Musings on Photography

Black and White

Posted in software by Paul Butzi on March 5, 2008

Back when I did everything with a large format camera, I did everything in B&W. When I got the EOS-5d, I made a sort of mental commitment to really give color a try – do everything in color for a year, at least. It was a rough adjustment.

But lately, I’ve been thinking about B&W again. That decision between B&W and color feels different now that I feel comfortable doing both B&W and color. Maybe I’m ready to live with one foot in each world.

Back when I was doggedly pursuing B&W in the wet darkroom, I had lots of honed technique to apply to the problems of printing B&W. Then I switched to printing in the digital domain, and all of a sudden I was plunged into the world of scanning, etc. Just before I bought the 5d and switched to a digital workflow start to finish, I’d done a bunch of experiments with tweaking development of B&W film to get the best possible scans, and I had on my plate the idea that I’d try shooting E6 color film, scanning that, and then converting the color scans to B&W in photoshop. Color transparency film scans well, and preserving the color info through to photoshop would have let me, perhaps, leave about two pounds of color filters behind. Plus E6 film processing is something I felt I could entrust to a lab.

Well, the 5d came along and put an end to all those experiments. But the thought that you might get really good, malleable results in B&W by exposing in color and doing the translation to monochrome in the digital domain was stuck in my mind. The noisefree nature of the images from the 5d gives you B&W that has that large format creamy goodness.

The problem I run into is this: when I’m converting from color to B&W using the channel mixer in Photoshop, I’ve got four sliders to fiddle with, and it always seems not very productive to go searching around in four-space trying to find the optimum result. Today it occurred to me that the problem of optimizing results by searching through an n dimensional space of parameters is not that hard – computers run simplex searches to find optimums in multidimensional spaces all the time. And now I’m wondering how hard it would be to write a simple script or plugin for photoshop that would take a color image, and do the hard work of fiddling with the color mixer to get the ‘best’ result, either automatically or guided by a human driver.

Anyone have any experience with writing scripts or plugins for Photoshop?

17 Responses

Subscribe to comments with RSS.

  1. Nick said, on March 5, 2008 at 6:56 pm

    The most recent episode of photoshop user TV has a great tutorial on creating…. agh, have to look up the term. OK, they don’t list the term on the website. Basically, you can record a series of actions. Turn on this feature, make the adjustments you want, and then you can run those same series of actions on any other file…. not sure if you can apply that to multiple files.

    You’ve peaked my interest. I think I’ll look into it… Unfortunately I’m not going to be able to get to it for a couple of days… (maybe tomorrow, but unlikely.)

  2. Tommy Williams said, on March 5, 2008 at 7:32 pm

    It sounds like you live in ACR and Photoshop but you might consider Lightroom for your RAW conversions. It has what–to me–seems a much more intuitive way of handling conversion from color to grayscale and even has an “auto” setting for adjusting the relative balance of the 8 colors in the grayscale conversion.

    ACR and Lightroom have the same RAW conversion engine but the workflow, for me, is far better in Lightroom than with Bridge and ACR.

  3. Anil said, on March 5, 2008 at 7:39 pm

    I use a Action to convert from color to B&W and I dont change anything after that. Works just fine for my needs. Can attach and email you the PSCS3 action if you would like to try it out.
    The action is really quite simple and uses a combination of Channel Mixer, Selective Color and Color Balance.

  4. JeffK said, on March 5, 2008 at 8:02 pm

    Check out Dave Beckerman’s latest stuff.

    He’s sold his film cameras and is using a 40D doing exactly what you desire. He’s using Lightroom for the conversions.

  5. Ed Richards said, on March 5, 2008 at 8:04 pm

    Use the new black and white converter in CS3. You can even do it as a layer and keep tweaking. As for a script – the action will not be too hard, but programming the robot to watch the screen and pick out the best conversion is going to be pretty time consuming.:-) In other words, with the black and white converter, conversion is an aesthetic judgment, not an optimization.

  6. ejeschke said, on March 5, 2008 at 10:20 pm

    If I examine each channel (R, G, B) separately first, I find it gives me great clues about how to narrow down the channel mixing settings immediately.
    I find that one channel often jumps out at me with the interesting tones in the right places. Often I will just pick one of the channels and go with that, bypassing the mixer entirely.

    If there is more than one interesting channel, then it usually goes something like–its the blue or red channel that draws my attention. Green, being the dominant player (in Bayer) often is the lowest noise and adds good fill as luminance. It becomes the dominant “stock” upon which I graft the interesting parts of the B & R channels.

    Eric Jeschke

  7. Colin Griffiths said, on March 5, 2008 at 11:23 pm

    Use Lighroom! It’s very easy to ring the changes and compare results. Alternatively Photokit contains some filters that allow you to combine filter strengths in a single image using masks.

