B/W Conversion Methods - More Thoughts

[John R Smith]

Just to take this topic a little further, if I may. Since I asked the original question here on the LL Forum, I have spent a lot of time downloading and evaluating various bits of software, as Mark and others advised, to see for myself what options are available for the B/W worker these days who has to convert colour digital files to monochrome as a first step in the process of producing a decent B/W print. For the last several years I have been using a sort of hybrid method for my photography – shooting B/W film, developing and then scanning the negs to 16-bit greyscale TIFFs, and then working with those files on the PC in a photo-editor before printing from an Epson R2400. So I was always working in greyscale from the outset, and never had to encounter the problem of converting a colour file. Now, with my new MF digital back, I am faced with a totally new set of decisions and choices, which I am finding very confusing.

The difficulty I have is that you chaps here on the Forum are so knowledgeable on the mechanisms of digital editing that when you do give advice it is often quite difficult for me to assess its value, because the underlying technical stuff can be just gobbledygook to me. OK, so that’s my problem, not yours. It rather reminds me of a dear friend back in the ‘70s who had a Nikon F and took hundreds of photos, always at f8 because he was firmly convinced that this magic number somehow gave the best results. That’s quite likely where I am now, with my L channel conversion. I suppose that I should mention here that at this stage I am not trying to do anything more than I was already doing with my negative scans, I’m just trying to get as good a result as I was getting from film. I was perfectly content with the fact that my film scans were a pretty inflexible end product, and that I could not do a lot more to them than basic levels, contrast, dodging and burning. Pretty much the same as I used to do in the wet darkroom, really.

I carried out an interesting little experiment a couple of nights ago. Having downloaded and played around a bit with trial versions of all sorts of applications and plugins (as recommended here and realising that it would take many weeks, not days, to get to grips with something like ACR), I started out with a new neg (sorry, I mean file) as a colour RAW from the camera and processed it to 16-bit greyscale using five different methods –

1) 3FR>Phocus>16-bit colour TIFF>PS>Lab color>extract L channel>16 bit gray TIFF

2) 3FR>Phocus>16-bit colour TIFF>PS>Convert to BW Pro>16 bit gray TIFF

3) 3FR>Phocus>16-bit colour TIFF>PS>Nik Silver Efex Pro>16 bit gray TIFF

4) DNG>ACR desat and levels>PS>16-bit gray TIFF

5) DNG>Raw Therapee desat and levels>PS>16-bit gray TIFF

OK, so I don’t really know what I’m doing (manual in one hand, mouse in the other). In ACR and RT I tried to do all the edits in RAW, as you folks were saying. What I did do was start from scratch each time without any reference to my previous attempt at all, and just try to get the picture looking “right” to my eyes, then save it. Then I printed them all out as workprints, and pinned them up to think about them. The really strange thing was that they all looked virtually identical – there were differences, but extremely slight. No more than the variation I would have expected to get in my darkroom over half an hour or so with the dev temperature dropping a bit. Obviously I was not trying to do anything fancy with filtration, black skies, etc (that’s not my bag), just a completely “straight” print. So I’m not at all sure which was “best”, but I do know which were the easiest.

In principle I can quite see the point that doing as much processing in RAW as possible must be highly desirable, and the least likely to cause banding, posterisation etc. I suppose that up to now, if I have a picture which needed a great deal of levels or contrast adjustment I have just dumped it anyway, because it was a rubbish shot and I should have done a better job in the camera. So I tend to print only those frames which are well-exposed, punchy, and need very little editing. The rest get binned. With my Hasselblad files, the TIFFs that Phocus produces are really gorgeous with no editing at all (provided that your exposure is on target), so the temptation is just to start with them.

So I’m struggling on, this is just to keep you up to date.

John

[stamper]

I don't think you will ever find the "best" method. Like everything else it is subjective. If you keep searching you will become frustrated. I am not saying that you should give up. Just settle on one that "appeals" to you and go for it.

[Alan Goldhammer]

John,

One resource that I have found to be useful is:  Leslie Alsheimer and Bryan O'Neil Hughes, "Black and White in Photoshop CS4 and Photoshop Lightroom" covers all aspects of black and white transformation using the two key Adobe software products.  The book is pretty thorough on all the possible conversion methods but does not include the NIK program.  It's also reasonably priced.  I've found that you can drive yourself crazy moving back and forth between different conversion approaches and that ultimately you need to compare what comes out of the printer; not what might be appealing on your computer screen.

alan

[John R Smith]

Yes, Alan, I am going to have to buy some books on this stuff. Next step is to try out Capture One - I just downloaded a trial version - but the problem is it may not run on my aged PC. CS4 wouldn't run properly, although strangely ACR did, so I could save the file out of ACR and open it in PS7 OK. The wierd thing is, my old software loads and edits these huge 16-bit files (230MB) much quicker than the new stuff does, and never crashes. Whereas CS4 trial version would not run for more tha five minutes without a total crash.

