Exposure.

Steen Bay said:

Steen Bay said:

Ah, I see your point. I'm not sure how to answer that, other to say that the same rules should apply. If the the total light is the same but the image is' stretched' to correct for the distortion, then the area is changing and by definition there should be some amount of fall off. I don't have enough expertize to determine how much, or even to prove if my reasoning is valid to begin with, but welcome anyone else who does!

My original post was in reference to a vary basic optical system, trying to dispel the misnomer of the inverse square law breakdown for close objects. If what he said was true, then there would always be an enormous and unavoidable (even at small apertures) fall off on every lens. For example, for a 35mm lens on FF, the AOV to the long edge is something like 54 deg. If my calculations are correct the distance to the edges would be something like 40% longer than to the center, leading to over 50% fall off in light at the long edge!

Click to expand...

Don't think that your calculations are correct. In the OP Joe says 5 feet to the center and 6.6 feet to the corners (32% longer distance) with a 25mm lens (on FF I suppose). Not quite sure, but I think that what he's arguing is pretty much the same as my point about rectilinear lenses. But it has nothing to do with the distance to the subject (like Joe says), only the FoV/AoV is important here. The wider the lens, the more the image will be 'stretched' with a rectilinear lens (in order to keep straight lines straight), and the more it will 'mess' with the inverse square law.

Click to expand...

I wasn't referencing Joe's example in my calculation, just an arbitrary one. Was my calculations correct for the estimated loss at the edges for my example?

In any case, compare the slight change in the area due to the distortion (stretch) for a typical rectilinear lens and you will probably find it to be not very significant (just a guess). Perhaps its just a matter of trading rectilinear image performance at the expense of slight light loss at the edges.

I don't believe this was Joe's claim though. The principle that Joe was siting (I believe) is one for non-optical systems (no lenses). If you are simply measuring the light falling on an object using a light meter (Illuminance), your result will be skewed as move closer to a non-point source. It doesn't apply for optical systems.

Andre Affleck said:

Steen Bay said:

Andre Affleck said:

Ah, I see your point. I'm not sure how to answer that, other to say that the same rules should apply. If the the total light is the same but the image is' stretched' to correct for the distortion, then the area is changing and by definition there should be some amount of fall off. I don't have enough expertize to determine how much, or even to prove if my reasoning is valid to begin with, but welcome anyone else who does!

My original post was in reference to a vary basic optical system, trying to dispel the misnomer of the inverse square law breakdown for close objects. If what he said was true, then there would always be an enormous and unavoidable (even at small apertures) fall off on every lens. For example, for a 35mm lens on FF, the AOV to the long edge is something like 54 deg. If my calculations are correct the distance to the edges would be something like 40% longer than to the center, leading to over 50% fall off in light at the long edge!

Click to expand...

Don't think that your calculations are correct. In the OP Joe says 5 feet to the center and 6.6 feet to the corners (32% longer distance) with a 25mm lens (on FF I suppose). Not quite sure, but I think that what he's arguing is pretty much the same as my point about rectilinear lenses. But it has nothing to do with the distance to the subject (like Joe says), only the FoV/AoV is important here. The wider the lens, the more the image will be 'stretched' with a rectilinear lens (in order to keep straight lines straight), and the more it will 'mess' with the inverse square law.

Click to expand...

I wasn't referencing Joe's example in my calculation, just an arbitrary one. Was my calculations correct for the estimated loss at the edges for my example?

Click to expand...

I think it's more like 12% longer distance to the long edges (not corners) in your example.

Andre Affleck said:

In any case, compare the slight change in the area due to the distortion (stretch) for a typical rectilinear lens and you will probably find it to be not very significant (just a guess). Perhaps its just a matter of trading rectilinear image performance at the expense of slight light loss at the edges.

Click to expand...

If shooting with the 16-35mm on FF the effect isn't significant at the 'long' end, but at 16mm it is.

Andre Affleck said:

I don't believe this was Joe's claim though. The principle that Joe was siting (I believe) is one for non-optical systems (no lenses). If you are simply measuring the light falling on an object using a light meter (Illuminance), your result will be skewed as move closer to a non-point source. It doesn't apply for optical systems.

Click to expand...

Let's hope then that Joe will clarify what his claim was. ;-)

Steen Bay said:

Steen Bay said:

In any case, compare the slight change in the area due to the distortion (stretch) for a typical rectilinear lens and you will probably find it to be not very significant (just a guess). Perhaps its just a matter of trading rectilinear image performance at the expense of slight light loss at the edges.

