Sensors and Sensibility : The Leica M (246) Monochrom

Each morning, a sound with the ironic name of “silk” punctuates the dawn and snatches me cruelly from a blissful slumber. I fumble for the iPad, shut off its alarm and immediately launch my email app. Nothing jolts me back into a state of consciousness quite like staring into the cold, hard reality of an inbox full of fresh new problems — each vying for my immediate attention.

A couple months ago, one of these wake-up emails bounced across whatever sequence of satellites connects Wetzlar Germany with Vancouver Canada. It was, of course, from Leica. They don’t email me often, but when they do, I know it’s because they want something.

Normally, receiving email from anyone who wants something qualifies as “undesirable.” But when Leica wants something, it usually falls into one of two categories: Either 1) they want me to speak at a Leica Akademie, or 2) they want me to test some new piece of gear. Both scenarios require an extensive amount of work, yet both are actually quite enjoyable.

This day’s email fell squarely into Category #2. “Do you have time to test a prototype of the new Monochrom?” it asked.

Truth is, I didn’t. I’d gotten embroiled in way too many projects and was already burning the candle at both ends.

“Absolutely, I have time,” I lied. “Send it to me.” Hey, I could always buy more candles to burn, but how many opportunities do I get to test a brand new Monochrom prototype?

Leica’s request was that I compare and contrast the image quality between three cameras: the original M9-based Monochrom (MM9); the latest Type 240 color M (M240); and this new M Monochrom, designated as Type 246 (M246).

Anyone who’s aware of my current photographic leanings will probably think I’m a curious choice for such a task. After all, I’m not exactly Mr. High Fidelity when it comes to photography. My photos are often grainy and/or noisy, haphazardly focused, dubiously exposed and unorthodoxly framed. But here’s the thing — the look of my photos is actually a conscious and creative choice, and not (as internet forums might suggest) one born of incompetence. As an engineer, I’m actually quite adroit at testing and analyzing products both clinically and objectively. But as a photographer, I’m free to ignore whatever the logical half of my brain suggests. I’m also free to appropriate left-brain logic and apply it creatively with the right-brain. And it’s this latter function that will dictate the direction of this article — I will first analyze the clinical performance of the M246, then use that analysis to suggest how I might be able to use it creatively.

Shades of Grey (Literal Version)

Any time I work with a new monochromatic imaging surface (film or digital), I like to see how it maps colors to tonality. Such knowledge was particularly important for this series of tests, since I would be pixel-peeping and comparing the outputs from both color and monochromatic cameras. In order to do this, I needed to make sure that any M240 color conversions would closely match the tonality of the M246 and MM9. I did this same thing back in 2012, when I wrote A Fetishist’s Guide to the Monochrom. At that time, I had to create a custom Lightroom profile to make the M9’s BW conversion match the output of the original Monochrom. Would I need to do the same here?

I shot the same X-Rite Color Checker chart under the same lighting conditions (and using the same lens) with the three cameras I’d be comparing: The MM9 and the M246 had very similar tonality. For kicks, I decided to see how the M240 file would look if I simply applied Lightoom’s default BW conversion, as shown below. Note that I was not evenly remotely concerned with focus, resolution or noise in this test, so the “softer” M246 image is indicative of nothing other than my carefree technique.

Surprisingly (and unlike the old M9), Lightroom’s default conversion of an M240 color file matches the tonality of the M246 quite closely. Sure, there are minor differences, but no wholesale zone shifts. Certainly, there’s nothing that should impact my ability to compare images from a BW-converted M240 file with those from either an M246 or an MM9. So, for the remainder of this article, all M240 files have been converted to black & white using Lightroom’s default BW conversion.

The Setup

In order to carefully compare images from three different cameras, I knew that my usual modus operandi of hand-holding and scale-focussing would not suffice. In order to perform these tests, I had to remove as much human variable from the process as possible — particularly considering that the human was me. So every comparison involved the following method:

1) Each camera photographed the same scene, under the same lighting conditions, using the same exposure.

