The “IQ advantage” of full frame

I used on a recent trip both the 50D (with my sharpest crop lens, the Tamron 17-50/2.8) and the 5D2 (with small primes) for landscape photography. When I compared similar images from the two systems, I was struck by how much better the FF results always were (more than expected from the resolution advantage alone). This has been discussed in a recent thread:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=33144156

At least some people seem to be convinced that such a “FF advantage” is a myth and I decided to finally do a controlled test exclusively with sharp prime lenses.

All shots from tripod from the same position with LV focusing and mirror lockup. All 1/500s at f/5.6 and ISO100. All processed from the raw files in DPP with standard picture style settings, daylight WB and without any further corrections or sharpening.

The top row is the 5D2 with the 85/1.8 left and the Tamron 90/2.8 macro right.

The bottom row is the 50D with the 50/1.8 left and the 60/2.8 macro right.

Here the resized full frames:



The 5D2+85 has a very minor FOV advantage over the 50D+50.

In contrast, the 50D+60 has a significant FOV advantage compared to the 5D2+90. Apparently, the 60/2.8 has a slightly longer effective FL, at least at infinity focus.

Now 100% pixel crops from the image center:



Obviously the 5D2+85 is much better than the 50D+50. I’m not saying that the 50D is “bad” but this is simply no contest and it completely confirms my field results using the Tamron 17-50 on the 50D. BTW, my 50/1.8 is extremely sharp, actually my only prime that beats the Tamron 17-50 at f/5.6 in the center.

Is this only because of the higher pixel count of the 5D2? The answer is no: The 50D+60 puts the same number of pixels on the subject as the 5D2+85 but the FF result is still better. Fine, low contrast textures have a “three-dimensional look” that the crop system can’t match. To my eyes the difference is still striking and immediately apparent.

I don’t know what causes this “FF advantage”. As I said in the previous thread, differences in raw conversion and the larger FF pixels being easier on the lens may play a role.

if opinion you're expecting, the difference is trivial

The Medium Format would be significantly different

cheers
Have fun

Both in terms of accutance and microcontrast. Wonder how much of this is due to lens and anti alias filter.

Thank you for providing this test.

--
Jon Angelo Gjetting
Website: http://www.gjetting.com/
twitter: http://twitter.com/Gjetting
blog: http://gjetting.blogspot.com/
facebook: http://profile.to/jonangelogjetting

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Robbie
--
Canon Can...Can you??

nothsafoto said:

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Click to expand...

I don't think that's a "fatal flaw". I think this test was done correctly because it reflects exactly how a photographer would have worked in real life. In real life, to get the same equivalent framing, I would have used two different focal lengths to achieve the same "equivalent" focal length on these two formats. In this case, it was done with two different lenses, but you could have done the same thing with a single zoom lens. Doing the "focal length equivalence" thing is simply a real-world necessity when trying to get the same framed shot with these two formats, and therefore I think it's fair that tests reflect this. Besides, shooting with two different sharp prime lenses stopped down to optimum apertures, then only looking at the central sweet spot of these lenses, shouldn't kill the validity of the test. In the dpreview tests, the reviewers regularly perform tests with sharp prime lenses stopped down to optimum apertures with the different respective lenses for the different respective camera brands.

nothsafoto said:

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Click to expand...

You are joking, right?

Did anybody dispute that the camera with higher pixel density will resolve more detail with the same lens?

mattr said:

nothsafoto said:

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Click to expand...

You are joking, right?

Did anybody dispute that the camera with higher pixel density will resolve more detail with the same lens?

Click to expand...

don't use a 17 -50 against primes/ Also, the 7d is looking great

mattr said:

nothsafoto said:

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Click to expand...

You are joking, right?

Did anybody dispute that the camera with higher pixel density will resolve more detail with the same lens?

Click to expand...

he is not joking, he is just adamant. Nothing could change his mind, at least not logic and proof.

You have to go great lengths to show something that is mere a common sense. You won't be able to convience him, there will be always something wrong.

--
::> I make spelling mistakes. May Dog forgive me for this.

zxaar said:

he is not joking, he is just adamant. Nothing could change his mind, at least not logic and proof.

Click to expand...

I think you are right. I got this impression already in the last thread.

But the response here still came a bit as a shock to me... ;-)

I get the reasoning behind this post, but if the lenses are different in the comparison, your results may partially be due to the difference in the lenses as well as the sensors.

I do agree however that the 5D Mark II has lower noise levels and it does appear to be sharper than the 50D at the pixel level even though the 50D has the 1.6x crop advantage.

(I have both of these cameras and the 5D Mark II is my primary camera for birds/wildlife. Usually the 5D Mark II is on my 600mm f/4 and the 50D is on the 400mm f/5.6 for BIF.)

Alan

mattr said:

zxaar said:

he is not joking, he is just adamant. Nothing could change his mind, at least not logic and proof.

Click to expand...

I think you are right. I got this impression already in the last thread.

But the response here still came a bit as a shock to me... ;-)

Click to expand...

he says he see no advantage to full frame. I asked him to name a single apc cam that outperforms D3 on high iso (one of the full frame advantage). He just ignored it. (he has no intention of logic and proof, just winding people up).

There are many advantages to full frame, people in world are not fools to pay so much for large sensors.
anyway thanks for the tests.

--
::> I make spelling mistakes. May Dog forgive me for this.

MAC said:

don't use a 17 -50 against primes/ Also, the 7d is looking great

Click to expand...

The OP stated it was his "sharpest crop lens".....get some sharper glass. =)

--
http://www.pbase.com/clknight
Colin

"I don't WANT to believe."

You'll never convince people who aren't open to convincing.

Let them be happy.

--

“There is only you and your camera. The limitations in your photography are in yourself, for what we see is what we are.” Ernst Haas

http://garyp.zenfolio.com/p518883873/

aayjay said:

... but if the lenses are different in the comparison, your results may partially be due to the difference in the lenses as well as the sensors.

Click to expand...

I think that's most likely not a significant issue. I used 2x2 different lenses. My own experience, as well as all test reports I know of, indicate that every single one of them should perform extremely well in the center at f/5.6.

There is no perfect solution to this problem. Moving closer with the same lens on full frame will change perspective and possibly other things such as atmospheric turbulence (I specifically wanted to do the test at long focus distance for landscape photography).

Otherwise what's the point of a more expensive body with less features? (Thinking 7D here.)

One problem is that most attempts to provide "proof" are flawed. It's just basic science to attempt to remove as many variables as possible and I seldom see anyone putting much effort there.

Different lenses are a variable as are the JPG parameters used or how the "same" JPG parameters mean something very different in different bodies. There's no question that your 5D samples above have more punch and contrast but that could be due to all kinds of optical or PP reasons independent from the sensor or AA filter.

I have no doubt that FF is superior to APS-H and it's superior to APS-C and it's superior to the next smaller one and so on. The problem is that it's impossible to show that advantage here on these screens. Even in those rare cases when someone uses "good science" to eliminate variables.

I believe FF has better IQ but the proof can only be seen in large prints done by a pro.

Julio said:

Otherwise what's the point of a more expensive body with less features? (Thinking 7D here.)

Click to expand...

Well, there are other advantages of course. Lens selection, viewfinder (the two main reasons I bought the 5D2), noise, shallow DoF... All debated to no end on these forums!

I'm sensing some denial, and I'm not talking about a river in Egypt.

nothsafoto said:

The fatal flaw in this test is the different lenses used. You should have mounted the 50mm on both bodies and cropped the "FF" image to match the APS-C image in post processing. Do this and post the resulting images (both full and 100%) and I will be interested to see if there is as much diffence as we see here.

Robbie
--
Canon Can...Can you??

Click to expand...

--
http://www.jeffseltzerphotography.com

I do believe that one will get more keepers with the full frame than with the crop but that is not to say you won't get great pictures from the crop but you will get them less often. I think it is because the full frame camera reads the light and detail better more often. I know with the crop sometimes I would get a flat looking picture and I'd know that I did what was right but it would bug me and i concluded the problem was a lack of sensitivity of the camera and I was right. I found that if the lighting conditions were really good then it would not much matter which camera I used but for marginal conditions the full frame has the edge which is why I would get more keepers. I have shot some pictures with the 5D that I didn't like but that occurred a lot less than with the 350D. I think really good glass is a tad more important with the crop than with the full frame and when someone with a crop gets stellar pictures on a regular basis it must be the glass!

Thanks for doing this test. As was already mentioned above, some people won't believe and will find any reason to ignore this obvious result. Cropping the FF image to compare with an APS shot with the same lens is an idiotic waste of time of course, since no one would throw away half of the image in real use, but a comparison with the same lens taken from different distances to match the FOV might be slightly more convincing.

Also wondering why you used mirror lock up with these focal lengths and shutter speeds? Does it really make a difference?

Thank you for providing these images. This clinches it for me. I was trying to decide between a 50D and one of the 5D's. I think the path is clear now I make large prints (11x14 up), and the 5D/5d2 is the only option.