle, and then design a “condenser” that brute-forces your way to 50mm f/0.7 by shortening the focal length and condensing the light. Basically it’s adding a 0.7x teleconverter that gives 1 f-stop."

I don't know why Sigma's Foveon sensor idea never caught on. Its advantages are not quite as clear cut as Sigma would have us believe, but I don't know if the lack of heavy weight development is because there are fundamental drawbacks or just that Sigma holds the patents. Just speculating, but I think the Sigma/Foveon thing had a lot to do with the late Yamaki senior's dream of becoming a mainstream camera maker and he wanted to hold the aces.

Either way, it's now moot as the CMOS-Bayer concept and it's various derivatives now seems to have overtaken it. Foveon's main advantage is resolution (and that's one of the things I was referring to when I said pixels and lpmm were hard to compare) and we now have more of that than we can really use. Canon seems to be saying with the 5D3, and I agree with them, that 20m-ish pixels is plenty. Lots of cameras now have that, and more.

I don't care much for your tele-compressor idea...

 

[woof woof]

No I don't think you understand. For the same photo, on a full frame you need 1.6 times the focal length, 1.6 times the aperture and 1.6*1.6 the ISO. This gives the same shutter speed for the same picture with the same signal to noise ratio. In theory.

I understand I was just trying to make the point that smaller formats can have an ISO or shutter speed advantage. Maybe it's just how my mind works as I usually shoot in aperture or manual and in reality I probably wouldn't have the same shutter speed or ISO when shooting with a smaller format because, for eg...

MFT 25mm, ISO 100, f4 = faster shutter speed than FF, 50 ISO 100 f8
That's all I wanted to say

 

[whom]

I understand I was just trying to make the point that smaller formats can have an ISO or shutter speed advantage. Maybe it's just how my mind works as I usually shoot in aperture or manual and in reality I probably wouldn't have the same shutter speed or ISO when shooting with a smaller format because, for eg...

MFT 25mm, ISO 100, f4 = faster shutter speed than FF, 50 ISO 100 f8
That's all I wanted to say

Ah right, and yes, you'd be right. I see now. But FF has pixels four times the area. It would require ISO 400 for the same shutter speed BUT because of the increased area that should (in my simple theory based world) results in the same levels of "noise".

 

[whom]

I don't know why Sigma's Foveon sensor idea never caught on. Its advantages are not quite as clear cut as Sigma would have us believe, but I don't know if the lack of heavy weight development is because there are fundamental drawbacks or just that Sigma holds the patents. Just speculating, but I think the Sigma/Foveon thing had a lot to do with the late Yamaki senior's dream of becoming a mainstream camera maker and he wanted to hold the aces.

Either way, it's now moot as the CMOS-Bayer concept and it's various derivatives now seems to have overtaken it. Foveon's main advantage is resolution (and that's one of the things I was referring to when I said pixels and lpmm were hard to compare) and we now have more of that than we can really use. Canon seems to be saying with the 5D3, and I agree with them, that 20m-ish pixels is plenty. Lots of cameras now have that, and more.

I don't care much for your tele-compressor idea...

Why don't like the compressor idea?! Imagine sticking a medium format lens with compressor on a Canon body and getting the same effect as the medium format would? A whole new world of possibilities.....

But wouldn't an idealised Foveon sensor have a significant noise advantage for the same megapixels? In a Bayer array all the red that lands on a green goes to waste etc etc. In the perfect Foveon case, all the light is absorbed. I know that this isn't the case because they aren't ideal, but you get my drift. I've not done a lot of reading on this so could be hopelessly misinformed.

 

[HoppyUK]

Ah right, and yes, you'd be right. I see now. But FF has pixels four times the area. It would require ISO 400 for the same shutter speed BUT because of the increased area that should (in my simple theory based world) results in the same levels of "noise".

Alan's caveat is the in 'reality' word, and he has a point. M4/3rds has some fundamental advanatges that can work for you. On the other hand, when push comes to shove at high ISO, larger sensors come to the fore.

Bottom line for me is the lens MTF thing, that works independently of the sensor and always favours larger ones, ie better sharpness.

Why don't like the compressor idea?! Imagine sticking a medium format lens with compressor on a Canon body and getting the same effect as the medium format would? A whole new world of possibilities.....

But wouldn't an idealised Foveon sensor have a significant noise advantage for the same megapixels? In a Bayer array all the red that lands on a green goes to waste etc etc. In the perfect Foveon case, all the light is absorbed. I know that this isn't the case because they aren't ideal, but you get my drift. I've not done a lot of reading on this so could be hopelessly misinformed.

Any kind of focal length or field of view converter has compromises, mostly bad ones. Why would you want to stick on a medium format lens and some wacky tele-compressor when you can have a nice EF-S lens and all the inherant advantages that go with that?

