First and foremost, let's be very clear: I am not a rangefinder shooter and certainly not a rangefinder reviewer. But I'm more excited than most people about the new Leica M (Typ 240) camera for one simple reason. It has live view and focus peaking so at long last I can, if I want, actually focus a Leica camera. (I have a vision problem that prevents me from focusing a rangefinder accurately.)
But like a lot of people I viewed Leica's move to a CMOS sensor, rather than the CCD used in the Leica M9 and Leica M-E cameras, with a bit of trepidation. That 18-megapixel CCD had more resolution than one would expect from an 18-megapixel camera. Despite the sensors many limitations, I was concerned that a 'modern' 24-megapixel CMOS sensor might actually be a step backward on the resolution front.
So, while I'm not a reviewer, I am a tester and have access to a nice Imatest lab. It seemed a good idea to compare the MTF50 of the new M Type 240 against the older M9 with the same lenses.
I'm pretty sure I'll be the only person to do this for one simple reason. The new base on the Leica M makes setting the camera up for Imatest incredibly difficult. It wasn't easy on the previous Leicas, but setting up the M-240 took around 2 hours at just one focal length.
So what you have here is a comparison of MTF50 using the Leica 50mm f/1.4 Summilux ASPH lens. Since the same lens was used for all three tests and for each, multiple focus bracketed images checked and the sharpest kept, it should give us a nice comparison of differences in the sensors.
For those of you wanting numbers on other lenses, I hope someone checks for you. It won't be me -- the M type 240 is just too difficult to set up. But this should, at least, give us a comparison of system resolution with the same lens between the new M-240 and M9.
I'll just give the results as tables with MTF (measured in line pairs / image height) in the center, averaged over 13 points, and averaged in the 4 corners. (Both horizontal and vertical resolution are measured at each point.) We'll measure at f/1.4, f/2.8 and f/5.6 for each camera.
|Leica M (Type 240)||f/1.4||740||630||590|
|Leica M (Type 240)||f/2.8||1070||860||770|
|Leica M (Type 240)||f/5.6||1140||990||860|
Well, obviously my initial concerns were incorrect. The new M-240 resolves at least as well (measured by MTF 50) as the M9 did. It probably is just a bit better. I don't want to split hairs - the differences in the center and overall are pretty small and probably of no, or very little, significance.
The difference in the corners, though, does appear to be approaching significance. (One thing to note, in determining the overall weighted average, the corners count only 25% as much as the center, and half as much as the mid points, so the corner difference gets masked a bit in the 'average' number.)
Why would there be a corner improvement larger than the improvement in the center or midpoints? My first guess would be that Leica, those masters of on-sensor microlenses, have improved the microlenses on the new sensor. But that could be entirely wrong.
It might also be that the corners are better with a 50mm lens and won't be as different at other focal lengths. Or perhaps this camera was just perfectly in tune with this lens. (We only have one right now, so I can't do a comparison.) Hopefully, someone else will decide that's worth further investigation.
But for any of you who were, like me, a bit hesitant about the new sensor, it's really good. Judging by what real reviewers like Steve Huff, Ming Thein, and Sean Reid are saying and showing, this is rather redundant anyway. The images are awesome.