Does crop "bring you closer" ?-- test inside 400mm 5.6

I read your post and was just clarifying for you that the 5DII's viewfinder is indeed no where near 1.6x larger and is in fact almost the same size, thus the common claim that full frame bodies have much larger viewfinders is an over statement.

As for the lens affecting the magnification, I don't see how that can be possible (I understand how the diopter affects it). The magnification is in reference to the image projected on the sensor and has nothing to do with the magnification of the lens.

On second thought, you're right about the image size at the sensor being the same regardless of lens used. I certainly don't mind you helping clarify about the viewfinders

I guess it's another instance where practice makes the theory complicated: If you didn't use the same lens at the same settings each time you measured a viewfinder, it'd complicate doing that comparison across systems. I see what you're saying pushes back against even that claim somewhat, though, since you can always compare the print with the viewfinder image. The prints would naturally be good across different systems.

I guess they have some sort of rig set up that uses calibration charts and cameras hooked up to the viewfinder and that's why they use the same lens (50mm in Canon's case).

Yeah, they'd need to have that locked down, like a test bench for shooting a rifle if you will, just for the sake of stability. The 50mm lens just helps them keep distortions low and is a stable enough reference point for purposes that may go beyond the viewfinder. But I think you're right that just having the taken photo is probably more than enough to get a very close number for the viewfinder coverage (after all, it would be rather useless if the viewfinder showed more than 100% of what the photo showed).

Edwin Herdman wrote in post #13104807
(after all, it would be rather useless if the viewfinder showed more than 100% of what the photo showed).

It is sometimes nice to see the space around the image to decide on composition but it's a PITA if you are in a hurry or if the scene is too dark for the guidelines to be visible.

Aye, that's one of the reasons people love rangefinders so much for street shooting.

peter10 wrote in post #13080974
thanks for doing this, but...

tbh to my eyes the 7D is better, but I guess we wouldn't really compare the two like this, we should compare the 5d with a focal length lens that gives the same equivelent on the crop. I suspect the 5D would then be better.

Not in the OP's context. It is subject-oriented (with limited focal length and fixed distance), not image-oriented. The 5D2 gives the better images in many cases, except that the 5D2 has a lot more horizontal banding at ISOs of 12800 and above, and more noise in the shadows of low ISOs.

Compared to what?

I've had a wide range of APS-C Canons, and the 7D is by far the best at high ISOs. Sure, the K5 and D7000 are better, but nothing Canon is.

I dunno. I'm honestly not super biased towards Canon, but Nikon's 'painting/smudging' version of D7000 high ISO is not appealing to me, I'll take gritty, honest noise every time.

JonK wrote in post #13080350
Just to see if my hunch/experience were correct (without relying on other peoples experiments) I wanted to see if there was this elusive magnification awesomeness to my 7D + 400mm combo that my 5DII + 400mm combo lacked.

It appears there was bias, demonstrated by this quote. 1st a "hunch/experience" and 2nd the use of "elusive magnification awesomeness".

JonK wrote in post #13080350
Its true, a 1.6X crop makes everything seem closer.

I guess, if one wishes to view "things looking bigger" as "things looking closer".

JonK wrote in post #13080350
Its mostly due to the field of view making the user feel as though the lens is somehow optically reaching longer than it would on a 35mm frame.

The word "mostly" intrigues me, and makes me wonder what else gives the user that "feel".
I see 4 ways in which the user gets this "feel":
1. By looking through the viewfinder.
2. By looking at the picture taken on the LCD.
3. By looking at the picture taken on a computer.
4. By looking at the printed picture.

#1:
The sensor size doesn't affect what's seen through the viewfinder, but the magnification of the viewfinder does. So what the user sees is dependant on the lens and the viewfinder. So any extra magnification/reach/cl​oseness is independent of the sensor size.
The viewfinder design will determine what the user sees. For different cameras it may be:
a. Varying lengths of dark tunnel with a picture at the end.
b. The same size picture.
As has been shown in this thread, the user sees very much the same size viewable display area on the 5d mk II and 7D, so there is little varying tunnel effect affecting perception. But the 7D viewfinder magnifies more and displays less of the optical pathway coming out of the lens, thus showing the user what the sensor will record.
So with the 7D crop camera there is more magnification for the viewfinder path than on the 5D mk II.

#2:
When viewing the picture taken on the LCD the 7D will fill the frame with less of the subject than the 5d mk II will.
The crop sensor will have recorded less of the optical pathway, but will have recorded it in more detail, resulting in roughly the same amount of total data.
The subject has been magnified more on the crop camera, but the detail is difficult/impossible to see on the LCD, so the quality of that magnification cannot be determined.
I use the word magnification as the ratio of LCD size/framed subject size, and as the subject is bigger on the LCD with a crop camera, it has been magnified more.

