Nadiah. Leica M-Monochrom, 50/2 APO
The final part in this triptych aims at examining the differences between Bayer and non-Bayer sensors. Part one was the review of the M-Monochrom; part two, the APO-Summicron 50/2 ASPH.
Note: I’ve been informed by Leica that both camera and lens are prototypes, and there may be changes between now and the final release product.
Operational and system differences aside, the aim of this portion is solely to look at how the sensors render images in black and white, and to examine pixel-level files and resolution advantages of the non-Bayer sensor. It’s frequently claimed that the Bayer matrix removes between 30-50% of the real resolving power of the sensor – i.e. a non-Bayer sensor of the same pixel count will have somewhere between 1.5x and 2x the resolving power. Since the M-Monochrom’s sensor is 18 MP and full frame, what better to compare it against the 36MP (2x) also full frame Nikon D800E? Both cameras have no antialiasing filter, which evens out the playing field somewhat. I’ll also go on later to look at noise, tonal rendition, and ultimate image potential, which is to say, what I can do with those files in the conversion.
A note on testing methodology: for the direct A-B comparisons, both cameras were shot in lossless compressed RAW and converted via ACR 6.7 final release. The D800E files were converted to black and white with a straight desaturation in ACR, and the M-Monochrom files upsized via bicubic smoother to match the output resolution of the D800E for the real image comparisons, and the D800E downsized for one set of the noise/ resolution comparisons to see if the comparison holds both ways. If sharpening was applied, it was applied consistently to both sets of images (and very minimally at that). The lens used on the MM was the 50/2 APO-Summicron-M ASPH (50 AA), the best lens that Leica currently makes; I didn’t have anything comparable to use on the Nikon (a Zeiss 2/50 Makro-Planar would have been perfect) – the closest thing I had was the Nikon 45/2.8 AI-P pancake, which is actually a fairly competent lens. The 45P resolves well at the focal plane, but lacks the flat-field and cross-frame consistency of the 50 AA – for all but the portrait comparisons, both lenses were shots at f5.6 or f8 to achieve maximum resolution. The D800E was focused with live view and magnification, and where possible, the camera moved to match framing (obviously impossible for the distant shots).
As usual, go by what I say; do not make any conclusions from the actual images (which are there for illustration purposes only) – I’ve been looking at many full size, uncompressed images on a calibrated monitor.
Resolution and pixel acuity
Full test scene. Leica MM, 50/2 APO
Center crop. Full size
Corner crop. Full size
It’s pretty clear that the MM is very much holding its own against the D800E; or perhaps that should be the D800E is holding its own against the MM – at least in the center. The corners tell a very different story; this is more a testament to the resolving power and cross-frame consistency of the 50 AA than anything. Even at f8, the 45P lacks the bite and crispness of the 50 AA; note especially definition of the crane cables. Slight magnification differences aside, I’d say the M9/ 50 AA combination is resolving ever so slightly more than the D800E and 45P; look closely at the antenna sticking out of the roof box. It may be a different story if I’d had a better lens, but I doubt we’re going to get much more center resolution out of any combination on the D800E – I certainly haven’t seen it with any of my other lenses, including the 85/2.8 PCE.
Second scene, full image. Leica MM and 50/2 APO
Center crop. Full size
The second scene appears to be much closer in terms of resolution; it was shot at f5.6 on both lenses. The 50 AA actually has a slight advantage here as the lens is now three stops down from maximum, but the 45P is only two stops down. Both cameras resolve the foliage well, and texture in the pavement and road is retained – just. If you take a close look at the motorcycle’s wheel spokes and license plate, it seems like the MM is once again resolving a hair more detail, but there’s really not a lot in it – in fact, it could well be false detail due to aliasing at this point. I wouldn’t pick one combination over the other at this point.
I did a number of other comparisons of various scenes, and could only conclude that the MM resolves at least as well as the current state of the art 36MP Bayer sensor. But for the most part, there’s not a lot in it – I would not pick one camera over the other on the basis of resolution alone.
Noise
First set of noise crops. MM enlarged to match D800E native resolution. Full size. I didn’t bother with ISO 320 because they both looked almost identical to the ISO 640 crops, which is to say essentially noise-free. Highlight and shadow recovery were both set to 10/10 for each camera.
It’s hard to say which way this comparison should go – one one hand, the MM has a much higher native ISO than the D800E (320 vs 100), and no added noise from the de-Bayering; on the other hand, it does use a CCD rather than CMOS sensor, which is known for having a higher noise floor to begin with. Once again, resolution appears to be a toss-up between the two cameras; the D800E clearly retains more useable resolution at higher ISOs.
At the pixel level, the D800E begins pulling away from ISO 1250; the MM is probably a stop behind by ISO 2500, and nearly two stops behind by the time we get to ISO 5000. I’d put ISO 5000 as being okay on the D800E, with ISO 10k being useable for emergencies. This lowers by a stop on the MM. It’s interesting to note that despite the MM exposure being slightly brighter – the exposure settings for both cameras were identical – the noise affects not just the shadows (as with the D800E) but also clearly encroaches on the midtones, too. Lowering the exposure a fraction on the MM may have helped, but it wouldn’t reduce the amplitude of the noise – there are clearly noisy pixels that have been amped far enough that they are affecting the fine detail structure of the image.
Low ISO crops. Full size
High ISO crops. Full size
For the second set of noise tests, the D800E’s files have been reduced to match the size of the MM. I’m also looking at the highlight portion of the image. I’d say there’s no difference in noise or resolution to ISO 1250; the D800E’s files downsize reasonably well, but you can see some stairstep artifacts on the fine detail of the label – this is more likely a Photoshop artifact than a reflection on the resolving power of the camera. The story for high ISO is once again similar to before: the D800E has less noise, by 1.5-2 stops again. Curiously, the downsizing (bicubic smoother) has also reduced acuity of the D800E slightly.
