Beginner Principles of methods for B+W conversion

ianp5a

ianp5a

Messages
4,717
Name
Ian
Edit My Images
Yes
I'm interested in discovering some of the principles behind useful methods for Black and White conversion.
I've been experimenting with the colour channels to good effect. But I'm curious to find out more.

I'm not interested in any automated programs or dedicated plug ins. But the direct methods themselves.

Thanks in advance.
 
I use colour channels in Photoshop, but only as a starter, I then adjust levels, dodge or burn to suit the image.
 
Thanks everyone.
But shooting black and white film doesn't help me discover digital conversion techniques.
 
I tend to use gradient map for a more contrasty conversion - but the real key is shooting the right picture for conversion in the first place. Thinking in B&W is a real bitch
 
Speaking as a b/w film photographer, to a digital b/w photographer (maybe a little like Queer Eye for the Straight Guy?).

I feel that digital can make great b/w. It get's an unfair bad reputation, because of poor conversions, and even poorer presentations.

Personally, if I want to make b/w digital, I use channel mixer. The RGB sliders (unless I want otherwise), I'll slide around in different portions to make 100% (or I think 200% in some programs?). My favourite method is to shoot in raw mode, then open up the raw images using my rather out dated but open source UFRaw software. I convert using the Channel Mixer there, selecting to generate a monochrome .jpeg. I like to think of the channel mixer as being rather like applying colour tinted lens filters - but post process! Just like lens filters, different colours will affect the b/w image differently. Channel mixers suffice for my needs. If I do shoot in RGB .jpeg, I simply use the Channel mixer in the open source Gimp software package to generate b/w, instead of UFraw.

I'm sure that modern bloatware packages have posher ways to deal with b/w conversion. Doesn't Photoshop and it's ilk have a tool that you can select different b/w tones for different colours?

Going off subject, but following that Queer eye theme again. In my opinion, good digital b/w photographers dedicate themselves to b/w, rather than dabble in it. They may shoot in RAW, or RGB .jpeg, but always with b/w in mind - learning to ignore colour information, and to instead concentrate on lines, subjects, depth, shadows, etc. That I think is the secret behind good digital b/w. In addition, if they share online, they don't mix b/w images up with colour images. They present at least a series of b/w without the visual pollution of colour images. The worse offence (in my lowly opinion), being to put a colour image up next to the same image after a b/w conversion and ask "which is better".
 
Great. I've been converting Raw in Raw Therapee using the channels in combination with saturation. As well as the usual histogram tools. It's not always clear that I've over-cooked one slider if any side-effects are not visible until another slider has been increased. I'll just have to keep experimenting with them. There's probably a preferred order in which to adjust them.
 
Last edited:
Even if you shoot RAW, it can still be instructive to use some physical filters on the camera when shooting B+W. There are two reasons. 1: When you get a feel for what you want to do, the physical filter will give you the best exposure. For example, a red filter will make blue skies look dark. Without the filter the camera may end up metering for the blue sky (which you will want to be dark) and not the foreground. The filter will let you get a correct metering for the colour channels you actually want to use. Make sure to turn off auto white balance if doing this. 2) The view through the viewfinder will better reflect the end result so that you can learn how different filters effect contrast.

Considering that people are regularly getting rid of "useless" film related filters, you could pick up a set of red, green and blue for not too much money. I wouldn't worry about orange or yellow - they are very useful for film, but their effects will be well within the flexibility of a raw file when making adjustments.
 
ianp5a said:
Great. I've been converting Raw in Raw Therapee using the channels in combination with saturation. As well as the usual histogram tools. It's not always clear that I've over-cooked one slider if any side-effects are not visible until another slider has been increased. I'll just have to keep experimenting with them. There's probably a preferred order in which to adjust them.
Click to expand...

When I owned and used digital cameras, I used the Nik Silver Efex plug in to convert images to B&W. It's really powerful, but also easy to overcook the image and end up with something that's a bit too HDR looking. It has a few unique controls such as 'soft contrast' which can create some lovely effects. It's probably nothing that Photoshop couldn't do, but just a lot easier. Worth checking out if you're into digital B&W.
 
No thanks. I'm asking about the principles, and not about specific programs that probably don't work anyway on my PC.
 
ianp5a said:
Great. I've been converting Raw in Raw Therapee using the channels in combination with saturation. As well as the usual histogram tools. It's not always clear that I've over-cooked one slider if any side-effects are not visible until another slider has been increased. I'll just have to keep experimenting with them. There's probably a preferred order in which to adjust them.
Click to expand...
I'm no expert, but as a guide ... if you want to brighten up light skin, and reduce blemishes - bias towards red. If you want to make blues (such as a blue sky) darker, and the clouds stand out more, then bias towards red. If you want to enhance blemishes and wrinkles on aged faces, bias towards blue. I think that blue has a positive effect on foggy landscapes, but don't take my word for it - experiment. If you really want to brighten up foliage, then bias towards green. Each image might be better biased one way or the other, and it really has to be to your own discrimination and tastes.
 
Last edited:
The basic principle is that when an image is desaturated, the brightness of the colour value (the L value of the LAB colour space) is mapped to the corresponding value between pure white and pure black. If you're using an 8 bit system you will have 256 possible levels of grey, 65,536 with 16 bit etc. (Photoshop actually uses 15 bit precision, but you get the idea). If you want a completely linear response, desaturating the image doesn't bias any particular colour. I.e. a green with a lightness value of 50, will look exactly the same as a red with a lightness value of 50. Hence if they appear next to each other in the B&W image, they will merge together. In the real world, film (i.e. real film) doesn't have a linear response curve so, for instance, it may be more sensitive to red which would result in those tones being lighter in the final image. So if you shot those two same colours of red and green, they would not merge together in the final image.

Film photographers use coloured filters to bias the image towards certain colours and completely change the distribution of tones in the image. For instance, like Paul said above, adding a red filter would allow that wavelength through and make that tone lighter on the final image. This tends to make (caucasian) skin lighter and will darken the sky as it stops blue light hitting the film. To simulate these effects in software, move the slider for that colour higher and reduce the others to keep the same overall brightness level. Reducing the blue channel will have a similar effect. If you know the response curve of a particular film (i.e. exactly how it responds to each wavelength of light) you could create a digital version by setting your channel sliders to the exact values.

In summary, whenever you convert from a colour to B&W image, the software is deciding how it maps the lightness values of the coloured pixels to create the grey tones. The channel mixer, or whatever equivalent you have in your software, allows you to bias that mapping in an infinite number of ways. How you do that is part of the creative process and there is no right or wrong way. You have to look at the final B&W image and decide what looks good to your eye.
 
Thanks. Mostly colour pseudo filtering. But a mention of noise reduction of individual channels.
 
You must log in or register to reply here.
Back
Top