2) Each camera photographed that scene using the same lens, which I rotated between all three cameras. In order to deliver maximum fidelity to each sensor, I used Leica’s stellar APO-SUMMICRON-M 50mm f/2 ASPH lens.

3) To reduce the possibility of external vibration, each camera was mounted to the same Gitzo tripod (a relic from a past career in landscape photography, if you can believe it).

Needless to say, such rigidity means this article’s accompanying photos look nothing like the quirky, grungy, exposure-guesstimated product that usually populates the ULTRAsomething site. Because of this, I’ve decided to divide my Monochrom test article into two parts. This first part will compare camera images from a purely clinical standpoint. By analyzing these differences, I’ll be able to predict how best to employ the new Monochrom M246 in the field. A follow up article will then take a look at my field-analysis of the M246, and whether or not my theoretical findings proved valid.

Shades of Grey (Figurative Version)

For my first round of tests, I chose to compare images at each camera’s base ISO. This means ISO 320 on both the MM9 and M246, and ISO 200 on the M240. Though this theoretically gives the M240 a 2/3 stop imaging advantage, it represents the “best case” scenario for any photographer trying to extract maximum image quality from each camera.

As part of the test process, I always took a second shot with the M240 — this time at ISO 320, which put the camera on equal ISO footing with the two dedicated monochromatic cameras.

I spent two days lugging three cameras and a tripod all over Vancouver, shooting hundreds of “test” frames illustrating various scenes. Upon analyzing all these shots, I came to the conclusion that they essentially all illustrate an identical set of similarities and differences. For this reason, I’ll use just one representative scene, and take various 100% crops from within that scene to demonstrate my discoveries.

Below is the scene as photographed by all three cameras (with the M246 version being the one displayed here).

What follows is a series of comparisons between cameras, using different crops from the above scene.

Comparison 1: Try as I might, after examining hundreds of frames — all shot with each camera at its base ISO, I could see very little resolution difference between them. Sure, the MM9 has only 76% as many pixels as the other two cameras, but it seemed no less adept at resolving detail. Comparisons between the M246 and M240 yielded results that I would consider “inconclusive.” Sometimes, the M246 seemed to resolve slightly more detail, and sometimes the M240 did. In reality, most differences boil down to the influence of “outside variables,” such as atmospheric conditions, wind, or even the subtle contrast differences inherent in the tonal variations between the way each camera mapped color to luminosity.

Comparison 2: This next comparison attempts to remove another pair of variables from the process. Specifically, the M240’s ISO has been increased to 320 in order to match that of the two Monochrom’s. Also, the MM9 file has been upsized by 15% using Photoshop’s Bicubic Smoothing algorithm, so as to match the overall pixel count of the two newer cameras.

Upsizing the MM9 files make them appear a tiny bit “softer” than either the M240 or M246 files. This is to be expected. The 2/3 stop increase in the M240’s ISO has also had a subtle effect on resolution. It’s not always immediately apparent, and is something that will very likely never be seen in a print (much less on a massively downsized web image). As such, it’s only likely to concern pixel-peepers and nerds who write camera reviews (cough).

Comparison 3: This is simply another crop, taken from a section near the center of the frame. Here, the softness of the up-sampled MM9 image is a bit more apparent, while the resolution differences between the M246 (at ISO 320) and the M240 (at ISO 200) continue to appear essentially non-existent.

Comparison 4: Same crop as above. Only now, the M240 frame has been shot at ISO 320, matching the ISO of the other two cameras. In this case (and at 100%), the slightly degraded resolution of the M240 file is apparent.

It’s important to keep in mind that the images I’m showing have had no sharpening and no noise reduction. In other words, they may be somewhat representative of how the sensor performs, but they’re not that representative of how people actually work. The following comparison shows how the resolution differences appear to be minimized significantly when a tiny bit of sharpening is applied using the Pixel Genius PK Sharpener plugin. Note that default, automatic values were used, and that the differences could be narrowed ever more by optimizing the sharpening in each file.