John

[Guillermo Luijk]

The difficulty I have is that you chaps here on the Forum are so knowledgeable on the mechanisms of digital editing that when you do give advice it is often quite difficult for me to assess its value, because the underlying technical stuff can be just gobbledygook to me.

Perhaps the best solution is to apply film and colour theory. I will just translate something a friend (Manuel Portillo) posted in a forum some days ago about B&W conversion (BTW all his B&W images look great):

"(...) but I think all those plugins are not necessary to get a decent B&W.

I think it's better to understand how filters work, whether physical or electronic, and how the human eye perceives luminosity for the different colours in the visible spectrum.
Starting from this point no special software is needed to obtain a good B&W according to your taste, clearly differentiating tones, adjusting them to the eye's perception or exagerating them to your like.
With Photoshop there are one thousand ways and more than 5 tools to achieve the same result.

An example:

1. Colour image developed with ACR's default values, close to camera's JPEG.
An uninteresting picture in this state that was thought from the shot to become a B&W image enhancing the sky. The camera captured in a low contrast, without any dramatism nor interest, much duller than the scene was.




2. Camera process and conversion to B&W. Everything is plain gray, like if we used film without any optical filter at shooting time.




3. Here converted to B&W in Photoshop from the RAW file, imitating the effect of a high contrast red filtering in B&W film.



I just used two tools in Photoshop to convert to B&W.
The "black and white" tool, starting from the "high contrast red filter" preset, that violently darkens the blues, cyans and greens, and enlightens the reds and yellows. The same thing that a red filter would do on a B&W film.

3.1. I adjusted the blue, cyan, green, red and yellow sliders so that the gray differences in the different colours fitted my taste.
3.2. I increased accutance (local contrast) in the bush area because colours didn't display well differentiated and contrasted in spite of the filtering.

As you can see, it is not complicated at all to obtain a custom B&W (I don't mean better or worse, that is relative), but you can enhance contrasts and differentiate colours in a much simpler way than you would shooting film."

Regards

[EduPerez]

My intuition leads me to think that the L channel technique should give results quite similar to B&W film; after all, what I would expect from B&W film is capture the amount of light coming from each part of the scene, and that is the theory behind the L channel. So, if you want to start with a workflow similar to using B&W film, this should be a good starting point.

RawTherapee's desaturation works precisely by zero-ing the A and B channels in LAB, by the way.

[john beardsworth]

My intuition leads me to think that the L channel technique should give results quite similar to B&W film; after all, what I would expect from B&W film is capture the amount of light coming from each part of the scene, and that is the theory behind the L channel. So, if you want to start with a workflow similar to using B&W film, this should be a good starting point.

RawTherapee's desaturation works precisely by zero-ing the A and B channels in LAB, by the way.

I'm not convinced intuition is helpful here! The L channel technique gives results that reflect the L element in the LAB method of recording brightness and colour. That may, by chance (OK a bit more) be similar to the results with a particular B&W film with no lens filter - but that's coincidental. It's best to think of the method as producing a pleasant, neutral result, but while having little opportunity to interpret the image as part of the conversion stage.

Equally, "if you want to start with a workflow similar to using B&W film" is not a good starting point. The original poster should start with a colour image and want the best possible B&W result - not be distracted by pretending he's using B&W film.

The original poster is still focussing far too much on conversion tools, as if it's just a matter of tool A or tool B producing qualitatively the best conversion, whatever that is. Guillermo got it right by emphasizing how the same technique, applied with judgement, results in very different conversions of the same image. Notice in the "red filter" result how the tonal separation is much greater - how the clouds stand out from the sky. That separation is a key aspect of the conversion stage - equally if not more important than "quality". Maybe  that effect seems too strong, so reduce it - you have that creative control with the more modern techniques.

John

[John R Smith]

Equally, "if you want to start with a workflow similar to using B&W film" is not a good starting point. The original poster should start with a colour image and want the best possible B&W result - not be distracted by pretending he's using B&W film.

John

Good points made above, but in some ways I think EduPerez has come nearer the mark for what I am trying to achieve. For me, colour is the distraction. I suppose I am "pretending" I'm using B/W film, fair enough - that's because I really don't want to see colour at any part of the process, but I am forced to by this digital back. If I could have bought a monochrome digital back I would have done. At the point when I am taking the photograph is when my decisions get made - when it's done, it's done. If I have done things right, the negative should print itself, really. And to be fair, my good shots from the CFV via Phocus do pretty much print themselves, with the same sort of work as I would do in the darkroom - basic paper exposure (gamma and levels), paper grade (contrast), some local dodging and burning (burn and dodge tools), and off to the printer. Great. After all, this is just my evening relaxation, quality time stuff, kicking back after a day's work with a beer and a nice shot. I don't want to raise my anxiety levels agonising over whether or not I should have applied 3% less luminance noise reduction or something, or have to spend hours reading heavy tomes about colour space transformation and gamut mapping.

The problem I have with all these methods which involve doing lots in the RAW converter, and more particularly with the things which mix the RGB channels in various ways, or do film "emulation" or whatever, is that there is too much choice. There is so much choice that I find it impossible to decide which is "best". I am paralysed by choice. When I shoot with film, once it is developed and scanned to grayscale, there is very little choice. That is actually the way I like it. Sorry, folks, I suppose I'm just a lost cause.

John

[stamper]

Your are not a lost cause. As stated in my post#2 pick one and go for it. I have tried a lot of methods and IMO it doesn't really matter which one you choose? Why? The answer is because after you have chosen one the real work begins. Layer blend modes, curve adjustments and other tweaks  - imo - make or break the image. As long as you don't desaturate an image or just change it to grayscale the conversion method will not be apparent after you have enhanced the image. I dare anyone to look at a B&W image after adjustments have been done and state without guessing which conversion method has been used?

Look at this image and state which conversion method was used?

[John R Smith]

Stamper

Well, no of course I couldn't tell what conversion method you have used at all, quite true. I can tell that you have given it a hell of a lot of contrast in the process, though, and not worried too much about losing your highlights. It's a very nice picture, but conceptually almost the opposite of what I am doing, where my pictures are very much dependent on tonal separation in the mid-greys, to get depth from front to back (sorry about the sharpening artefacts, I haven't got the hang of downsizing these files for the web yet).

John

 [attachment=20527:D03_08_O...an_River.jpg]

[Ken Bennett]

There is one line of the original post that caught my eye, and may explain a lot of the confusion here:

So I was always working in greyscale from the outset, and never had to encounter the problem of converting a colour file.

This is simply untrue. You encountered the "problem" of converting color to black+white at the moment of exposure on your film. That "problem" is the basis of untold thousands of pages over the last 150 years dealing with b+w film photography, the use of filters, exposure and processing techniques, etc. The photographer has many tools available to manage the conversion from color ("real life") to grayscale (b+w film). Pretending that there is some magical "correct" black and white interpretation of any real life scene -- and that you can automate the process in digital photography -- well, it's just not that simple.

I agree completely with Guillermo that the proper application of color theory gives the most control over the b+w conversion process. This is true whether you are shooting b+w film or color digital and converting in Photoshop or some other application.

And, I agree with several others that you should pick one technique and really learn how to use it. You may, at some later point, decide that you want to try some others, but then you'll be in a much better position to judge what you like and dislike.

[John R Smith]

This is simply untrue. You encountered the "problem" of converting color to black+white at the moment of exposure on your film. That "problem" is the basis of untold thousands of pages over the last 150 years dealing with b+w film photography, the use of filters, exposure and processing techniques, etc. The photographer has many tools available to manage the conversion from color ("real life") to grayscale (b+w film). Pretending that there is some magical "correct" black and white interpretation of any real life scene -- and that you can automate the process in digital photography -- well, it's just not that simple.

Well, I ignored the moment of exposure part of it because I didn't want to muddy the issue. My camera kit of course contains a full range of colour filters in all the sizes for my various lenses, and has for the past forty years. In recent years, however, I have tended to use them less and less, partly because modern films seem to hold the skies in better than they used to without a lot of help. I've never seen the moment of capture to be a "problem" in that regard, really, although of course sometimes I get it wrong. Mostly, with film, I just shoot with a light yellow filter to pop the clouds a bit, and that's it.

John

[soboyle]

I find the tools in lightroom, or the B&W adjustment layer in photoshop to
work very well for making B&W conversions.
You can tweek how the various colors in the image render to B&W, and can come back and adjust these later if so inclined.

[stamper]

Stamper

Well, no of course I couldn't tell what conversion method you have used at all, quite true. I can tell that you have given it a hell of a lot of contrast in the process, though, and not worried too much about losing your highlights. It's a very nice picture, but conceptually almost the opposite of what I am doing, where my pictures are very much dependent on tonal separation in the mid-greys, to get depth from front to back (sorry about the sharpening artefacts, I haven't got the hang of downsizing these files for the web yet).

John

 [attachment=20527:D03_08_O...an_River.jpg]

The contrast was deliberate and I added grain to give it a gritty look. A matter of taste? I do B&W which is a lot less contrasty. There is no "right or wrong" in B&W because it is in essence an art form and not an out of the camera type image that the purists like to claim is the only way of doing photography? Less likely to get criticized for over manipulating an image. Good luck in your endeavours.