#3:
When comparing results on a computer it is important to remember the purpose of photography, to create pictures people look at.
So it makes sense to define the comparison, and then allow a good number of people to vote on the results. Given that people looking at pictures will be subjective, one would expect a variety of responses.
In this case I think there are two obvious comparisons:
a. Display the pictures the same size, filling the screen, with no cropping/resizing. In this case the crop camera will result in a picture with greater magnification (as defined above). Less of the subject will be visible, and the subject will be larger than from the full-frame camera.
Given the roughly equal number of pixels in each picture, this comparison could also be done at 100%, enabling the detail/quality to be compared.
b. Create the same subject framing. This is applicable if one wishes to create a picture of say a bird.
It seems to me a fair way to do this is to crop the full-frame picture to match the framing of the crop picture and then resize the crop picture to match the size of the FF picture. This involved shrinking one picture rather than enlarging the other.
The resultant pictures can be viewed to fill the screen or at 100% to compare detail/quality.

#4:
A similar choice to #3. Either:
a. Print each picture on the same size paper and compare.
b. Crop the FF picture, and print both to fill the same size paper.

JonK wrote in post #13080350
We have established that it doesn't.

I don't know who "we" is, but I think I've shown that the magnification of a crop camera is greater than a FF camera.
The question the test attempted to answer is whether that extra magnification results in a better/same/worse picture quality than using a FF camera for the intended outcome.

JonK wrote in post #13080350
But then the argument comes in that a camera like the 7D APS-C sensor with 18MP has higher pixel density than the 5DII 35mm 21MP frame, which is true. So, we've been told it will have a magnification because there are more pixels under the lens and that will mean more data and... so on.

Not more data, the same amount of data, recording a smaller part of the optical pathway, ie what's coming out behind the lens.

JonK wrote in post #13080350
I was bored this morning and did a quick test. I used a manfrotto tripod and ball head because I can remove the lens and put it back on the tripod without moving where it's looking. So all that was done was the bodies were swapped. I focused my cameras on a tree/bird feeder using live view so no MA issues are here.

This raises the "reality limit" question.
I have done a lot of tests in order to help me understand what creates the best pictures, and what affects that ultimate picture quality.
One conclusion I reached is that using a tripod, any tripod, will result in a sub-optimal result. To create the very limit in quality, it is necessary to use say 25kg of dry compact sand under the camera, and the same on top, forming a very solid high mass stable platform.
Which is great if you can do that, and your subject is equally still, etc etc.
Which is why I say this raises the "reality limit" question. When hand-holding a camera taking pictures of moving subjects, is it possible to get close to what is technically possible? I would say currently yes, close, but only close.
Lots of light, high shutter speed/flash, best aperture for the lens, good hold, perfect focus on the right spot, closest framing etc etc

As pixel density increases, that "reality limit" will become more important, to the point where very few pictures taken will come close to what is technically possible.

JonK wrote in post #13080350
The results speak for themselves:

Having read all 16 pages, it appears that the majority think the 7D results are slightly better with the same framing.
For me they are so similar, and with unknown factors affecting the test, that I'd say they were effectively the same.
The ability to focus and see the subject when taking the picture would have been better on the 7D because of the extra viewfinder magnification.

I use a 60D with 70-200mm. I do mainly portraiture.
I'm used to using the 200mm end when taking head and shoulder pictures, as it gives a narrow POV, and the noses/ears don't stick out I'm used to being at least about 5m (15ft) away.
I recently put my lens on a friends 5D mk II, and was very surprised how much closer I had to get to the model. To me uncomfortably close.

I say this as someone hoping to be convinced to buy a FF camera as a 2nd camera. Maybe a FF would be best with my wide-angle lens.

The other issue not mentioned on this thread is that of equivalence. It's a while since I've read this, but for anyone with a day to spare LOL, it's a good read:
http://www.josephjames​photography.com/equiva​lence/

liquidstone wrote in post #13083385
I suggest using an Av of f/6.3 or f/7.1 rather than f/8 on the 400 5.6L, so diffraction effects at the pixel level won't mask the detail gathering ability of each sensor.

Doesn't matter - the details on the beak captured by the 7D is smeared together on the 5Dmk2 shot. Other lens settings or a sharper lens can't do anything about that.

JonK wrote in post #13101084
When I get a free minute I will do another test. A few co-workers I've showed this to have agreed that the 7D does not look more detailed but just larger in some shots.

You started this thread with the claim that you were not biased, having both cameras.

The last half of the thread (or more than that since the tread is growing), you are busy defending your view.

The scientific approach is to look at all the evidence. We have already seen evidence where details captured by the 7D just is not available with the 5Dmk2 (which is the only one I have, and should then be prejudiced to prefer).

You decide to ignore that, claiming that you should go out take new shots to "tell the world".

Take the new shots to convince yourself. But you can not ignore the money shot. Both shots was in focus. And no movement fuzz. And one very clear winner. What we do not know, is if the 7D was limited by the lens, or if it could have won even greater with a sharper lens. But with or without rescaling - some of the information just was not caught by the 5Dmk2 sensor.