Comparison 5: The next comparison again uses the same image, and again selects a different crop region. The goal for this comparison was not to see which camera had the highest resolution, but to see which camera had the nicest gradation between shades of grey. On first glance, the MM9 appears to give the smoothest gradations at base ISO. However, it’s important to remember that the file’s overall micro-contrast has been softened by the upsizing process, which has the effect of greatly minimizing the visible noise. In reality, files from the MM9 (that have not been upsized) have a nearly identical noise floor as the M240 and M246 cameras.

The M246 at ISO 320 and the M240 at ISO 200 appear nearly identical. Noise levels are incredibly low in both cameras, though I can perceive a tiny bit more noise in the M240. Don’t worry if you can’t — I’ve had the benefit of looking at hundreds of images and doing blind comparison tests. I would guess that, at base ISO (200 on the M240, 320 on the M246), I’ve been able to correctly identify which file came from which camera 75% of the time — and the way I can do this is to look at the noise levels. It can be subtle, but it is perceptible under certain conditions.

Comparison 6: This is the same crop as displayed previously, only the M240 has now been shot at ISO 320 to match the two Monochrom cameras. Here, we can definitely see more noise in the M240 image. Whether or not this matters to you is entirely your decision. Frankly, to me, it doesn’t matter. Nor am I sure I’d be able to see it in an actual print. Some might even prefer the noisier file because it gives the illusion of increased detail. For example, one could easily fool themselves (and others) into believing that the noise seen on the distant mountain is actually a more detailed representation of its “trees”.

Descent Into Darkness

Once I’d seen how a mere 2/3 stop increase in ISO speed began to make M240 and M246 files start to diverge in appearance, the next step was obvious: test the cameras at steadily increasing ISO speeds. So, one evening, I set up a tripod in a safe location and watched night descend upon Vancouver. As the sun set, turning first to dusk and then to night, I photographed the scene each time the required ISO speed increased: First to ISO 800; then 1600; 3200; and finally 6400.

Here is how the scene looked at the beginning of the evening:

And here’s how the scene looked as shown in the final ISO 6400 test:

(Note: Both overview images courtesy of the M246).

As before, a single LEICA APO-SUMMICRON-M 50mm f/2 ASPH lens was shared between all three cameras — Each set of 3 photos also used the identical aperture, shutter and ISO speeds.

The first crop, shown below, is from the upper right section of the frame. Each column corresponds to a particular camera: MM9 on the left; the new M246 in the middle; and the M240 file on the right. Each row correlates with an ISO: 800 on top; then 1600 and 3200 in the middle; and ISO 6400 on the bottom. Note that, for this particular comparison, I kept the MM9 file at its native resolution, and did not up-sample it to match the M246 or M240.

At ISO 800, noise is present though not overly problematic — though it does display a mild impact on the overall resolution, particularly in the shadow regions. This is most apparent in the M240 file, and slightly apparent in the MM9. As the ISO speed rises, so do the differences between cameras.

The second crop, shown below, is from a section located nearly in the center of the frame. Again, the MM9 is on the left, the M246 is in the middle, and the M240 is on the right. Also as before, ISO values range from 800 on the top row to ISO 6400 on the bottom.

Crop 2 depicts an area with more micro detail but less subtle gradation than Crop 1. The results, however, are identical. The new Monochrom M (Type 246) is nothing short of astounding at high ISO.

Again, it’s important to point out that the images displayed in the previous two examples have had no noise reduction or sharpening applied. The M240 files benefit substantially from the application of color noise reduction, and the result of that noise reduction (coupled with some optimized sharpening) is indicated in the file below. With proper processing, the fidelity of the M240 file can be brought mostly in-line with the M9-based Monochrom — though the quality and evenness of the MM9 noise remains (to my eye) superior. Neither file comes close to matching the fidelity of the M246.

