dpreviewTV: EF-M 22mm F2 not "ideal" for the M6 Mk II?

nnowak wrote:

RLight wrote:

nnowak wrote:

RLight wrote:

The EF-M 22 outresolves the 20mm Sony, even at a MP deficiency on DXO Mark where the original EOS M (18MP) vs the 24MP A6000 per DXOMark...

https://www.dxomark.com/Lenses/Compare/Side-by-side/Canon-EF-M-22mm-F2-STM-on-Canon-EOS-M-versus-Sony-E-20mm-F28-on-Sony-A6000__989_819_1093_942

.

This whole lens not sharp enough for sensor business applies to everyone. But arguably least to Canon.

For a example, the EF-M 11-22 outrsolves the Sony 10-18 by 20%, once again being tested on an inferior 18MP body.

https://www.dxomark.com/Lenses/Compare/Side-by-side/Canon-EF-M-11-22mm-F4-56-IS-STM-on-Canon-EOS-M-versus-Sony-E-10-18mm-F4-on-Sony-A6000__1170_819_1084_942

And lastly, the Sony 16-50, again, is outresolved with the MP advantage...

https://www.dxomark.com/Lenses/Compare/Side-by-side/Canon-EF-M-18-55mm-F35-56-IS-STM-on-Canon-EOS-M-versus-Sony-E16-50mm-F35-56-on-Sony-A6000__990_819_1082_942

First, what do Sony lenses have to do with the ability of EF-M lenses to resolve 32mp?

Simple. If you're going to grade a compact lens on performance, grade it against others in class.

DPReviews comments had nothing to do with Sony or any other mirrorless system. Their comments on lens resolution were in regards to the difference between an M6 and an M6 II.

True. But, this is another what’s good for the goose is good for the gander. DPR and DPR TV have a way of being inconsistent with reviews and this although a logical segue they’re doing on questioning the 22 resolving power in the wake of a super resolution sensor, should’ve been asked of those less powerful Sony lenses against the less resolution 24MP sensors. It’s not. Because the Canon is fodder here in getting the treatment first it’s pertinent to give context of other options that you have (the OP) as that context is lacking in the discussion. Other non-Canon options are more lens-limited. I’m going to go out on and limb and say Nikon, Sony and Fuji, the other APS-C players will not have the same question asked. And, that’s dirty brand politics being played by DPR. It’s a peeve of mine when that sort of thing happens, as it does around here.

Otherwise, what you're really saying is you need bigger glass, at which point I vote buy a FF, but that defeats the point of an M. The point I'm drawing is to draw the readers attention to EF-M glass quality, for it's size/weight is quite high which the implication by DPR TV is that it's not good enough.

Not good enough for 32mp. Nothing in the comment had anything to do with "not good enough compared to Sony".

Now to the point of being able to resolve 32MP, as others have stated, any of these lenses can resolve it.

Plenty have stated it, but it is incorrect. Every lens has a physical limit to the finest detail that can be sharply projected onto the image sensor.

The real question becomes one of how much gain can one expect?

This is the real point. A lens that is soft will just have that softness better recorded with a higher resolution sensor. A final print will display no more detail than the same image with a lower resolution sensor.

Someone would have to do some benchmarks as EF-M glass doesn't get much love in that regard with metrics. DXO has some, but, they stopped testing after the 11-22, and never tested on the M3 or newer either so it's a stretch here of well this guy is better than that guy, which I just did, but doesn't directly answer the question to your point. It just gives you an idea what to expect if you're considering it vs another system, which the OP is coming from Sony I might add. If, you're gaining more lens resolution, and more sensor resolution, you can expect more resolution. Now can you expect FF resolution? As I said, EF-M glass needs more love in benchmarks

You are arguing something completely different from DPReview. Your suggestion is that Canon is better than Sony because the EF-M 22mm f/2.0 is a better lens than the Sony 20mm f/2.8. This is a completely different discussion from DPReview's contention that the 22mm can not properly resolve 32mp.

.

Optical limits has a couple more lenses tested, and with the M3, but, they lack the Sony 20mm. This is a hard question to answer with out some assumptions or fill in the blanks for the 22 in particular.

Second, a higher resolution sensor makes a lens look worse, not better. Switch your comparison to the 20mp a3000 and the Sony lenses get much closer to the EF-M lenses.

Sounds like DXO has anomalies (what a surprise) as 20MP should resolve less detail and score more poorly in sharpness, yet to your point when I switch it, ditto. That should be flipped. Good catch. The EF-M is still sharper, btw.

No, it should not be flipped. You are misunderstanding the impacts of sensor resolution on lens testing. The simplest example is CA testing which is reported in pixel widths. On a low resolution sensor, a lens may only score 1 pixel width of CA. Put that same lens on a high resolution sensor and it scores 3 pixel widths of CA. The lens looks worse on the high resolution sensor because the high resolution sensor is better able to record the defect.

Here is another way to look at the issue.... If I measure myself in one meter increments, I am 2 meters tall. If I measure myself in one centimeter increments, I am 170 centimeters tall. I did not get shorter between measurements, but the higher resolution scale more accurately captured my height.

Third, you absolutely positively can not compare lens tests that were shot with different cameras. Different cameras means different image sensors, different micro lenses, different resolutions, different AA filters, different RAW conversions, etc. You can only compare lenses shot with the same camera. Here is the same lens tested on two different 20mp cameras with different results

https://www.dxomark.com/Lenses/Compare/Side-by-side/Tokina-AT-X-Pro-SD-11-16-F28-IF-DX-II-Nikon-on-Nikon-D500-versus-Tokina-AT-X-Pro-SD-11-16-F28-IF-DX-II-Canon-on-Canon-EOS-7D-Mark-II__872_1061_871_977

Fourth, DXO scores are utter nonsense and virtually meaningless.

Lots of people say that, yet, I need a common benchmark to prove the point with numbers. DXO gives numbers and has the 20mm to compare it against.

Every system, including the M system, has great lenses and bad lenses. No system is comprised entirely of great lenses or entirely of bad lenses. The optical tests on various websites help you determine which lenses are the good ones and which are the bad within a specific system. The tests absolutely positively are not meant for comparing different systems and each website usually has a disclaimer specifically stating this.

Too bad they don't have Fuji... We've had this discussion before.

It is because DXO Photolab does not support Fuji X-Trans files. All of the data on DXOmark is just a repackaging of the testing that DXO does when profiling its RAW conversion software.

I feel optical benchmarks are lacking out there.

Yes, they are lacking, but I am not sure what use it would provide beyond silly brand bashing arguments. The real questions are "what focal lengths do you need?" and "which system covers those focal length with suitable quality?" Whether the Sony 16-50mm kit lens is better or worse than the Canon 15-45mm kit lens is irrelevant. If your primary need is a really, really good standard zoom, it should be obvious that neither of these options will meet your needs.

It's quite frustrating when trying to gather what one can expect. A best answer, is the best I can give. It's not "the" answer though to your point by any means. Far from it.