So now that my goal is to shoot more black and white in the digital age, how do I obtain the best B&W image? Does one revert back to film for the analog effect of film for the highs and low or obtain the latest Leica M Monochrom that has a monochromatic sensor or implore. Should one continue to shoot in color and use Silver EFX Pro and/or the Lightroom B&W conversions for the best B&W results? This has been my exploration over the last several months.
I must admit that I have always had the feeling that shooting an image in color and then converting it to black and white felt like cheating. I mean, I am not really taking B&W image, its kind of an after effect. I am not working in medium, not thinking in terms of shades of gray, contrast, light and dark, red filters, highlights, etc. It must be that the Leica MM is the answer. Shooting digital but with a B&W only sensor is back to basic. But boy is it expensive and the lenses even more so. But I put my order in and I waited, waited, waited and waited. Finally I just got impatient and went with a converted monochromatic Canon 5DII. I figured it had the following advantages over the Leica MM; cheaper, less expensive lenses, better ability for longer exposures, through the lens autofocus, etc. Obviously, I had never used a Leica or rangefinder camera and therefore did not have that much invested in the Leica mystique.
Thorsten Overgaard has an excellent web site on the new Leica M Monochrom (MM) and in comparing the shooting styles of Overgaard and Thein (from my previous post) they have a very similar “photojournalists or reportage” approach, one that I rather like. But to the point, Overgaard has buried down in his Leica MM pages a discussion called “How 18 MP becomes 36 MP” in which he discusses the implied advantage of a monochromatic sensor over one with a Bayer array as, for example the advantage of the Leica MM over the Leica M9. As the heading indicates, he puts forth the belief that a monochromatic sensor without the Bayer Array will significantly outperform the same sensor with a Bayer Array. In essence, the argument is that a monochromatic sensor reads luminance only and is not concerned with color conversion caused by the color filters on top of the photosites. This monochromatic sensor sharpness advantage should also work with the converted Canon 5D II sensor just as well as on the Leica MM since it has no Bayer Array.
So advance forward I find myself in front of an old German barn in Pennsylvania, just south of Carlisle on a sunny morning with the converted Canon 5D Mark II and a Nikon D600. On the Canon is the Canon 24-105mm f/4 zoom and on the Nikon is the Nikon 24-120mm f/4G zoom; approximately equivalent lenses. The Canon is on a tripod and I am hand holding the Nikon. Here is the scene:
Now I get home and I compare the images from the converted Canon and the D600; 21.1 megapixels versus 24.3 megapixels respectively and I could not tell the difference. (The Canon is also without the anti-aliasing filter and the Nikon has it, so the canon should be slightly sharper as a result.)
First of all, the only sharpening I did was to apply screen sharpening to the jpg files using Photokit Output sharpening and I applied the same sharpening to all images. There was no individual sharpening done to either of the raw files. The enlargement is at 100% and it is hard to present a clear reading of the images as the process of converting to jpg and placing on the web site has a tendency to blur the images a little. In any case they are very close with perhaps there is a slight edge going to the 5D II. However, as mentioned above, the Canon has no AA filter so it should have a slight, but there is not the significant increase that Overgaard talked about.
Separately, as Overgaard mentioned in his article is the test results from Lens Rental where they tested the difference between the Leica MM and the Leica M9 on Imatest. (MM MM Good). To quote Lens Rental “The difference between the two cameras was significant, but just barely so. In other words, you should be able to see a difference in a reasonably large print, but you’d probably have to look at them side-by-side to tell which was the MM and which the M9 (Assuming you made the M9 image black and white, of course).” (In a future blog post, I will also compare the Leica MM to the Nikon D600 and the Canon 5D II. I am presenting this information in the order of my discovery.)
Why is this so? Why is there not the big difference between the sensor without the Bayer Array and the one with it? I believe there are two separate issues at work. First of all, the Overgaard theory, and I use him to represent others how put forth similar contentions, is predicated on the fact that a Bayer sensor will receive less stimulus than a monochromatic sensor because there can be scenes were there will be less of one color cast. “Now we can understand what Leica has accomplished by creating a monochrome sensor equipped camera. The red and blue cells in a Bayer camera are used to primarily derive color information while in the MM all pixels are used solely for luminance information, and therefore the sensor has greater resolution than an equivalent color sensor. In that regard the MM’s resolution is at least equivalent to 24 Megapixels in a Bayer camera, other factors excluded.” – quote from “Say Hello to Henri” on Luminous Landscape; “
Since the sensor is divided into red, green and blue, it is possible to be taking a picture of a city scene after sunset on a clear evening in which there is a dominance of blue. Since blue pixels only represent 1/4 of the total pixies of a sensor, shouldn’t the image be less sharp than a monochromatic sensor that uses all of its pixels all the time? This is the question I asked Thom Hogan. His response was: “The red, blue, and green sensitivities overlap a bit, so in blue light you’d get lots of blue response, some green, little red. Indoors in reddish light (incandescent) you get lots of red, some green, and little blue. White balance is generally done as a rotation of red and blue around green. Thus, in blue light you lower the blue values and raise the red values. So what happens is you get noise from the red channel, opposite in incandescent light. It’s rare that a channel is completely zero. It’s almost impossible to make the green channel zero. But there are cases where you can get the appearance of reduced resolution, as I outline in the sensor section of my books.” (which would be the case of almost totally red and black in the image) So here is the answer put simply, it is almost impossible to have any of the channels go to zero which means that in virtually all cases a 24 megapixel camera will act like a 24 megapixel camera. The reality of the scene with the barn presented above is that there was no unique lighting situation to handicap the Nikon’s Bayer sensor and as a result there was no inherent advantage for the monochromatic sensor, other things being equal.
I believe another element that plays a roll in compromising the difference between the two sensors is that a gray scale closely matches the color grid. For example green is a mid tone gray. Since there are twice as many green pixels as red and blue and as Thom Hogan pointed out, it is almost impossible to have the green pixels go to zero, there will be no advantage for a monochromatic sensor. Below is an image of color pencils and the image converted to a gray scale. Notice how the colors, particularly green fall into the areas where the Bayer sensors do their work.
So, my unscientific conclusion is that there is not significant advantage to purchasing a monochromatic body over a colored sensor in terms of sharpness all other things being equal; i.e.., same manufacturer and no anti-aliasing filter. But how about in producing good black and white? Well, in my next post I will demonstrate how shooting with a monochromatic is a disadvantage rather than an advantage over a colored sensor.