Converting photos to black&white

Today I was finally giving b&w photography a serious try, or rather: converting normal photos to b&w using my trusty Photoshop elements (and iPhoto). Why should you want to do more than press the "b&w" button or use the "desaturate" function? Because when you lose all the colors you need some level of control over the process to e.g. turn unimportant stuff darker or highlight the important parts.

An excellent resource on how to do that is available at cambridgeincolour.com. And although most of the techniques require the full Adobe Photoshop, one works beautifully with Photoshop Elements 4.0. And it gives you just enough control over the process to make experimentation fun .

How do you do it? You create two hue/saturation adjustment layers. Here's how to do it (shamelessly copied from camebridgeincolour.com. See this site, it's great!)

1. Create a layer with Layers > New Adjustment Layer > Hue/Saturation...
2. Set the mode of this layer from "Normal" to "Color" - I first did not read this completely, and as a result nothing happened at all :-).
3. Set the saturation of this layer to "-100". The image turns black & white.
4. Create another layer. Make sure this layer is between the first one and the background layer containing the actual image.
5. Now you can play around with the hue slider and watch the effects this has on your image. You can adjust the effect by increasing the saturation. If you're very daring or want to achieve a special effect, you can also edit specific color channels this way.

I was, frankly, amazed about what this technique can do. Below you see three images of the "Frauenkirche" in Dresden. Left: the original color image. Observe how the different stones have their own texture and their own color. The next image shows this image turned b/w using iPhoto's b&w effect. Much of the texture is lost. The last image shows the same image turned to b/w with some adjustment using the technique above. Quite a difference, right?

The original image

Automatic conversion; much of the detail is lost

Conversion with the "layer method". Much more structure is retained.