Generalities

Basically, what these files (and hundreds more like them) have shown me is this:

1. With each camera at its base ISO (meaning ISO 320 for the M246 and MM9, and ISO 200 for the M240), there is very little difference in image quality between these three cameras. Resolution, noise and contrast are extremely similar, and all three cameras deliver outstanding image quality. And yet, when I make a blind comparison between two images — one shot with the M240 and the other with the M246 — there is enough of a difference that I can still select a favourite image. And 75% of the time, my favorite file comes from the M246. Obviously, subtle differences in focusing accuracy and other variables will affect my preference, but 75% is not insignificant.
2. Since files from the original Monochrom are 24% smaller, it might not necessarily be the best choice if you need to make extremely large prints. Though as previously shown, at low ISO’s, the original Monochrom certainly holds its own against the two newer cameras.
3. That said, there’s a certain indefinable “something” that I like about the MM9 files — particularly when shot near base ISO. I’ve not yet figured out what this is, but it appears the MM9 files have slightly more out-of-camera contrast than the M240 or M246 files, which seems to give them a tiny bit more initial “punch.” Of course, post-processing will easily level any playing field, so I suspect my quasi-preference for the look of MM9 files at low ISO is something that would fail to find its way into a final print.
4. With the M240 set to the same ISO (320) as the M246, it’s quite easy to tell the images apart. The M240 is noticeably (though not offensively) noisier and, as such, has slightly less detail. In general, this difference is only visible when a file is viewed at 100%, and without any noise reduction or sharpening applied.
5. Files from the M240, when converted to B&W, get rather noisy and rather blotchy at higher ISO values.  Files from the MM9 also get noisier, but the MM9 noise is like a fine dusting of static, rather than clumps of de-mosaicing blotches. Noise from the M246 is extremely fine-grained and totally unintrusive. ISO 6400 from the M246 remains wonderfully clear and imminently usable (at least for my photography).

So which camera is “better?”

You don’t really expect me to answer that, do you? “Better” is a subjective term. What’s better for one photographer might not be better for another. I will, however, tell you what these results mean to me, and how I plan to exploit them in the coming weeks.

A Master Plan

I suspect that the M246’s low-light capabilities are going to fundamentally alter the way I use a digital camera. The reason for this is likely not what you expect — I don’t intend to alter my balance between daytime and nighttime photos, nor do I expect to start taking more photos in dimly lit interiors. In fact, I’ll still use this camera for the purpose of taking photographs outdoors and in broad daylight. So why am I saying its low light capabilities will alter the way I use a digital camera? Bear with me…

When I’m out shooting candid, “found” photos in public places, I have absolutely no control of my environment. My subjects are in motion. I’m in motion. A photo opportunity might appear 1 meter to my left, quickly disperse, and a new opportunity might suddenly present itself 5 meters to my right. I have precious little time to worry about composition, and almost no time to worry about background. Lighting conditions? They are what they are, and they change frequently between deep shadow and bright sunlight.

Because of these constraints and conditions (and because of my own personal preferences), the last few years have seen me gravitate back to film for shooting on the streets. Film (particularly black & white film) is forgiving. I can miss focus and miss exposure, and the picture still succeeds. The same isn’t always the case with a digital sensor — where errors from my hasty, reactionary shooting technique get magnified by digital’s clinical precision.

