Why full frame cameras 'expose' the flaws in 'lesser' lenses?

[Terry Rogers]

I have a question for something I've read and heard multiple times but am not sure about why.

I read that modern full frame cameras 'expose' or 'reveal' the flaws in lesser lenses more than crop cameras. To me this doesn't make sense save a couple caveats.

I would expect a 20-24mp full frame sensor to 'hide' certain lens defects due to the lower pixel density compared to 18-22mp crop sensors. I would expect greater sharpness from the same lens on the full frame sensor compared to the crop sensor because of the decreased pixel density. I don't know how CA is affected by this.

I would expect corner sharpness to be decreased on the full frame sensor compared to the crop sensor, but that could also be accomplished by shooting wide and cropping in on only the ff sensor.

Could someone explain to me why a full frame sensor would 'reveal' the deficiencies in cheaper lenses compared to the same lens on a crop sensor?

Thanks

 

[Zv]

In respects to corners and mid frame the FF sensor would show more of the image circle than on crop therefore you would see a decrease in sharpness, more vignetting and more CA as thats were these things tend to hang out. The net effect of all this is decrease in IQ.

Compare the same lens on a crop sensor and some of that just simply gets cropped away while the centre, which is sharp now extends to the mid frame.

However, I do wonder if my 7D shows more lens flaws than my 5D2 as, like you said due to increased pixel density it ought to. I like clarification on this point too.

I think since FF are more sensitive to light it may also show more flaws as a result but I don't know how.

 

[Pi]

I read that modern full frame cameras 'expose' or 'reveal' the flaws in lesser lenses more than crop cameras. To me this doesn't make sense save a couple caveats.

There is a good reason that it does not make sense to you because it is false, in the majority of the cases.

First, we are talking about EF lenses only. In the center, FF sensors enlarge the image 1.6x less. If that was the only factor, then FF would resolve 60% more in the center. But there is also pixel density/AA filter. To make it simple, assume the same pixel count. Then the FF advantage in the center, same lens, same f-stop, would be from 0% (ideal lens) to 60% (very soft one). In reality, it goes up to 40%, depending on the lens and how you measure it.

The meaningful comparison however is for equivalent f-stops and equivalent FL's (same AOV). Then FF has even greater advantage for soft lenses, like 85/2 vs 50/1.2, where FF resolves even more than 60% more. When you stop well, then diffraction is the real limit, and then asymptotically, FF and crop have the same resolution.

In the borders, FF "reveals" the flows of the lens in the corners. Well, it still enlarges them 1.6x less; and for equivalent lenses, you have to stop the FF more. Then (depending on the lens), most of the times, you get a better resolution in the borders as well. A notable exception usually cited is the 17-40 in the borders but TDP shows otherwise (the lens at 17mm is focused there too close, BTW). Still, exceptions exists, I guess.

It only makes sense to me to compare not the same lens on FF vs. crop but the lens that you may use on FF vs. the one you may use on crop. Then the 17-40 in the borders is better than the 10-22 on the wide end, according to the TDP crops (again, focused too close).

Here is a graph of the resolution (in the old units DXO used until recently, MTF-20), in the center, 50/1.4 on the 7D, vs. the 85/.8 on the 5D2, in equivalent f-stops. I was planning to draw a similar graph for the border performance but DXO decided to hide their data.

 

[Mt Spokane Photography]

I read that modern full frame cameras 'expose' or 'reveal' the flaws in lesser lenses more than crop cameras. To me this doesn't make sense save a couple caveats.

There is a good reason that it does not make sense to you because it is false, in the majority of the cases.

+1

FF sensors will always produce a sharper image on a given EF lens. I would not believe everything I hear. Its true that there is more viginetting.