Intermediate conclusion: downsize or upsize files to match, it won’t make any difference.
Dynamic range, tonality and a quick word on bokeh
Reference bears again, full size image from the above crops.
Crops; full size
Although there is a huge amount of subjectivity introduced by the B&W conversion method, it’s safe to say that in general, a straight desaturation results in the lowest contrast image. Yet the MM images always land up being less contrasty than the D800E’s; it doesn’t appear that this results in there being more dynamic range – in fact, I’d say there’s if anything slightly less useable dynamic range (look at the noise in the lens barrel that isn’t there on the D800E image). What I’m seeing is a different tonal response curve that’s more shadow-biased; it’s probably something to do with the inherent differences in sensor architecture more than anything else. Is one better than the other? Only you can answer that, because it depends very much on your intended output.
I don’t want to talk too much about bokeh, because that’s a property of the lens, not the camera, but in the crops, both lenses are delivering a pleasing out of focus rendition.
Output potential
And now is a very good time to talk about output potential: what can the cameras actually do, when the files are processed properly, in a real-world scenario? The portraits were lit by a 1000-LED daylight balanced panel (not that it matters for B&W conversions), with brightness adjusted to give a reasonable exposure to simulate daylight or indoor lighting – 1/90th at f2.8 ISO 640 or thereabouts. I’ve put a fair amount of work into the output of both cameras – basically, enough to the point that I’d be happy with the finished image.
No matter what I did to the D800E’s images, they were always slightly contrastier – especially in the skin tone highlights, which required quite a lot of tweaking to avoid borderline harshness. The MM’s files just feel tonally smoother – look at the frame contents in the bottom left corner of the image, and the model’s dress. (Clicking on any image will bring you to a the Flickr landing page, from which you can view a larger version.)
The tonal difference mentioned above – manifested as a sort of ‘lightness’ if you will – is again apparent here. I personally find the MM’s rendition a bit more pleasing to the eye, but there really isn’t much in it. Full size crops of the image follow below (you will need a screen with more than 1200 pixels across to view them); take your pick for resolution – I can’t say that one has more than the other. Microcontrast is slightly better on the MM/ 50 AA image, though.
Conclusion
Often when we are writing reviews, comparing gear, or reading reviews, it’s very easy to get carried away and land up making a huge deal out of small differences. It’s pretty clear – to me, at any rate – that both cameras are capable of producing outstanding image quality, minor differences aside. Yes, the D800E does offer a stop more useable ISO, but then the MM makes up for it with lower shutter vibration and an easier method of focusing (if you’re using manual focus lenses on the D800E) – these somewhat cancel out when you’re handholding. The biggest difference again is going to be in the method of working – I’ve continuously found the D800E requires a bit more care to get the most out of it – it doesn’t really feel like a casual, fluid camera in the same way as the Leica Ms do. However, even though it’s more fiddly to focus, it’s also a lot easier to determine whether the image is in focus or not – using both back to back really reinforces how poor the MM’s screen is. And as mentioned in the MM review – you’re going to have to recalibrate your internal vision to see luminance values rather than contrasting colors and perceptual luminance. The MM does not see in the same way as you are used to with traditional B&W conversions, which take into account some of the color information when performing the conversion.
Tonal rendition is a subjective thing; some may prefer the D800E and others the MM; personally, I feel the MM’s files have a bit more luminosity to them – it’s difficult to describe, and it could very well be a lens thing; this is definitely an endearing trait. If you do a lot of black and white work, I’d seriously consider adding the MM to your arsenal; just make sure you also have the right lenses to do it justice. For the rest of us who are content to make conversions from our conventional Bayer cameras – with the channel mixing flexibility that enables – I’ll be posting an article on black and white conversion options in the near future. Stay tuned! MT
Coda: There have been a huge number of people asking why I chose to use the ‘inferior’ 45P against the 50 AA. I want to clarify this logic here, and I continue to stand by the results of this test.
1. Aside from the single corner crop included out of curiosity, the center performance of both lenses at f5.6 or f8 at the pixel level is as good as I’ve seen out of any lens.
2. This is a sensor comparison. So we look at the center resolution of the sensor, which is the same as the edge resolution. We look at noise, dynamic range, tonal response etc – note I did not include color or microcontrast (those are also influenced by the lens). The former three properties aren’t.
3. Yes, I could have used a worse lens on the MM or a better lens on the D800E. But the reality is that nobody pays me to write these things, so I wasn’t about to go out and buy a lens I didn’t need for the sake of one test.
4. Finally, it’s a real world comparison. If I did have the MM and 50 AA, I probably would look into the camera cabinet and try to decide between that and the D800E/ 45P combination if I wanted that focal length. In the end I would select on a) noise, if I needed low light performance, b) if I needed color and c) weight. Both combinations are capable of stunning images. Both are also capable of utter rubbish. The biggest difference is the photographer, not the camera.
The M-Monochrom is available here from B&H, and the D800E is here from B&H and Amazon.
____________
Visit our Teaching Store to up your photographic game – including Photoshop Workflow DVDs and customized Email School of Photography; or go mobile with the Photography Compendium for iPad. You can also get your gear from B&H and Amazon. Prices are the same as normal, however a small portion of your purchase value is referred back to me. Thanks!
Don’t forget to like us on Facebook and join the reader Flickr group!
Images and content copyright Ming Thein | mingthein.com 2012 onwards. All rights reserved