One of film’s most useful attributes is, for me, the way it gently rolls off the highlights in an overexposed area. Consider the following scenario: I’m shooting on a shady section of the street, and my subject is in that shade with me. Behind my subject, and all around it, there is brightness — sky; reflective buildings; the sunny side of the street. If I were to take this photo using a camera’s built-in light meter, my subject would appear in silhouette — mercilessly darkened, so that the rest of the scene didn’t overexpose. But my subject is the subject, and I want it to be properly exposed, and not necessarily the background. Because of the way BW film rolls off the highlights, I’m free to overexpose a scene (thus properly exposing my backlit subject). Enough highlight detail remains to prevent the image from clipping to pure white. But with digital, this hasn’t necessarily been possible. Digital cameras do not gently roll off the highlights. They clip them — and all the more so on a monochromatic camera. If I use a digital camera and want a backlit subject exposed properly, the resulting photograph will possess a plethora of blown highlights. If I want to protect against blowing the highlights, then digital cameras require I do the opposite of what I do with film: Instead of exposing for my backlit subject and overexposing the background (as I would do with film), digital cameras require that I expose for the background and underexpose my subject, which I then must try to brighten in post-processing.

You see where this is going now, don’t you? Cameras with poor low light performance have noisy shadows and very little shadow detail. If my subject actually resides in those shadows (as is often the case “on the streets,”) then I need to brighten that subject in post-processing. But bringing up the shadows brings up any noise that resided within those shadows. Also, shadow noise tends to seriously obfuscate actual image details. The result? With most digital cameras, I have to choose between one of two evils: huge expanses of blown highlights, or a gritty, blurry, low-res subject. Take a look at my photo oevre for the last few years, and you’ll see which evil I’ve partnered with.

But the M246 is likely going to change all of this. Its stellar low light performance means I’ll be able to play digital contrarian, and “expose to the left.” This will keep my highlights in check. And because of the M246’s low-light capabilities, I know there will be enough clean and meaningful shadow detail that I’ll be able to pull up the brightness of my subjects in post.

But there’s more: Even in the absence of strong backlighting and high contrast, I’ll still be able to take advantage of the M246’s low light performance in ways that might not necessarily be intuitive. Remember when I mentioned how both I and my subjects are usually in motion? Better low light capabilities mean I’ll be able to shoot at higher ISO’s, thus using a shutter speed that will be fast enough to counter any motion blur. And remember when I said subjects can instantly appear at unexpected distances? This is precisely why I often scale focus a camera — quickly guessing my subject’s distance and rotating the lens to match. Alas, my guesses are not always accurate. So to improve the likelihood that my subject will actually fall within the lens’ depth-of-field, it would help if I could use a smaller aperture. Smaller apertures generally mean shooting at higher ISO speeds — not a problem with the M246.

Conclusion

As this is only stage 1 of my Monochrom M246 investigation, it’s still too early to draw any definitive conclusions. But I’ve seen enough to know that I want to take the M246 out onto the streets, implement my master plan, and see if it really does become the “game changing” camera I expect it to be for my brand of candid, observational photography.

So, until I check back in with that report, here’s a fun little photo to ponder: I was working in my studio late one night, and had just quickly patched up a sound designed to drone away in the background while I got some work done. The only external light source in the room came from a pair of IKEA book lights — each with only a single LED. They provided enough light for me to see the synth’s jacks, knobs and wires; but not enough light that I could actually read much of the text imprinted on its panels.

“I wonder what those cameras would do with this scene?” I thought. Ten minutes later I had my answer.

As with the “Cityscape : Crop 3” image shown earlier, this one shows a direct comparison between the three cameras with all noise reduction and sharpening disabled. The M240 column shown at the far right represents the “best” I could make that file look — I fine tuned the color noise reduction and the sharpening to get what I believe was the best result. This shows how much image data can still be extracted from such a noisy M240 file, but it also shows how far removed the quality is from an unaltered, unprocessed M246 file. By the way, I consider it a minor miracle I was able to include the MM9 image at all — it was so dark, I couldn’t see enough contrast in the viewfinder to focus accurately. It took a few tries to get the focus right. The M240 and M246, on the other hand, offer Live View, which made it extremely fast and easy to focus on the synth. Score one for modern technology!

So that’s it for now. The next article in this series, Sentences and Sensibility, is far less pedantic and contains photos that seek to determine whether or not the M246 is capable of stooping to my level…

CONTINUED IN PART 2…