brainiac wrote:
Now imagine the image size of those two originals. The 5D2 image is huge by comparison. It started with more pixels, and then got uprezzed! By misunderstanding that issue you have proved my point that this is a very misleading comparison.
Perhaps Tom could show us the full frames of each camera so that people can understand a little better what they are really looking at, namely 2.4% of the 5D2 frame against 6.2% of the 7D frame (please correct me if I calculated that wrong).
I am biassed as I am almost always focal length limited, but I do know the differences in total image size are grealty different.
thedigitalbean wrote:
It is if in the real world you are focal length limited. There is some merit to the comparison for the focal length limited shooter (i.e. one who cannot get closer to one's subject nor can afford a longer lens). This may not be everyone, but there are plenty of them, especially wildlife and bird photographers.
...but if you want to compare reach it is important to explain that you have discarded 37.5% of the 5D2's image quality before making the comparison. Tom didn't explain that and people are already drawing the wrong conclusions.
brainiac wrote:
...but if you want to compare reach it is important to explain that you have discarded 37.5% of the 5D2's image quality before making the comparison. Tom didn't explain that and people are already drawing the wrong conclusions.
Ah I see, I figured it was obvious thats what he was trying to do, perhaps it wasn't.
Thanks for that - it is indeed quite a well made comparison. I do wonder whether the 7D might fare a little better with high sharpness glass like a Zeiss ZE or something. The lenses he used are good, but who knows if they saturate the 7D with detail. Although he didn't admit it, his f2.8 sample with the 50 f1.4 on the 7D does appear to be a little sharper than the f8 shots in the main test. The 7D is hard on lenses for sure.
Even though the comparison of 7D images with different apertures shows that f5.6 is noticeably sharper than f/8, Roland has refused to use DoF and diffraction equivalent apertures in the direct comparions between 7D and 5DII.
The use of different RAW converters:
His comparison of DPP versus CaptureOne shows that C1 is softer than DPP. That does not match anyone else's experience with those two converters and 7D. CaptureOne always gives sharper results for the 7D than DPP. For the 5DII, there is no difference in sharpness between the two converters. He must have messed something up there.
brainiac wrote:
...but if you want to compare reach it is important to explain that you have discarded 37.5% of the 5D2's image quality before making the comparison. Tom didn't explain that and people are already drawing the wrong conclusions.
(italic boldface added)
Agreed that good explanations are important. I understood the intended purpose of this test (at least, what it means to me, as a focal-length limited occasional bird photographer), but it's very easy to read a post quickly, perhaps miss a detail or two, apply one's own interpretation/preconceptions/biases/etc to what is being said (in contrast to what the OP really means to say), and let one's own interpretation solidify before the OP's meaning even has a chance to gell in one's brain.
That said, I feel that brainiac's statement that I noted above needs qualification. If one is focal length limited and is interested in comparing reach, it's quite possible that the discarded "37.5% of the 5D2's image quality" (i.e., the part of full-frame outside the 1.6 crop) is irrelevant for one's photograph, and it is fair to compare the 1.6 cropped portion of the 5D2 sensor to the 7D sensor.
On another note, I will also agree that it is absolutely crucial to nail the focus if the comparison is to be meaningful. I think that many who think that 7D or 50D images are not significantly better than up-rezzed 40D images may have been lead astray by less than optimally focused comparison photos. Of course, to really take advantage of all those 18MP, one really needs to practice good technique.
Agreed that good explanations are important. I understood the intended purpose of this test (at least, what it means to me, as a focal-length limited occasional bird photographer), but it's very easy to read a post quickly, perhaps miss a detail or two, apply one's own interpretation/preconceptions/biases/etc to what is being said (in contrast to what the OP really means to say), and let one's own interpretation solidify before the OP's meaning even has a chance to gell in one's brain.
That said, I feel that brainiac's statement that I noted above needs qualification. If one is focal length limited and is interested in comparing reach, it's quite possible that the discarded "37.5% of the 5D2's image quality" (i.e., the part of full-frame outside the 1.6 crop) is irrelevant for one's photograph, and it is fair to compare the 1.6 cropped portion of the 5D2 sensor to the 7D sensor.
On another note, I will also agree that it is absolutely crucial to nail the focus if the comparison is to be meaningful. I think that many who think that 7D or 50D images are not significantly better than up-rezzed 40D images may have been lead astray by less than optimally focused comparison photos. Of course, to really take advantage of all those 18MP, one really needs to practice good technique....Show more →
Yes - all well said. In addition, it's clear that higher density sensors increasingly narrow the technique window. To see the advantage of the extra pixels you have to tick a lot of boxes:
- superb lens
- very accurately focussed
- high shutter speed
- plenty of light for detailed (lower) isos
- aperture below diffraction limit
This highlights the number 1 full frame advantage: it's much more forgiving with all of these factors and can save you a lot of money on lenses for an equivalent result UNLESS you are reach limited in your typical application.
The has some validity to me, for real life shooting, if you are trying to get a subject to fill the frame, this sort of comparison would be useful:
If there was an alligator infested moat in front of the statue, then you would have to uprez the 5D2 file thanks to the crop factor of the 7D, using the same lens.
You can't always walk forward or backwards to fill a frame in a given shot. The 'more pixels on a duck' argument.
I would like to add one more reason to have a crop camera, besides the focal-length limited situation.
It fits right in here, and you mentioned the savings in optics by using full frame.
With a set of lenses, you get another set of lenses "for free" if you have both a full frame and a crop camera.
Especially with primes, the extra set of FOV's may come in handy. Now, that point would be cancelled if you could get the same image quality just by cropping and uprezzing from the full frame body. Tests like this are suited to find out just that.
Tom, I hope you didn't take the critique too hard. Your test is not worthless, thank you for bringing it on as is. Testing is is a tough game, and if everything is not right, you get the flame. And of course, the less flaws in a test, the more useful it is.
brainiac wrote:
Yes - all well said. In addition, it's clear that higher density sensors increasingly narrow the technique window. To see the advantage of the extra pixels you have to tick a lot of boxes:
- superb lens
- very accurately focussed
- high shutter speed
- plenty of light for detailed (lower) isos
- aperture below diffraction limit
This highlights the number 1 full frame advantage: it's much more forgiving with all of these factors and can save you a lot of money on lenses for an equivalent result UNLESS you are reach limited in your typical application.
So very true Richard. The 7D for me sees primarily one use, to get more reach out of my 800 when I simply cannot get any closer to those tiny shorebirds. And in those cases yes, all those boxes do get ticked
timpdx wrote:
The has some validity to me, for real life shooting, if you are trying to get a subject to fill the frame, this sort of comparison would be useful:
If there was an alligator infested moat in front of the statue, then you would have to uprez the 5D2 file thanks to the crop factor of the 7D, using the same lens.
You can't always walk forward or backwards to fill a frame in a given shot. The 'more pixels on a duck' argument.
Thanks for posting
You can't always walk forwards or backwards, but you can always either zoom or switch lenses. If you're using the correct lens for each camera to frame the pictures identically, then the 5D2 always puts more pixels on the duck.
If you're not framing the pictures identically, then comparisons are rather moot.
Even though the comparison of 7D images with different apertures shows that f5.6 is noticeably sharper than f/8, Roland has refused to use DoF and diffraction equivalent apertures in the direct comparions between 7D and 5DII.....
I realize that the 18mp of the 7D will certainly show diffraction quite well, but f/8?
Any image that I would choose to print large (mostly landscapes) would likely be shot at least at f/8. Thus, for me, this "diffraction limited test", if you want to call it that, is still indeed very meaningful to me in the real world.
I don't mean that f/8 isn't useful on APS-C. If that is what it takes to get the DoF you need, that's fine. The DoF and diffraction you get at f/8 with APS-C is identical (as a fraction of the whole image, at equal output size) to what you get with f/13 for FF. People don't use f/13 to showcase what resolution they can get from a FF camera, do they?
Fat Dave wrote:
You can't always walk forwards or backwards, but you can always either zoom or switch lenses. If you're using the correct lens for each camera to frame the pictures identically, then the 5D2 always puts more pixels on the duck.
If you're not framing the pictures identically, then comparisons are rather moot.
So, where do I get (at the same price and weight as the 300 /4 IS) an EF 480mm f/6.3 IS lens, the "correct" lens to use on the 5DII here?
The same place you get the 53mm f/0.9, 31mm f/0.9, and 11mm f/2.5 TS-E you'll use on the 7D. etc etc etc
Obviously there are not exact equivalents moving back and forth. However, if one is trying to do a "comparison", one should frame identically, or as close to it as possible.
I have to jump in here and offer nothing useful.......
I'm an engineer too, like many of you seem to be in this thread... -- I, however, instead of trying to work my passion as if it were my profession, rather like to take refuge in protography from the daily drudge of calculus, ballistic coefficients, kinetic potentials, and other things that make one's brain hurt.
Geez people... photography is art and as such it is very subjective.... So think of the comparison this way -- Look, Ma -- no math!
I have an assignment to shoot...let's see... some Rolex watches for a Jewelry story catalog.. I will shoot this assignment with either a 7d or a 5d.....
I choose my trusty 24-105L for this.... its awesome sharpness, ability to focus to 1.5ft, and its zoom range appeal to me. Of course, IS never hurts!
I set up my lighting and "scene" or layout of products WITHOUT regard for any camera -- just like I would do it if I had already chosen the camera. I don't care at this point.
I compose the scenes in the viewfinder the same way with each camera and shoot a series of test shots..... meaning, I take the same shots with each camera -- I don't care about sensor size or pixel density. If I move up or back, its not to compensate for larger VS smaller pixels at their relative densities within their respective sensors --- I only move to make the same composition in the viewfinder. I'm just as likely to change focal length to achieve the same thing. Since this is a comparison exercise, I would take comparable test shots by both moving and changing focal length with each camera. I only care about the finished product.
I either print or view the results on screen along with some assitants and maybe the client and choose the best. Done.
If I had a 5d2, I would do this tonight and post it tomorrow for you all to choose....Unfortunately, I only shoot crop bodies......
Epologue with an old cliche: "The more things change, the more they stay the same".....
Back in the day before the Web was even a twinkle in a Swiss researcher's eye.. I remember being at a local camera club meeting and witnessing a heated conversation betwen several chemists... one of those great Kodachrome vs Velvia flamefests... They were arguing passionately about how a one film emulsion's silver halide particulate distribution interacted with the color dyes to produce much better images than the other....... and how a 21 chemical process with a .25 degree temperature tolerance was so easy to screw up by the mere mortals who were running the labs....... and oh yes, my favorite... they argued endlessly about how invalid all suggested testing methodologies were..... WHY
BECAUSE Kodachrome and Velvia are (were?) very different films!!
alundeb wrote:
So, where do I get (at the same price and weight as the 300 /4 IS) an EF 480mm f/6.3 IS lens, the "correct" lens to use on the 5DII here?
