I've been pursuing digital B&W in earnest for almost 12 months now and although I'm still not convinced it surpasses B&W film. However, this view needs to be couched in what is "my" framework for evaluating digital B&W. B&W to me is about the tonality; good strong blacks and whites with midtones that don't "grey" out the image. On the other hand, I dislike hard "pepper & salt" B&W images, devoid of midtones with absolutely no shadow details. Hopefully this provides an outline as to what I'm chasing...
Recently, Leica release its M Monochrom dedicated B&W version of the M9, but as much as I'm both a B&W fan and an avid Leica M user, I have reservations about this camera. The first thing it does especially when paired with Leica glass, is offer oodles of resolution - and that is where my problem begins! To me, B&W is an abstraction and uber sharpness just doesn't feel right or necessary. In addition, in recent weeks, even Leica has been recognising the work of a number of named photographers who were early adopters of the M Monochrom. This started me evaluating again...
The images above are a small crop from one of these photographers images, which I am only using to navigate to my next point about digital B&W conversions. The top crop is the original (in the sense this is how the original photographer chose to render the image). To my eye this offers an image too smooth for a B&W feel so I started looking at generating simulated grain. The following 3 crops have had nothing more than grain added via three different applications.
The grain filter supplied with PS CS5 is not too bad at all; it lacks some subtlety in its application but overall it was acceptable. The only "got ya" and its a show stopper is that it only works on 8-bit images.
Nik Silver Efex seems to be everybody's favourite for B&W conversions. I tried it and have been using it in parts for most of the last year. It's been a like, but mostly hate, relationship as I find it useful for somethings but never capable of producing a consistent "look" B&W conversion to begin working from. I custom work every image, however, because digital is so malleable I find it very difficult to maintain a consistent look across conversions with Nik Silver Efex. In addition, which is more retrospective now, I feel a good part of the problem is 1) Nik Software having a "half-hearted" attempt at providing a list of film stock appearance which in turn 2) is separate from the treatment of the image. In the end I'm never happy with the output.
Given I'm referring to the crops above, I have only seriously investigated the simulated grain component recently on the back of looking at Leica M Monochrom images. The original upper image - especially the complete image - was turning out to be my B&W image equivalent of fingernails down a blackboard.
A downside I was seeing with Nik Silver Efex, specifically as it related to grain simulation, was no ability to control grain in the highlights - especially as they started to blow out. As I hunted around for options I came across DxO Labs FilmPack 3, which solved a number of issues I wasn't even looking to solve at the time. Basically, the software provides a number of pre-canned B&W film stocks (also has transparency and colour neg offerings too), and in the expert version, which I would recommend, allows you control over the intensity of the effects as well as colour filtering.
So what did DxO FilmPack 3 solve... From the lower crop above, for me, it gave the best simulated grain available, which comes back to my initial framework of what B&W is to me. I've come to the conclusion that one important aspect of B&W tonality comes from the grain structure. When its not there, even fine grain, something just doesn't feel right. The other issue it solved was achieving a good and consistent digital conversion-base. I don't get too caught up in what specific film stock the selection is trying to emulate as this in itself is somewhat impacted by the colour rendering of the image being offered to the software to begin with. Rather, I tend to choose between four or five that work for me tending to increase grain on fine structure films stocks or reducing it on the grainier ones. The result being that I can consistently get very good base conversions before starting any custom work.
The outcome of this has resulted in firstly purchasing the software well within the 31 day trial limit, and secondly - and probably more importantly - I'm now redoing EVERY digital B&W conversion I've done over the last year. The difference is that good; output looks much more like film and for some conversions that I could "never" get right, DxO Filmpack solved.
With no affiliation to DxO, in my opinion, this is the best B&W conversion software on the market today...
Recently, Leica release its M Monochrom dedicated B&W version of the M9, but as much as I'm both a B&W fan and an avid Leica M user, I have reservations about this camera. The first thing it does especially when paired with Leica glass, is offer oodles of resolution - and that is where my problem begins! To me, B&W is an abstraction and uber sharpness just doesn't feel right or necessary. In addition, in recent weeks, even Leica has been recognising the work of a number of named photographers who were early adopters of the M Monochrom. This started me evaluating again...
The images above are a small crop from one of these photographers images, which I am only using to navigate to my next point about digital B&W conversions. The top crop is the original (in the sense this is how the original photographer chose to render the image). To my eye this offers an image too smooth for a B&W feel so I started looking at generating simulated grain. The following 3 crops have had nothing more than grain added via three different applications.
The grain filter supplied with PS CS5 is not too bad at all; it lacks some subtlety in its application but overall it was acceptable. The only "got ya" and its a show stopper is that it only works on 8-bit images.
Nik Silver Efex seems to be everybody's favourite for B&W conversions. I tried it and have been using it in parts for most of the last year. It's been a like, but mostly hate, relationship as I find it useful for somethings but never capable of producing a consistent "look" B&W conversion to begin working from. I custom work every image, however, because digital is so malleable I find it very difficult to maintain a consistent look across conversions with Nik Silver Efex. In addition, which is more retrospective now, I feel a good part of the problem is 1) Nik Software having a "half-hearted" attempt at providing a list of film stock appearance which in turn 2) is separate from the treatment of the image. In the end I'm never happy with the output.
Given I'm referring to the crops above, I have only seriously investigated the simulated grain component recently on the back of looking at Leica M Monochrom images. The original upper image - especially the complete image - was turning out to be my B&W image equivalent of fingernails down a blackboard.
A downside I was seeing with Nik Silver Efex, specifically as it related to grain simulation, was no ability to control grain in the highlights - especially as they started to blow out. As I hunted around for options I came across DxO Labs FilmPack 3, which solved a number of issues I wasn't even looking to solve at the time. Basically, the software provides a number of pre-canned B&W film stocks (also has transparency and colour neg offerings too), and in the expert version, which I would recommend, allows you control over the intensity of the effects as well as colour filtering.
So what did DxO FilmPack 3 solve... From the lower crop above, for me, it gave the best simulated grain available, which comes back to my initial framework of what B&W is to me. I've come to the conclusion that one important aspect of B&W tonality comes from the grain structure. When its not there, even fine grain, something just doesn't feel right. The other issue it solved was achieving a good and consistent digital conversion-base. I don't get too caught up in what specific film stock the selection is trying to emulate as this in itself is somewhat impacted by the colour rendering of the image being offered to the software to begin with. Rather, I tend to choose between four or five that work for me tending to increase grain on fine structure films stocks or reducing it on the grainier ones. The result being that I can consistently get very good base conversions before starting any custom work.
Same location: top is HP5+, lower is Nex 5n + DxO Labs Filmpack 3
The outcome of this has resulted in firstly purchasing the software well within the 31 day trial limit, and secondly - and probably more importantly - I'm now redoing EVERY digital B&W conversion I've done over the last year. The difference is that good; output looks much more like film and for some conversions that I could "never" get right, DxO Filmpack solved.
With no affiliation to DxO, in my opinion, this is the best B&W conversion software on the market today...
