Wallet full of $100 bills

[Jon Ruyle]

In that case, we'll have to agree to disagree, I suppose. Look at the photozone.de MTF numbers for these two lenses -50mm f/1.2L @ f/1.2 on5DII: 2473 LW/PHvs. the85mm f/1.8 @ f/1.8 on5DII: 3082 LW/PH (the 85 1.8 is ~25% sharper by the numbers) on the same body.

But that's comparing the lenses wide open and uncropped. The result is exactly as expected: the slower uncropped lens wins. This isn't what I meant by "switching the lenses". I meant, instead of making the 50mm faster and cropping it, make the 85mm faster and crop it.

The two systems in question have different sensors - they aren't going to take the same pictures. The only reasonable way to perform that comparison is to use the same lens with a different distance from the subject, and live with the fact that perspective and compression are different.

They wont take exactly the same pictures, of course. One will have better IQ. But you can compare two systems, one FF and one APS-C in which each system has the same framing, same working distance, same DOF, same compression and same noise. Yes, the two systems will use different lenses, and this effects the outcome. But if I consistently get far better IQ taking the same picture (modulo IQ) with cheaper lenses using one of the two systems, that's the one I want.

You cannot change both lenses and bodies and expect to compare the two systems.

This is the core of our disagreement. I say you not only can change the lens, you *must* change it. You must compare systems with the same *effective* focal length and *effective* f numbers. You seem to feel we must compare systems with the same actual focal length and same actual f number.

You need to perform two independent comparisons each with one constant and one variable.

I agree. But I think you're using the wrong variables. If we change the point (3,4) in Cartesian coordinates to the point (6,8), we've changed two variables because x and y both changed, right? Maybe. But in polar coordinates, I've kept theta the same, and only changed the variable r.


Keep everything else the same and change only the sensor. Yes, you have to change the lens, too. That's too bad. Try to choose lenses that aren't too different. (Don't compare a cheap zoom to an expensive prime, for example). One system will consistently win, and the lenses you need to perform the test for that system will consistently cost less.


If you insist that it makes sense to compare IQ in systems that take pictures with different DOF, compression, exposure time, etc, then I agree that we must agree to disagree (don't read that ten times fast [])


Anyone else out there want to weigh in? Is anyone else still there? Can you say "hijacked thread"? My apologies, John.

[Sheiky]

Anyone else out there want to weigh in? Is anyone else still there?

Sigh....haven't been active the last few days and I missed A LOT on this thread obviously [:P] Outlook almost considers this thread as SPAM []I'm not sure if I do want to catch up though [A] It's a lot of technical talk, but I might give it a try when I got more time...like two weeks time [8-|]

[neuroanatomist]

But that's comparing the lenses wide open and uncropped. The result is exactly as expected: the slower uncropped lens wins.

You're saying that the 85mm f/1.8 'wins' because it's slower, and becauseit's a longer focal length (no cropping). Yes, the 85mm f/1.8 does win, for sharpness. But not because it's slower and not because it's uncropped/longer. It 'wins' because it is a sharper lens (as I and many reviewers have said, because Canon designed the 50mm f/1.2L to favor bokeh, not sharpness).


Let's extend the comparison in the same direction - compare the 85mm f/1.8 with a lens that is just about exactly 1.6x slower and longer - the 135m f/2.8 Softfocus. In that comparison, the shorter, faster lens is a bit sharper - it still 'wins', with both lenses wide open. Why is the 135mm f/2.8 softer? If you tell me that the slower, longer lens is softer because it's just not as well-designed a lens in terms of sharpness, well...you're right. Of course, if that's your reason...you'll also have proved my point about the 50 1.2L vs. 85 1.8 comparison. If you have another reason, well, I'd be interested to know what that is!


Now, the 135mm f/2L is slower and longer than the 85mm f/1.8, and it's also sharper wide open than the 85mm f/1.8 - but again, because it's simply a sharper lens, not because of the aperture or focal length. The 135mm f/2L is sharper than the longer, slower EF 200mm f/2.8L II for the same reason, it's just a damn sharp lens!

But if I consistently get far better IQ taking the same picture (modulo IQ) with cheaper lenses using one of the two systems, that's the one I want.

So, a slower, longer, cheaper lens will give better IQ on a FF than a faster, shorter, more expensive lens will give on a crop body. In that case, someone with a crop body who needs the 'normal' FF-equivalent angle of view shouldn't even consider the 35mm f/1.4L that Keith just showed us is very sharp on a crop body. Instead, they should go FF and just get theslower, longer, cheaper EF 50mm f/1.8 II - that lens will have better IQ on a 5DII than the35mm f/1.4Lon a 7D. At least, that's what your statements indicate. If you can prove to me that the nifty-fifty has better IQ on a 5DII than the35mm f/1.4Lon a 7D, I'll buy a 5DII - and I'll give you my 7D!


The other factor is that a system has many parts and features. The faster frame rate and better AF tracking are significant differences between the 5DII and the 7D. If the 5DII with it's better IQ misses a shot because the subject was running toward you too fast for the AF to track it, or misses the perfectly-timed action shot of the burst because it's frame rate is half as fast as the 7D, better IQ doesn't help.