  8. My Camera World said, on March 6, 2008 at 8:31 am

    In some ways the B&W film format was easier.

    You looked at the scene and tried to understand the gray tones.

    You selected a filter to enhance or recede some colour elements to suit to artistic vision

    Click. The easy part.

    In the darkroom, contrast selection (lens and paper and developing times), some dodging or burning to fine tune specific elements.

    I used to spend a lot of time in Photoshop trying to squeeze the most out of an image I guess only because I now have more tools.

    I would use Channel mixer (filter selection above) , or I would create 3 separate grayscale images form each channel and then add a curves layer for each of these and mask the best elements.

    Now with Photoshop CS3 I love the new B&W tool that allows you to click on any colour element within the image and drag left to darken or right to lighten just the grey-colour equivalent to set up the basic image.

    Then as before back to mask to fine-tune any elements

    Niels Henriksen

  9. Joe Reifer said, on March 6, 2008 at 8:34 am

    The Photoshop CS3 Black and White Layer method is just as easy as the Channel Mixer, and a lot more flexible. I like the DXO Filmpack Photoshops plugins on a layer as well.



  10. Gordon McGregor said, on March 6, 2008 at 9:21 am

    I’ve written photoshop scripts using the exposed object model. It isn’t too difficult – even throwing up bits of GUIs and the like isn’t that hard to grasp. All the tools and documentation is available on the Adobe website too, there are a few languages you can pick to do it in.

    They support a few languages, one PC biased, one Mac biased and javascript which should work on both. In fact one of my first ever blogposts was a javascript photoshop thing to layer multiple images. Should give you an idea of what the code looks like. If you need more pointers let me know.

    On the B&W thing, I find I can do the more straightforward stuff with either lightroom or the channel mixer. However, when I really get down to doing a B&W that I care about, I end up taking a bit of the scene from one channel, then find I like another part of the scene from another channel and then maybe a third area needs a blend of bits of multiple channels.

    So I end up with 3 layers, with 3 masks. Lets you do the channel mixer thing but vary across the image, rather than being tied to a global adjustment.

  11. Dennis Allshouse said, on March 6, 2008 at 12:10 pm

    Seems like the tough part of automating is developing the rules to judge what is the best conversion, and then how would you “look at” the image to apply the rules?

  12. Dennis Allshouse said, on March 6, 2008 at 12:11 pm

    Do you have rodis in bloom already?

  13. Paul Butzi said, on March 6, 2008 at 12:20 pm

    Seems like the tough part of automating is developing the rules to judge what is the best conversion, and then how would you “look at” the image to apply the rules?

    I wouldn’t try to automate the rules to judge what is the ‘best’ conversion. If you’re doing a simplex search in 4-space, you’d just put up 4 versions of the image, and let the user pick which one was the worst. Then you’d replace that ‘worst’ image with a new version, over and over again.

    In other words, let the viewer do the judging. It wouldn’t take to many iterations before you had it down just right.

    Do you have rodis in bloom already?

    It’s not a rhododendron; it’s much smaller and blooms much earlier. This one is also situated in a warm spot near the house that gets sun in the morning and in the afternoon. In fact, it’s right across the gravel path from the now famous Japanese Maple.

    I’d tell you what it is, but I can’t remember and the chipmunks seem to have stolen the copper tag that tells me.

  14. Hugh Alison said, on March 6, 2008 at 1:42 pm

    My 5D has pretty much been set on the monochrome picture style with the yellow filter since I bought it – always shooting in RAW – I then use Canon DPP as my RAW converter, sometimes change to orange or red filtration, tweak contrast and exposure, and export to a 16 bit TIFF.

    Only time is comes off the monochrome defaults is if I am shooting for someone else.

  15. Olli Wendelin said, on March 7, 2008 at 9:45 am

    I shoot people rather than landscapes. After some experimentation with channel mixer, I decided on a 60, 20, 20 RGB mix for a good rendition of skin tones. Then I built a simple photoshop action for “push button” B&W. This may seem simplistic, but the less I edit, the more I can shoot and enjoy life.

  16. Dave Kosiur said, on March 7, 2008 at 11:25 am

    I like using Lightroom for B&W conversions. In fact, I just finished a class at Santa Fe where that was the recommended method. One feature that’s great for B&W conversions is the targeted adjustment tool, after you convert to grayscale…

    It’s not an automated procedure, but one way of approaching image comparisons in Lightroom would be to make virtual copies of the color image and then apply a different B&W preset to each virtual copy, then view them side-by-side in the Library module.


  17. Karsten Schmidt-Hern said, on March 9, 2008 at 8:29 am

    Try Greg Gorman’s conversion:

    I don’t use it all the time, but it is very nice because it ensures repeatable results.



Comments are closed.

%d bloggers like this: