[Ikarus]

Diffraction kicks in at ƒ5.6 on both APS-c and Full Frame

With this, you imply that diffraction is invariant to the f-stop across sensor sizes, but this is not the case. Rather, the diffraction-limited depth of field is constant for all sensor sizes, so for the same diffraction, the f-stop scales with the crop factor. My point is that FF has the same amount of diffraction as APS-C at the equivalent stop, but an 1.5^2 lower light intensity, thus ~ 1 stop of extra headroom. Of course, one could stop down the APS-C to compensate for that, but at the smaller apertures, diffraction will become more noticeable.

[jsherman999]

If you have more than one format you know how each performs under given conditions. The equivalence numbers are meaningless.
.

How do you just know how each performs? Underneath your 'just knowing' is an ad-hoc application of equivalence, gained from (likely) experience.

I think the difference is that you prefer to leave it as vague an estimate as possible, which is fine, but it doesn't make the definitions under it all 'meaningless', it just makes them unimportant to you.

As far as shooting goes, I rarely apply equivalence either, for what I do. If you're hitting a perf limit though and you're wondering if lens X on camera Y might give you better results for shooting situation Z, you'd be better off applying it more precisely before you buy. Or, not. Equations generally don't care if they're used.

Alan Oleson summed it up nicely on dp:

"We have just found a simple and practical tool to explain most of the differences which others seem to think are unexplainable and just some weird thing you have to accept when changing formats."


.

[normhead]

Well actually is was an observation from having looked at many lens charts... not theoretical at all, and certainly no implications. As far as i know theoretically, without diffraction, lenses would get sharper and sharper as the aperture got smaller and smaller, at least in terms of acceptable sharp[ness as it relates to DoF.

My point is that FF has the same amount of diffraction as APS-C at the equivalent stop, but an 1.5^2 lower light intensity, thus ~ 1 stop of extra headroom.

Maybe I'm just dense but, you're going to have to give me a practical example.

[Parallax]

How do you just know how each performs?
.

Not speaking for myself, but rather for those who have come far closer to mastering photography than I, it's the same way a chef "knows" how much salt, or garlic to add. The same way an Indy car driver "knows" how much throttle to give in turn 3. The same way a composer "knows" whether or not a G minor 7th will sound better than a G minor.
It's the difference between art and science.
I think equivalence is more important to people who are more concerned with the science of photography than with the art.

---

Last edited by Parallax; 03-22-2016 at 05:17 PM.

[Winder]

I think equivalence is more important to people who are more concerned with the science of photography than with the art.

For people who are much better at Photoshop or even On One type software I think that is true. One reason I like FF is that is it easier to get the look I want. The RAW files are much easier to work with... for me and my skill level. Flikr and 500px are both full of phenomenal images taken whit APS-C an processed by people who really know what they are doing with layers and brushes.

Here is Ricoh's take on the subject (Advantages) of FF vs. APS-C.
Expression / PENTAX K-1 Special site | RICOH IMAGING

[Parallax]

My above statement isn't about advantages or disadvantages. As Norm has pointed out more than once; every format has an advantage (or disadvantage) over every other format, depending upon the situation.

[todd]

ƒ22

try to stay below f/11 if at all possible.

Have you guys tried deconvolution sharpening? I've been playing with it and am pretty convinced that it makes diffraction a non-issue.

[normhead]

OK, then tell me about it, I'm relatively clueless.

[todd]

OK, then tell me about it, I'm relatively clueless.

Here are the technical details. https://en.wikipedia.org/wiki/Richardson%E2%80%93Lucy_deconvolution

I've been testing it in Rawtherapee since being told it can be used to treat diffraction in another thread here on the board.

Rawtherapee wiki on it

I will give you a demo later when I have time as I don't like the one I posted in the other thread well enough. (I have another F22 image that will demonstrate it better, and I need to redo it without doing any other processing.)

[osv]

My point is that FF has the same amount of diffraction as APS-C at the equivalent stop, but an 1.5^2 lower light intensity, thus ~ 1 stop of extra headroom. Of course, one could stop down the APS-C to compensate for that, but at the smaller apertures, diffraction will become more noticeable.

"equivalent stop" being correctly defined there as equivalent dof, because when the actual dof is the same, so is the diffraction(ignoring the effects of pixel density on diffraction visibility).

it's an extra stop of headroom in terms of the numbers on the aperture scale of the lens.

[derekkite]

I'm putting this as a sticky on my desk top, so i never have to write it out again, anyone who'd like to suggest improvements feel free.

Equivalence

There are three components that affect an image. Field of view, and shutter speed and Depth of field. To produce an “equivalent image” all must be the same.

If we start with an image taken with an APS_c camera taken at the following settings.

ƒ5.6 , 100 ISO, 1/100s

To take the same image on a full frame sensor, the first thing you need to do is normalize the depth of field to ƒ8. If you don’t normalize the DoF you are taking different images an no comparison can be made.

Changing the shutter speed can affect motion blur in moving objects so to have the same image the shutter speed should be the same.

so we now have
ƒ8 to normalize DoF,
1/100s to normalize an motion blur

to compensate for stopping down to ƒ8 we have to increase the ISO of the full frame image, also one stop.
but in doing so, we create more noise, and let in half the light.

So the equivalent exposure on the FF will be ƒ8, 1/100s 200 ISO.

At 200 ISO the image will be created with half the total light that the 100 ISO image was created with, on the Full Frame, which happens to be the same amount of total light that the 100 ISO image will be created with on APS-c.

In the end, for an image taken with the same DoF, the same Field of View, the same time of exposure… the amount of light collected by the sensor will be the same, and the noise will be the same.

For any image that can be taken by both cameras, ( and that is about 87.5% if you count the 8 stops from ƒ1.4 to ƒ22 as your possibilities ) of all possible images, everything is the same to produce the same image.
Noise is the same
Total light is the same.
Depth of field is the same
Field of View is the same.
Shutter speed is the same.
Low light performance is the same.

That is equivalence.

The advantage to the full frame sensor is that it has one stop of narrow Depth of Field that cannot be taken on APS-c.

The advantage to APS-c is if and only if the Megapixels of both cameras are the same, there is 1.5 times more magnification with long glass or macro. That is the only way you get it all. The K-1 partially makes up for that by having 36 MP, but not completely. To completely make up the advantage the FF would have to be 51 MP as opposed to a K-3 24MP image. But the extra megapixels of the K-1 does negate any advantage of using a K-5 or any 16 MP camera, because it's the same number of megapixels in crop mode that the 16 megapixel camera is.

To get the thing that a camera excels at on one format, you have to give that up if you switch to the other format., which is why the majority of us who are contemplating a K-1 are contemplating running the two formats side by side.

That will be true, unless the pixels on the two different formats are the same size and efficiency, so that when you crop the larger format, you get exactly the same image you'd get on the smaller format. That would make the smaller format redundant, although possibly still desirable for size and weight considerations.

Norm, look at the depth of field charts. DOF doesn't change that much, but what changes is the proportion of the image that is in focus. If you have a scene where you want 2/3 of the depth in focus, yes you need to stop down on the K1, but if you want 28" in focus to isolate something, then the same aperture would give you the results you want.

It is the field of view that is very different. Your wonderful 200mm shots would need 300mm to be the same. My 500mm would be somewhere around 400mm or less in field of view terms, meaning something small on the K3 would be very small on the K1.

But if we manage to fill the frame somehow; either longer lens, closer in, then there is many more pixels and all the nice iso improvements and dynamic range to be enjoyed.

You don't buy the K1 to get the same shots better. You get the K1 to get different shots. I'm interested in the improvements in focus.

As a friend says, we really need one of each.

[Ikarus]

"equivalent stop" being correctly defined there as equivalent dof, because when the actual dof is the same, so is the diffraction(ignoring the effects of pixel density on diffraction visibility).

it's an extra stop of headroom in terms of the numbers on the aperture scale of the lens.

I'm not sure you're agreeing or disagreeing, but this is just what I said. My point is that the FF sensor still has one extra stop of headroom due to the light intensity being lower than on the APS-C sensor by a factor of 1.5^2 at a given equivalent exposure. The only reason I brought up diffraction is because I anticipated the counter-argument that one always has the option to stop down the APS-C sensor further in order to prevent the clipping of highlights.

[osv]

i think that we agree? i'm just clarifying the dof/diffraction portion of it for the normheads on this forum who make ridiculous claims like "Diffraction kicks in at ƒ5.6 on both APS-c and Full Frame"

that statement is totally wrong, as the cambridge diffraction calculator proves.

i find it rather fascinating that after so many threads on equivalence, people still come up with stuff like that.

[Ian Stuart Forsyth]

Norm, look at the depth of field charts. DOF doesn't change that much, but what changes is the proportion of the image that is in focus. If you have a scene where you want 2/3 of the depth in focus, yes you need to stop down on the K1, but if you want 28" in focus to isolate something, then the same aperture would give you the results you want.
It is the field of view that is very different. Your wonderful 200mm shots would need 300mm to be the same. My 500mm would be somewhere around 400mm or less in field of view terms, meaning something small on the K3 would be very small on the K1.
But if we manage to fill the frame somehow; either longer lens, closer in, then there is many more pixels and all the nice iso improvements and dynamic range to be enjoyed.
You don't buy the K1 to get the same shots better. You get the K1 to get different shots. I'm interested in the improvements in focus.
As a friend says, we really need one of each.

The difference we perceive in DOF when opening up one stop usually has very little consequences to the success in the final image in the DOF we are interesting in keeping IF, where we see the difference in blur is the ability to isolate the subject from the background better.
images taken at F/6.3 and F/4


I for one would give up this for 2.25 times fast shutter speed or 2.25 times more light.

I always find it interesting in that many chase the reach factor ( nothing wrong with this) at this point in the game is not that much of a difference between 36 FF and 24 cropped, when you consider how much you have to taxing the image circle of a lens for this little added pixel density its really at diminished returns. When we look at what we are trying to tax from the image circle to convert a 300 mm lens to a 530 equivalent FOV with a 24mp sensor to that of a 450 fov with a 16mp you don't get a heck of a lot more with 24mp. Where I would sooner invest my resources is at the greatest return on investment and that is at the other end of the spectrum where the lens is its sharpest ( using as much of the image circle as possible). Here you will see much better IQ when you are not FL limited and in places where I lose the reach with cropping a lens to a FOV @ 1.7x factor, that IQ is seldom the IQ I am chasing

[normhead]

Norm, look at the depth of field charts. DOF doesn't change that much, but what changes is the proportion of the image that is in focus. If you have a scene where you want 2/3 of the depth in focus, yes you need to stop down on the K1, but if you want 28" in focus to isolate something, then the same aperture would give you the results you want.

My own interpretation of both the DoF charts and shooting in the field under controlled conditions with a ruler in the frame to confirm DoF would suggest different.

I'm willing to go along with the notion that the larger format might give you a slightly different image, I'm sure it does, and you might want to have both.... in any case, but I'm a little hesitant to accept your statement that DoF doesn't change much, since one of the things I did while formatting this way of presenting the data was set up a ruler beside my flower garden and take a 50mm shot and a 35mm image, one representing a D810 and one representing a K-5. The change in DoF for the same aperture was almost double for the APS-c set up. That was close to minimum focus, so maybe not as relevant further away from the camera but to just say it isn't much, well that's possibly not true.

So I already did look at the DoF charts to confirm what I posted. I'm not sure what you want me to see.

I detect you're getting at something, I'm just not sure what it is. It blows me away you guys with both formats don't have images to demonstrate what you're talking about.

I've always said, if you can't fake a stop, you're not a photographer. So a one stop difference usually makes no difference to the final image, going in one direction or the other.

The difference between APS-c and FF is a little over 1 stop or 2.25 times if you want to get all dramatic. Hence my opinion that there isn't much difference between them.

I suspect that's why there are not a lot of comparison images that even show the difference between APS-c and FF. I suspect it's really hard to tell the difference.