To frame this discussion a little differently, I think it may help to differentiate between imagequalityand image<span>characteristics. 'Quality' comprises factors like sharpness, noise, color, and contrast. 'Characteristics' comprise factors like depth of field, perspective, and angle of view.
<div>


The 50mm f/1.2L on the 7D will have similar image 'characteristics' to the 85mm f/1.8 on the 5DII - but, the image quality will be different, because of the inherent differences in the lenses.


I think it's possible to 'simulate' differences in image characteristics by using different lenses on different bodies - butnotpossible to compare image quality by using different lenses on different bodies. My contention is that when comparing the 7D and the 5DII, themaindifferences (with the notable exception of noise) will be in the image<span>characteristics.


IMO, if you take an excellent prime like the 85mm f/1.2L II and put it on the 7D vs. the 5DII, the main differences (again) will be in the image characteristics, not the image quality (as I've defined them).


From that perspective, you're absolutely right aboutchoosing a system where thelenses you need to perform 'better' will consistently cost less. But better is relative. If 'better' means shallowerdepth of field, ok - the cheap, slow lens on FF wins. If 'better' means narrower angle of view to avoid the need to crop away resolution, the crop body wins, especially from a cost perspective at the supertelephoto end. In terms of image quality, the FF will win for noise, no question. Sharpness, color and contrast are all critically dependent on the lens, and much less dependent on the sensor.


Not sure if we can agree on that (I suspect the answer there is 'no'). But, we can agree to disagree, 'eh?
<div></div>
<div></div>
</div>

[Jon Ruyle]

I wont have time to reply fully until tomorrow (I've got to put the kids to sleep, then early to bed to get up at 4:30am ).


But one statement you made *really* caught my eye.

If you can prove to me that the nifty-fifty has better IQ on a 5DII than the35mm f/1.4Lon a 7D, I'll buy a 5DII - and I'll give you my 7D!

According to photozone, the 35 f/1.4 @ f/1.4 on an APS-C has the following MTF 50:


Center: 1873.5


Border: 1574


Nifty fifty @ f/1.8 on 21mp full frame:


Center: 2821


Border: 2460


Extreme border: 2359


Don't worry. I'm not ready to take your 7D yet []


We're comparing an 8mp APS-C to a 21mp full frame, which is unfair, I admit. If *only* they had done the test with a 15mp APS-C I'd have my proof!

[Fast Glass]

Ok, I have backtrack this thread and have come to a conclusion. FF is better than 1.6 so I will get a FF. Simple.[]\


Although using a longer slower lens is in a sense unfair but is inherente in the two optical systems. But is't that great? Same DOF for a slower cheaplens, andsharper for the same apparent DOF.


John.

[Jon Ruyle]

On 8mp APS-C, the 85 f/1.8 has similar results wide open to the 35 f/1.4. Photozone tests the 85 on both 8 mp and 15 mp APS-C.


On 15mp APS-C, the results are:


Center: 2267


Border: 2183


Assuming the 35 f 1.4 which has similar scores at 8mp also has similar scores on the 15mp APS-C, it would still fall far short. I admit that isn't proof. But, well, I pretty confident that the nifty fifty *would* win this. (But then, I was pretty confident of that before looking at photozone [])


Hopefully I didn't mess up any of this data... I'm in more of a hurry than usual.


I've gotta put the kids to sleep

[neuroanatomist]

We're comparing an 8mp APS-C to a 21mp full frame, which is unfair, I admit.

Yeah - PZ's wide open center MTF for the 85mm f/1.8 is 1813 with the 8 MP 350D, and 2267 on the 15 MP 50D.





If Keith's test with the 35L didn't convince you that the 7D can deliver images with (approximately) the same quality as the 5DII, I guess we have no choice but to agree to disagree...

[btaylor]

I haven't read too much of this thread (mostly because it's ENORMOUS!!!) but I will say one thing - I a little jealous that Keith gets to play with both the 5D MkII and the 7D.


[]

[Fast Glass]

If Keith's test with the 35L didn't convince you that the 7D can deliver images with (approximately) the same quality as the 5DII, I guess we have no choice but to agree to disagree...

Sensor only the 7D and 5D II are not even close in terms of sharpness WITH the same lens. I downloaded Bryansnoise cropsand I needed a lot moresharpening tostillget inferior results to the 5D II,downrezed to the same resolution and at native resolution. I used 85 units for the 5D II and I needed 110 units for the 7D.


John.

[Keith B]

I couldn't tell anything really from my test. The difference in magnification made it impossible to tell and therefore not a good way to compare them at all IMO. At first it seemed right but I'm kind of a dumb guy.


I'm definitely not a pixel peeper but I don't mind helping out if anyone has a different test they'd like to see. I saw the zoom theory mentioned. I don't have the mention 70-200 II but I have the mk I, 16-35II and 24-70 if anyone wants to see the results from any of those lenses. I think the 24-70 is pretty consistent through out the range.