As I say, I don't know why Foveon hasn't caught on. And I guess we never will, because it would need a major manufacturer to take it on and throw a lot of R&D at it. Can't really see that happening with so much vested interest in CMOS-Bayer (and Sony seems to be assuming the role of purveyor of sensors to all and sundry, except Canon) though now might be the time given old Yamaki's recent passing. Maybe someone like Samsung could give Foveon a punt if it really has legs? Don't hold your breath.

 

[whom]

Any kind of focal length or field of view converter has compromises, mostly bad ones. Why would you want to stick on a medium format lens and some wacky tele-compressor when you can have a nice EF-S lens and all the inherant advantages that go with that?

Well if you got yourself a Hasselblad lens and stuck a 0.5x on it with a 5D you'd be looking at turning a 110mm f/2 into a 55mm f/1.

To be honest, I thought that there would be faster medium format lenses than there turns out to be.... Ok so it isn't THAT useful of an idea...

 

[danbroad]

I think that is wrong. The focal length divided by the f stop gives aperture diameter (as seen when looking at the front of the lens, see wikipedia "entrance pupil") and that is the only constant here. So increasing the focal length causes the f stop to increase by the same amount.

So a 17-55mm F/2.8 on a crop is the same as 27-88mm F/4.5.

To put it another way, if you want the same look of as 24-70mm F/2.8 gets you on a full frame you'd need a 15-44mm F/1.8 on your 1.6 crop....!

Its not wrong; lens is lens. You don't change the focal length of a lens no matter whether you mount it on full frame, APS-C or a shoebox. Your lens, with its optical constants, remains the same. You're discussing angle of view, which most certainly does change.

 

[arad85]

To be honest, I thought that there would be faster medium format lenses than there turns out to be.... Ok so it isn't THAT useful of an idea...

Why do you need fast lenses when you have that huge sensor to give you DoF...

 

[HoppyUK]

Well if you got yourself a Hasselblad lens and stuck a 0.5x on it with a 5D you'd be looking at turning a 110mm f/2 into a 55mm f/1.

To be honest, I thought that there would be faster medium format lenses than there turns out to be.... Ok so it isn't THAT useful of an idea...

You carry on Sounds like a horrendous idea to me!

The main reason we don't see lower f/numbers on medium format lenses is it's optically hard to deliver low f/numbers over the larger sensor area, apart from the size/weight issues. No surprise that Olympus do f/2 zooms for M4/3rds, and there are (affordable) sub f/1.0 primes for that format too.

PS Entrance pupil and aperture/diaphragm are not the same thing.

 

[whom]

Its not wrong; lens is lens. You don't change the focal length of a lens no matter whether you mount it on full frame, APS-C or a shoebox. Your lens, with its optical constants, remains the same. You're discussing angle of view, which most certainly does change.

I am discussing angle of view, depth of field and signal to noise ratio. It is a theoretical comparison. I know a lens doesn't magically change, what gave you the impression I thought it did? My point is if you think to yourself "oh, I use a crop camera which makes my 100mm F/2 lens a 160mm F/2 lens" you would be wrong in my eyes. I would correct you and say actually it is the same as a 160mm F/3.2 lens because the ENTRANCE PUPIL (aperture diameter) is physically the same and defined as focal length divided by F number. This is of course for identical framing and out-of-focusness.

Why do you need fast lenses when you have that huge sensor to give you DoF...

Well why not? Even SHALLOWER DoF Less is more.

 

[whom]

You carry on Sounds like a horrendous idea to me!

The main reason we don't see lower f/numbers on medium format lenses is it's optically hard to deliver low f/numbers over the larger sensor area, apart from the size/weight issues. No surprise that Olympus do f/2 zooms for M4/3rds, and there are (affordable) sub f/1.0 primes for that format too.

PS Entrance pupil and aperture/diaphragm are not the same thing.

Is the entrance pupil not what the size the aperture appears to be when looking from the front?

 

[danbroad]

I know a lens doesn't magically change, what gave you the impression I thought it did?

....the fact that your reply started with " I think that is wrong?" when I mentioned optical constants?:shrug:

Anyway, I think we're referring to the same point, which is that using your 100 as an example on a 2x crop (for ease of math) then yes, your resulting image will display apparent focal length and depth of field similar to a final image taken by a 200mm on full frame.... (In fact, the image produced is a perfect crop slice out of the middle of a full frame 100 image; changing the angle of view did that, nothing more).

However, and here's a big advantage of crop systems, the shutterspeeds will still be f/2 because light is light and the aperture is the aperture. IIRC, Bob Atkins wrote a good article on all this, I'll try and dig the link later.

 

[danbroad]

http://www.bobatkins.com/photography/tutorials/crop_sensor_cameras_and_lenses.html

...there it is, knew I had it somewhere. There's plenty of confusion over this all over, but fundamentally consider smaller-than-full-frame sensors to be a sort of 'specialist middle bit'. 300mm lenses do not suddenly become 810mm lenses on my V1, but the angle of view changes so that it takes less subject to fill the viewfinder frame. Instead of those 10MP covering a whole full frame angle of view, they now cover just the middle bit [which in the V1's case is 1/2.7th of the whole/full-frame picture]. It's a type of 'specialist digital zoom'.