Click to expand...

If shooting with the 16-35mm on FF the effect isn't significant at the 'long' end, but at 16mm it is.

Click to expand...

Someone should do a test, not unlike what I proposed in my OP.

Steen Bay said:

Steen Bay said:

I don't believe this was Joe's claim though. The principle that Joe was siting (I believe) is one for non-optical systems (no lenses). If you are simply measuring the light falling on an object using a light meter (Illuminance), your result will be skewed as move closer to a non-point source. It doesn't apply for optical systems.

Click to expand...

Let's hope then that Joe will clarify what his claim was.

Click to expand...

Here's what joe claimed (which can be found in the OP):

As we move away from a uniformly lit scene (regardless of whether the light is emitted or relected), the amount of light passing though any given area in any given time interval is decribed by an inverse square relationship. I used this principle for my star example above.

This relationship is not exact, of course, since if we double our distance to the scene, we do not double our distance from every point in the scene. However, the reality is that the assumption of "uniformly lit" causes far more problems with applying theory to practice than this issue.

There is also the issue of background light. Let's say I take a pic of a doll on a chair. A manly, doll, mind you, 'cause I'm a manly guy (Steen will vouch for me).

Clearly, if we double our distance from the doll, we do not double our distance to the background, or even close to it. However, what happens is that, for a given framing, the proportion of the background that we see can be described with an inverse square relationship, so the same rule applies albeit for a different reason.

Feel free to consult your bible if you don't believe me.

Great Bustard said:

...to understand, just the first one will do. And I didn't resort to saying that anyone has to merely take my word for it anywhere , and resent the implication you made. In fact, I took great pains to work out all the mathematics.

So, yeah, poor form on your part Jay. Really poor form.

Click to expand...

Wow! Implications I made? I had merely stated that over and over I had cited other web sites, definitions, and magazine articles supporting my claims. I kept asking over and over why this and why that. All I saw from you was restatements of the same theory, which still happens to be wrong. I am so sorry you still don't see that, and I am sorry if you see my statements as implications that you need to take offense to. They were not intended that way.

Jay A said:

Jay A said:

...to understand, just the first one will do. And I didn't resort to saying that anyone has to merely take my word for it anywhere , and resent the implication you made. In fact, I took great pains to work out all the mathematics.

So, yeah, poor form on your part Jay. Really poor form.

Click to expand...

Wow! Implications I made? I had merely stated that over and over I had cited other web sites, definitions, and magazine articles supporting my claims. I kept asking over and over why this and why that. All I saw from you was restatements of the same theory, which still happens to be wrong. I am so sorry you still don't see that, and I am sorry if you see my statements as implications that you need to take offense to. They were not intended that way.

Click to expand...

Here's what you said:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36787083

While I feel that Great Bustard is also very knowledgeable, all he has done is bend anything thrown at him to fit his premise which happens to be wrong. It's almost as if, you catch him off guard and he will use that same premise to once again make what you have pointed out to fit to it, without ever accepting the fact that we have already stated that we have dismissed that premise to begin with. It's almost like a conversation going as follows,
"this is so because the bible says so."
" Well, I don't believe in the bible, so how do you explain THIS."
" Well, because the bible says so."

How else was I supposed to take it? Link and quote anyplace where I acted anywhere like you describe. In fact, Andre asked:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36781967

If the light desity remains the same with distance, why does a star appear dimmer than the sun?

and I gave a very detailed and direct answer:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36782088

and even cast doubt on my own response and gave a better answer further down:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36783718

So, please, tell me how else I was supposed to take your comments, and link and quote where I said anything like what you described anywhere .

Umm, ok Jay, but what did you think about my post?

Great Bustard said:

Steen Bay said:

Great Bustard said:

In any case, compare the slight change in the area due to the distortion (stretch) for a typical rectilinear lens and you will probably find it to be not very significant (just a guess). Perhaps its just a matter of trading rectilinear image performance at the expense of slight light loss at the edges.

Click to expand...

If shooting with the 16-35mm on FF the effect isn't significant at the 'long' end, but at 16mm it is.

Click to expand...

Someone should do a test, not unlike what I proposed in my OP.

Click to expand...

A simple brick-wall test is the perfect test to perform here, and that has been done lots of times, and it's not even necessary to do the test, since it's pretty easy to calculate how much the bricks in the corners are stretched/enlarged if the wall is shot with a perfectly corrected rectilinear lens. Well, that is, it's a relatively simple calculation if you have the right equations (and knowlegde), which I haven't, so maybe you could do the math?

The point here is of course, that if shot with a rectilinear lens (with sensor parallel to the wall), then all the bricks will have the same size in the image, meaning that with a UWA lens like 16mm on FF, then the corner bricks will be enlarged by a considerable amount (because they are longer away from the camera than the center bricks), and therefore lose brightness. So the question is, how much brightness is lost in the corners with e.g. 16mm and 35mm rectilinear lenses on FF? (Distance to the wall, and the actual FL and sensor size isn't important here, only FoV/AoV matters)

Great Bustard said:

Great Bustard said:

Great Bustard said:

I don't believe this was Joe's claim though. The principle that Joe was siting (I believe) is one for non-optical systems (no lenses). If you are simply measuring the light falling on an object using a light meter (Illuminance), your result will be skewed as move closer to a non-point source. It doesn't apply for optical systems.

Click to expand...

Let's hope then that Joe will clarify what his claim was.

Click to expand...

Here's what joe claimed (which can be found in the OP):

As we move away from a uniformly lit scene (regardless of whether the light is emitted or relected), the amount of light passing though any given area in any given time interval is decribed by an inverse square relationship. I used this principle for my star example above.

This relationship is not exact, of course, since if we double our distance to the scene, we do not double our distance from every point in the scene. However, the reality is that the assumption of "uniformly lit" causes far more problems with applying theory to practice than this issue.

There is also the issue of background light. Let's say I take a pic of a doll on a chair. A manly, doll, mind you, 'cause I'm a manly guy (Steen will vouch for me).

Clearly, if we double our distance from the doll, we do not double our distance to the background, or even close to it. However, what happens is that, for a given framing, the proportion of the background that we see can be described with an inverse square relationship, so the same rule applies albeit for a different reason.

Feel free to consult your bible if you don't believe me.

Click to expand...

No controversial claims here ("manly guy", excepted), so I'll just take your word for it. ;-)

Great Bustard said:

So, please, tell me how else I was supposed to take your comments, and link and quote where I said anything like what you described anywhere .

Click to expand...

Joe, you are right. I think your first post brilliantly answered the original issue of that thread. It was the second issue of object brightness where things got a little screwy IMO.

I think Jay's frustration stems from the fact that instead of admitting that you made incorrect statements, you changed the boundary conditions of the premises to account for your errors instead of just admitting you misspoke. Even if you dispute this, this is the impression you have left.

Look, everyone on this thread has made mistakes, including myself. You leave a far better impression admitting that you misspoke than by trying to backpedal and insist that the error doesn't exist or making it right by changing it's intent.

I am glad you admitted to "casting doubt" on your other response, but frankly, you should have cast a few more, and if you don't think so, I can gladly point you to those statements. And while I'm at it, I can point you to several of my own I should have cast doubt on. This is very tricky subject matter and using the right terms is critical to conveying the correct relationship between principles. This should be a discussion, not an ego contest.

Steen Bay said:

The point here is of course, that if shot with a rectilinear lens (with sensor parallel to the wall), then all the bricks will have the same size in the image, meaning that with a UWA lens like 16mm on FF, then the corner bricks will be enlarged by a considerable amount (because they are longer away from the camera than the center bricks)

Click to expand...

You have to be careful here. You are not considering perspective when you make these claims. I believe he intent of a rectilinear lense is to correct for barrel distortion and pin cushioning. The bricks in the corners do not need to be the same size if the perspective accounts for it, they just have to be rectangular (even though the perspective shifts this too).

Steen Bay said:

Steen Bay said:

Here's what joe claimed (which can be found in the OP):

As we move away from a uniformly lit scene (regardless of whether the light is emitted or relected), the amount of light passing though any given area in any given time interval is decribed by an inverse square relationship. I used this principle for my star example above.

Click to expand...

Steen Bay said:

This relationship is not exact, of course, since if we double our distance to the scene, we do not double our distance from every point in the scene. However, the reality is that the assumption of "uniformly lit" causes far more problems with applying theory to practice than this issue.

Click to expand...

Click to expand...

Why do you assume a uniformily lit scene in the first place? It makes no difference how uniform the scene is lit. The relationship is maintained regardless. You make this contingency in several places within your essay (very well done BTW).

Steen Bay said:

Steen Bay said:

There is also the issue of background light. Let's say I take a pic of a doll on a chair. A manly, doll, mind you, 'cause I'm a manly guy (Steen will vouch for me).

Clearly, if we double our distance from the doll, we do not double our distance to the background, or even close to it. However, what happens is that, for a given framing, the proportion of the background that we see can be described with an inverse square relationship, so the same rule applies albeit for a different reason.

Click to expand...

Click to expand...

Where is the part about the corners of the frame?

Steen Bay said:

Steen Bay said:

Feel free to consult your bible if you don't believe me.

Click to expand...

Click to expand...

Well I will consulted with your essay instead , which I think you used to draw your original statement.

As it turns out, for the same f-ratio and shutter speed, the density of the light falling on the sensor (exposure) will be the same regardless of framing, perspective, and format (for a uniformly lit scene with a framing such that the distance from the corners from the scene are "close" to the distance from the center of the scene -- 24mm EFL and longer will do nicely). /

Unless I am mistaken, we were talking about parts of the scene (corners and center) where distances to the camera are different, and it's affects on density of light at the sensor. You state this as an exception due to the inverse square law rule, and I don't think that is correct. If you measure the amount of light from the corner, it will indeed be about 60% less (for 35mm lens) than the from the center (due to the inverse square law). However, this difference is not seen at the sensor (even at very small apertures). The question is why?

The following is the very first question asked by the OP of the original thread:

Great Bustard said:

Great Bustard said:

Great Bustard said:

I'm tempted to think that a lens at 35mm f/2.8 would pass less light than at 50mm f/2.8 because at 35mm the aperture will be smaller (35mm/2.8 vs 50mm/2.8). Yet when I compose a scene with the same FOV and lighting, but change focal length, the meter calculates the same shutter speed regardless.

Click to expand...

Click to expand...

Click to expand...

By saying that "at 35mm the aperture will be smaller" he is stating that he realizes that even though both lenses are set at F2.8, there is still a smaller diameter to the diaphragm when shooting at this F-stop with a 35mm lens as compared to a 50. He is merely asking why that is...or I should say he KNOWS why that is but is asking why that difference in brightness happens. At least this has been my contention from day one. Some (myself included at first) tried to explain by stating that the F-stop is the result of dividing the aperture (or the diaphragm) into the focal length. But here too, the OP stated that he knew that but wondered why the difference in light happened to begin with. Since a different F-stop will mean a different brightness in the final image, I interpreted that question to mean "what causes that difference in brightness?"

The basic explanation that I kept arguing against was that it was the result of the inverse square law. I never did intend to argue about the inverse square law having an affect on light as it pertains to area of an object, either itself or on the sensor. But I DID keep arguing that the inverse square law was not responsible for making the object get dimmer as its light rays traveled to the camera. The law has nothing to do with the apparent brightness of the subject, and increasing distance to that subject does not affect that brightness. Andre's example photos of the reflectors support this. If they got dimmer as they traveled towards the camera, then how could 2 of them, equal size, equal lighting, but different distance both be the same brightness at the sensor?

That said, I kept seeing over and over again, mathematical examples and carefully worded statements seemingly meant to support your dispute. It was almost like, no matter what examples were used to disprove your premise, you would just come up with more mathematics and theory to support it. I finally came to the conclusion that either you were being very stubborn or that both you and I were talking about 2 different things.

I had admitted several times that I myself just did not understand where some of the responses were coming from and thought that perhaps two different arguments were being discussed. Finally it did dawn on me that some of you were talking about area and not brightness. I think Andre was the one who pointed this out. However, each time I tried to say this, I kept getting more statements saying that no I was wrong and the others were right, with more mathematics and statements illustrating something totally different than that which I was arguing.

I suppose that's one of the problems with trying to communicate in these forums, where statements get responded to in one area of a thread without the person responding realizing that that statement was otherwise addressed elsewhere in the thread.

Andre Affleck said:

Great Bustard said:

So, please, tell me how else I was supposed to take your comments, and link and quote where I said anything like what you described anywhere .

Click to expand...

Joe, you are right. I think your first post brilliantly answered the original issue of that thread. It was the second issue of object brightness where things got a little screwy IMO.

Click to expand...

Wait, wait, wait. You are saying all those comments were solely about the subthread beginning here:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36781967

and had nothing to do with the OP, and that we're all square on what was said in the OP? That's news to me! Do you know why it's news to me? Because no link, no quote, no nothing , accompanied the statement accusing me of saying "it's so because the bible said".

Andre Affleck said:

I think Jay's frustration stems from the fact that instead of admitting that you made incorrect statements, you changed the boundary conditions of the premises to account for your errors instead of just admitting you misspoke. Even if you dispute this, this is the impression you have left.

Look, everyone on this thread has made mistakes, including myself. You leave a far better impression admitting that you misspoke than by trying to backpedal and insist that the error doesn't exist or making it right by changing it's intent.

I am glad you admitted to "casting doubt" on your other response, but frankly, you should have cast a few more, and if you don't think so, I can gladly point you to those statements. And while I'm at it, I can point you to several of my own I should have cast doubt on. This is very tricky subject matter and using the right terms is critical to conveying the correct relationship between principles. This should be a discussion, not an ego contest.

Click to expand...

Tell you what, I'm going to take you up on your offer. Please link and quote those statements, because as I demonstrated here:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36789701

you most certainly ascribed statement to me that I definitely did not say, or imply.

This request has two parts. The first is to link and quote any statement I made that you can demonstrate to be in error. The second is to link and quote any statement I have made where I said anything like "Well, because the bible says it's so".

So, no, it's not an ego contest, but a discussion where people misrepresent what I said (link above) and a discussion where people say that I dodge questions, well, that's not really conducive for anything constructive.

Andre Affleck said:

Steen Bay said:

The point here is of course, that if shot with a rectilinear lens (with sensor parallel to the wall), then all the bricks will have the same size in the image, meaning that with a UWA lens like 16mm on FF, then the corner bricks will be enlarged by a considerable amount (because they are longer away from the camera than the center bricks)

Click to expand...

You have to be careful here. You are not considering perspective when you make these claims. I believe he intent of a rectilinear lense is to correct for barrel distortion and pin cushioning. The bricks in the corners do not need to be the same size if the perspective accounts for it, they just have to be rectangular (even though the perspective shifts this too).

Click to expand...

What we normally call barrel 'distortion' on a WA lens isn't really distortion (but 'pincushion distortion' is). The images we get from a fisheye lens (with lots of barrel 'distortion'), that's simply how things look, that's how the natural perspective actually is, according to the inverse square law. But for some reason it seems that our brain prefers the 'corrected' (or maybe 'distorted' is the right word) images that a rectilinear lens gives us. That's what a rectilinear lens does, it messes with the natural perspective (and the inverse square law), in order to keep straight lines straight.

I think you misunderstood the original question...

Jay A said:

Jay A said:

Jay A said:

Jay A said:

I'm tempted to think that a lens at 35mm f/2.8 would pass less light than at 50mm f/2.8 because at 35mm the aperture will be smaller (35mm/2.8 vs 50mm/2.8). Yet when I compose a scene with the same FOV and lighting, but change focal length, the meter calculates the same shutter speed regardless.

Click to expand...

Click to expand...

Click to expand...

By saying that "at 35mm the aperture will be smaller" he is stating that he realizes that even though both lenses are set at F2.8, there is still a smaller diameter to the diaphragm when shooting at this F-stop with a 35mm lens as compared to a 50. He is merely asking why that is...or I should say he KNOWS why that is but is asking why that difference in brightness happens.

Click to expand...

My question was why would 35mm f/2.8 meter the same shutter speed as 50mm f/2.8 for the same scene and composition. That doesn't seem like I said anything about a "difference in brightness" to me. I was thinking that perhaps 50/2.8 would meter a faster shutter speed as the diaphragm is larger. (I know now why that is not the case.)

Jay A said:

At least this has been my contention from day one. Some (myself included at first) tried to explain by stating that the F-stop is the result of dividing the aperture (or the diaphragm) into the focal length. But here too, the OP stated that he knew that but wondered why the difference in light happened to begin with. Since a different F-stop will mean a different brightness in the final image, I interpreted that question to mean "what causes that difference in brightness?"

Click to expand...

I never said anything about using different f-stops. As for brightness, if anything I was wondering why 35/2.8 would yield the same brightness as 50/2.8.

Steen Bay said:

Great Bustard said:

Someone should do a test, not unlike what I proposed in my OP.

Click to expand...

A simple brick-wall test is the perfect test to perform here, and that has been done lots of times, and it's not even necessary to do the test, since it's pretty easy to calculate how much the bricks in the corners are stretched/enlarged if the wall is shot with a perfectly corrected rectilinear lens. Well, that is, it's a relatively simple calculation if you have the right equations (and knowlegde), which I haven't, so maybe you could do the math?

The point here is of course, that if shot with a rectilinear lens (with sensor parallel to the wall), then all the bricks will have the same size in the image, meaning that with a UWA lens like 16mm on FF, then the corner bricks will be enlarged by a considerable amount (because they are longer away from the camera than the center bricks), and therefore lose brightness. So the question is, how much brightness is lost in the corners with e.g. 16mm and 35mm rectilinear lenses on FF? (Distance to the wall, and the actual FL and sensor size isn't important here, only FoV/AoV matters)

Click to expand...

PS - Maybe the light loss in the corners is even worse than I thought, and maybe the calculation is a bit more difficult too, because, not only are the corner bricks longer away from the camera (and therefore smaller) than the center bricks, but the (flat) corner bricks are also seen from an angle, which makes the 'effective' size/area even smaller.. no wonder that it's difficult to make rectilinear WA/UWA lenses with sharp corners and without vignetting!

Steen Bay said:

PS - Maybe the light loss in the corners is even worse than I thought, and maybe the calculation is a bit more difficult too, because, not only are the corner bricks longer away from the camera (and therefore smaller) than the center bricks, but the (flat) corner bricks are also seen from an angle, which makes the 'effective' size/area even smaller.. no wonder that it's difficult to make rectilinear WA/UWA lenses with sharp corners and without vignetting!

Click to expand...

Actually, I've been thinking about that. It's not merely that the corners appear dimmer because they are further away, but that a rectilinear projection represents them with greater area than it does the center.

That is, a brick in the corner is not only further away from the aperture, but it takes up more space on the sensor, which means that the light is diffused more.

JackM said:

JackM said:

JackM said:

f anything I was wondering why 35/2.8 would yield the same brightness as 50/2.8.

Click to expand...

Click to expand...

Click to expand...

Sorry if I misunderstood, but isn't that basically what I said? You were wondering why the same brightness occurred if the aperture changed (the diaphragm opening not the F-stop). No?

Great Bustard said:

Steen Bay said:

PS - Maybe the light loss in the corners is even worse than I thought, and maybe the calculation is a bit more difficult too, because, not only are the corner bricks longer away from the camera (and therefore smaller) than the center bricks, but the (flat) corner bricks are also seen from an angle, which makes the 'effective' size/area even smaller.. no wonder that it's difficult to make rectilinear WA/UWA lenses with sharp corners and without vignetting!

Click to expand...

Actually, I've been thinking about that. It's not merely that the corners appear dimmer because they are further away, but that a rectilinear projection represents them with greater area than it does the center.

That is, a brick in the corner is not only further away from the aperture, but it takes up more space on the sensor, which means that the light is diffused more.

Click to expand...

The center and corner bricks have the same size in the image (and on the sensor), if the wall is shot with a perfectly corrected rectilinear lens, but the corner bricks are stretched/enlarged, and they are 'made' out of less photons than the center bricks, meaning that both brightness and sharpness are lost.

Jay A said:

Jay A said:

Jay A said:

Jay A said:

f anything I was wondering why 35/2.8 would yield the same brightness as 50/2.8.

Click to expand...

Click to expand...

Click to expand...

Sorry if I misunderstood, but isn't that basically what I said? You were wondering why the same brightness occurred if the aperture changed (the diaphragm opening not the F-stop). No?

Click to expand...

you were talking about the brightness changing. (shrug)

For those who are interested: There is an interesting related discussion going on in another forum.

http://www.largeformatphotography.info/forum/showthread.php?t=68275

It raises some interesting questions. For example:

1. Assume that you are in a room with black walls. You are sitting entirely in the sunlight (or light from the open sky) coming through the window. Does the amount of light hitting you depend on how far you are from the window? Why?

2. Does the size of the window matter? (Assuming it's large enough to completely cover you).

Please note how civil the discussion is over there. I hesitated posting this here for fear that some of you might misunderstand some of their comments and create insulting posts in their forum. It would be embarrassing if some of the recent personal attacks from this thread spilled over into their forum.

Steen Bay said:

Steen Bay said:

Actually, I've been thinking about that. It's not merely that the corners appear dimmer because they are further away, but that a rectilinear projection represents them with greater area than it does the center.

That is, a brick in the corner is not only further away from the aperture, but it takes up more space on the sensor, which means that the light is diffused more.

Click to expand...

The center and corner bricks have the same size in the image (and on the sensor), if the wall is shot with a perfectly corrected rectilinear lens, but the corner bricks are stretched/enlarged, and they are 'made' out of less photons than the center bricks, meaning that both brightness and sharpness are lost.

Click to expand...

I'm a bit confused here. If the corner bricks are strectched/enlarged compared to the bricks in the center, how are they taking up the same size in the image and on the sensor?

I'll have to find a brick wall to take a pic of, I guess.