Do not defend your original view just because you happened to have started this thread. Defend it if you can show proof. And be ready to draw back your assertion when people posts clear proof to the contrary of your original view.

On the captured note, some of the lines on the beak did not have any space between the lines on the 5Dmk2 shot. The 7D did clearly capture line, space, line, space, line, space. I.e. the higher sensor resolution did manage to resolve features outside the limits for the 5Dmk2 sensor.

This thread is growing because there are lots of really stupid comments where people are trying to defend very subjective views or trying to attack others objective views. This is just moving the debate into a "random noise" direction instead of bringning more facts on the table.

Interesting facts would be: What lenses or lens+TC combinations (and aperture values) will let the 7D outresolve the 5Dmk2? People having a cheap lens can't assume that the 7D will give more "reach". But we do have proof that people with good enough lens will get more pixels on the target and that these extra pixels will capture extra details.

Next thing debated (but slightly glossed over) is that a 100% crop from a sensor with smaller pixels will show more noise. But if you have 10 million pixels on a bird or 4 million pixels on a bird, but print at similar size, each noise point on the print will be smaller. So more noise at 100% crop is only relevant when each sensor pixel is presented at the same size on the output medium. I.e. the 7D will have more noise when the extra sensor resolution is used for extra magnification.

In a situation where the center 22mm of the 5Dmk2 sensor is used for a full print, and the full sensor of the 7D is used for a same-size print, then the two photos will look quite similar in noise. The 5D print have larger pixels with less noise variance. The 7D pixels have more noise variance but each pixel is closer together. Running an averaging filter on the 7D capture to drop the excess resolution of the 7D, the pixel variance for that filtered print will drop. After all, a large sensor pixel will just capture the averaged noise of multiple pixels from a higher-density sensor. The actual noise level of the sensor depends on technology, and not pixel size.

JonK wrote in post #13102761
So then all 1.6x crop have better reach? Grab a crop body because it has better reach. I should break out my old 300D then.

Silly arguments won't win an argument. They just show that the scientific way to prove something has been thrown away to make room for personal opinions.

ktownhero wrote in post #13103119
Exactly. Which is why the 1/focal length rule for shutter speed doesn't change based on sensor size.

But the 1/focal length rule does change based on pixel size on the sensor, and all other operations that affects the total magnification from subject to final print - as defined by the fov the final print will have when viewed.

pwm2 wrote in post #13106570
blah blah blah

One thing I've come to realize on this forum is that everyone is a going to peep pixels, which they reserve the right to and in this specific case that is appropriate. Another thing I have come to realize is that there is an abundance of people who "see no difference" who are looking at images on low end TN panel 6-bit displays. Having shown some people here at work the same tests, 3 out of 4 of them said the 5DII image looks more detailed and the 7D image "looks bigger". I have no problem done non-biased tests, but on my display(s) (8 bit and 10 bit displays) the results seemed a little biased in favor of the 5DII. I'll test again, maybe I can convince myself otherwise.

JonK wrote in post #13106624
One thing I've come to realize on this forum is that everyone is a going to peep pixels, which they reserve the right to and in this specific case that is appropriate. Another thing I have come to realize is that there is an abundance of people who "see no difference" who are looking at images on low end TN panel 6-bit displays. Having shown some people here at work the same tests, 3 out of 4 of them said the 5DII image looks more detailed and the 7D image "looks bigger". I have no problem done non-biased tests, but on my display(s) (8 bit and 10 bit displays) the results seemed a little biased in favor of the 5DII. I'll test again, maybe I can convince myself otherwise.

1) You sure that the scientific process of considering receievd facts/views should summarize them as "blah blah blah"? It does seem like a slightly defensive stance.

2) Your discussion about 6-bit TN panels (I'm not using a 6-bit TN panel) shows how you mix color resolution and spatial resolution. The photo of the bill don't care if you have a 6-bit TN panel when it comes to presenting a clear difference in amount of features resolved.

3) You have shown some people here "the same tests" make me wonder - did you spend the same time letting them compare the bills as you let them compare the leaf shots? The scientific, unbiased, way would require that you did. And the scientific, unbiased, way would have you show them the two bill photos without telling which was which.

I agree that when I did look at your leafs, I thought the 5Dmk2 came out quite well, but felt it was hard to compare them.

With looking at the photos of the bills, it was not hard to compare. The 7D shots did win significantly. And I can not see any indication that the 5Dmk2 shot was at a loss because of the image being badly focused. For most of the features, it looked that both cameras captured the same amount of details. But for some parts of the photo, the 5Dmk2 was clearly (yes not slightly, but clearly) not being able to keep up.

All that reminds then, is to see if it can be scientifically shown that the 5Dmk2 photo was taken in a non-optimal way - since I have to assume that it wasn't post-processed with an intention of killing details.