Now, the depth of field at the same subject distance for any one lens is just the same regardless of the sensor because that is a lens/light characteristic... except now, the new '810mm' type angle means to get that same depth of field 'look' at 810mm full frame [should such a lens exist] would require an f/stop 2.7 times that of the 300mm lens I'm using.

e.g a 300/4 wide open looks like a 300/4 on full frame, but on the V1 the final image looks like I shot an 810mm lens at f/10.8 [or, realistically, I'd need a Sigma 800mm at f/11] on that same full frame camera. Nothing else has changed: you could get exactly the same effect by printing out the full frame image and cutting out the middle third.

 

[HoppyUK]

http://www.bobatkins.com/photography/tutorials/crop_sensor_cameras_and_lenses.html

...there it is, knew I had it somewhere. There's plenty of confusion over this all over, but fundamentally consider smaller-than-full-frame sensors to be a sort of 'specialist middle bit'. 300mm lenses do not suddenly become 810mm lenses on my V1, but the angle of view changes so that it takes less subject to fill the viewfinder frame. Instead of those 10MP covering a whole full frame angle of view, they now cover just the middle bit [which in the V1's case is 1/2.7th of the whole/full-frame picture]. It's a type of 'specialist digital zoom'.

Now, the depth of field at the same subject distance for any one lens is just the same regardless of the sensor because that is a lens/light characteristic... except now, the new '810mm' type angle means to get that same depth of field 'look' at 810mm full frame [should such a lens exist] would require an f/stop 2.7 times that of the 300mm lens I'm using.

e.g a 300/4 wide open looks like a 300/4 on full frame, but on the V1 the final image looks like I shot an 810mm lens at f/10.8 [or, realistically, I'd need a Sigma 800mm at f/11] on that same full frame camera. Nothing else has changed: you could get exactly the same effect by printing out the full frame image and cutting out the middle third.

That's kinda right Dan, but then it's not.

Nikon V1 has a crop factor of 2.7x, which is linear, therefore the area difference to full frame is 2.7 squared. Basically, you're not cutting out the 'middle third' with the V1, but an area less than 14% of FF. Your DoF calcs of f/4 vs f/10.8 are right though (as the f/ratio is already squared).

And depth of field is not just a lens/light characteristic. DoF is a concept that only exists in a final output image, ie print, of a specific size and viewed at a prescribed distance. The ISO standard that everyone uses is based on a 10in print viewed from a distance equal to the diagonal, ie about 12in, though it also works exactly pro-rata for other sizes if the viewing distance is adjusted.

Nikon V1 is interesting in the context of this thread though, as a good example of how the theoretical principles outlined by the OP can get overturned when a major manufacturer throws the kitchen sink at the development of a particular sensor/format.

I've used a V1 quite a bit, and I'm sure you'll agree that it punches way, way above its weight in terms of what we might expect from such a small sensor in terms of sharpness and ISO performance. I was frankly astonished by its image quality, and I don't think we've seen the last of what it can do with the massive hike in effective reach it can give, eg for birding.

 

[danbroad]

I was trying to keep it simple, for all concerned, including me... however, I do concede there are a couple of issues [the whole write up concerning DoF around the hyperfocal distance is particularly difficult to digest etc].

When referring to DoF I'm thinking more in terms of the final image, hence my last sentence. It's actually a real blessing for ultra-telephoto to have a little more leeway when focusing. I'm not sure how many shot's I'd lose with this kind of crop otherwise.

Anyway, come and join over at the V1 owner thread! I agree, it's a birder's secret weapon and more people are coming around to the idea as it gets a foothold...

 

[whom]

I was trying to keep it simple, for all concerned, including me... however, I do concede there are a couple of issues [the whole write up concerning DoF around the hyperfocal distance is particularly difficult to digest etc].

When referring to DoF I'm thinking more in terms of the final image, hence my last sentence. It's actually a real blessing for ultra-telephoto to have a little more leeway when focusing. I'm not sure how many shot's I'd lose with this kind of crop otherwise.

Anyway, come and join over at the V1 owner thread! I agree, it's a birder's secret weapon and more people are coming around to the idea as it gets a foothold...

Thanks for that link. I do understand what you are saying, I already understood what you are saying. But I only started this thread with one intention. What does it take to take an identical picture on a crop body as a full frame body. Not how does depth of field change with crop, what does it mean to have a smaller sensor etc etc.

I just wanted to reassure myself that the 5D2 I bought is worth the money EG I like having my 50mm F/1.4 on my full frame. On a crop I would need a 31mm F/0.88!!



Not this is for anyone confused reading this. My original point.

For matching picture on a full frame from the same location with the same shutter speed as you just took a picture with your crop camera...

Focal length*crop factor
F number*crop factor
ISO*crop factor*crop factor

If someone doesn't believe me, why not take two pictures at different settings (doesn't matter if you start with a crop already) using that written just above and see. It would be nice to see someone else's shots like mine on the first post.