Anti-Aliasing filter, how big is the impact on the field?
Saturday morning, early morning in the gym and some spare time available to test something I always thought about the effects of the anti-aliasing filter (AA filter from now) inside the cameras on final image. Why Leica cameras are so crisp and sharp compared to other brands with same resolution and same lenses? Yes, the sensor stack could be one of the elements, but for sure the main difference is the absence of the AA filter on the various M bodies and always inside the peers with same resolution.
So I took my old Ricoh GR with its collapsible 19mm lens and I tested the difference with the Sony a6400 with Batis 18mm mounted on.
Ricoh Gr (200 £ used) is the Ninja camera for street, it has a 16mpx APS-c Sensor without AA filter, on the other side the A6400 has a 24mpx APS-c sensor with AA filter; later I added my Sony A7r3 with his Full Frame 42mpx sensor without AA filter and my beloved Canon 24 70 2.8ii mounted on.
I print so much and I know for normal size prints, a resolution higher than 10mpx is useless (if you don’t need crops) and this was the final resolution of the test for the bunch of contenders: 4000px long size (almost 10.5 mpx on 3/2 aspct ratio). The correct way to test different magnifications of different sensors is by normalizing the resolution. The settings are 19mm 5.6 for Ricoh, 18mm 5.6 for A6400, 28mm f8 for A7r3 with focus (manual) on the right hill to avoid any field curvature.
There are 100% crops :
GR vs A7r3
The Ricoh GR is monster camera, almost 8 years in the market, followed by the GR2 (just wifi more) and GR3 with newer sensor and IS is still capable to wape out any other APS competitor with AA filter, also thanks to his supersharp lens. The main purpose is street photography, were is an undisputed champion but, we can see its big voice on landscape photography.
Yes, the Sony A6400 has a blazing fast autofocus and a great sensor with so much DR but is the third of the bunch about IQ.
The GR vs Sony A7r3 show us another point of view: after this test we have another demonstration about the marketing plans of Sony,Canon,Nikon, etc. to expand, adding the AA filter to cheapest products, the difference on IQ between medium/top products (a7r3) and prosumer products (a6400); I’m sure if tomorrow Sony make the a6400r without AA filter, the differences on the filed would be almost not relevant. I am also sure , if a new Sony a74 of 24mpx would come on this year withour AA filter, it would be a market revolution.
And, this is another BIG jump against Canon marketing ; they stopped for now to work on the new evolutions of cameras without AA filters and if you are a Canon user, 5dsr is the latest camera with almost (added and removed) AA filter.
I had a chance to test another point of view about this topic: how about the A6400 compared to the A7r3 in crop mode ?
That would be to test 24mpx APS-c with AA filter vs a 18mpx crop of the Full Frame sensor without AA filter. Curious about that? I did the test with the same process, resizing to 4000 long side and….
The result is the same of above, the AA filter kills fine details when compared to something captured without it.
And yes, crop mode of the big sister (18mpx for the A7r3, 26 for A7r4) is always better than the original APSC picture.