B&W Conversion - LAB Colour mode

[ciaran]

More tutorials can be found on my website: theWonderOfLight Dot Com
----

There are so many ways to convert an image to B&W, each offering different advantages, disadvantages, look and feel. Personally, I find converting using the Lightness channel in the LAB colour mode gives me the most pleasing results for my portraiture.

1. The first thing to do is convert the image to LAB colour mode (Image->Mode->LAB)



2. Like RGB, LAB colour mode has 3 channels. However, the colour information is split into channels A and B, and the Lightness/Luminosity detail is contained within the Lighntess channel. It is this channel which is used for the conversion



3. We need to remove the colour channels from the image. Some people physically delete both of them, but I find the easiest way is to simply select the lightness channel



Then convert it to Greyscale (Image->Mode->Greyscale). Note: This is not the same as just converting an RGB image to greyscale



4. The conversion is more or less done at this stage. But most people prefer to work with RGB images, so the next step in the conversion process is to convert the greyscale image back to RGB (Image->Mode->RGB)



5. The image is now a B&W image back in the RGB colour space, ready for further processing.



6. Personally I always find I need to boost contrast a little, so some curves/levels adjustment is usually required.



There are so many different ways we can convert an image to monochrome and generally I don't subscribe to the idea that one way is better than any other. For different types of shots, I may use a different conversion technique. What I've shown here is the results of a number of common techniques (I forgot to add desaturate). I don't detail how any of the conversions are done, but if you would like to know, please ask.




Sepia Toning

All of my B&W shots have some sepia in them, with varying degrees of strength. I prefer the effect where the toning is added to the shadow detail more so than the highlights. Below are some techniques for adding sepia to your shots.

Colour Balance

1. When you have monochrome conversion, add a Colour Balance adjustment layer.
2. Select the "Shadow" option in the Tone Balanc box at the bottom of the form. This will add the colour balance to only the shadow detail.
3. Enter in the following values for sepia "20, 10, -10" for the RGB values
4. You can then vary the effect by simply varying the layers opacity. I usually have it down and around 25% which is very subtle

Another option (and there's a trick here which can be used to apply an effect to shadows/highlights on any adjustment layer)

Sepia Photo Filter

1. Go to the channels tab
2. Control click on the RGB channel. This selects all the highlight tonal values in the design. It actually does a background conversion to greyscale and then selects all values from 128 through to 255
3. Invert your selection (Control-Shift-I). All shadow detail is now selected (0 through to 12
4. Add a photo filter adjustment layer with the selection still active
5. Choose Sepia
5. Sepia will be added propotionally to the shadows in comparison to the highlights.
6. Strength of the effect can be adjusted by varying the opacity of the adjustment layer.

Solid Colour Layer

1. Same technique as above to select only the shadow detail. Go to the channels tab and control click on the RGB channel (select highlights)
2. Invert your selection - Control-Shift-I (select shadow)
4. Add a "solid colour" adjustment layer. In this case you can pick any colour you want, but for sepia usually the RED channel has to be twice the BLUE
5. Change the mode of the new layer from "Normal" to "Colour"
6. Once again vary the opacity to determine the strength of the effect

Sometimes there isn't enough contrast between the shadow details and the highlight (i.e. most tonal ranges are in the highlight section or in the shadow range) in which case the sepia effect wont be as strong. One way of generating a "better" mask where there is clearer seperation between shadow and highlight is to use "Calculations"

1. Image -> Calculations
2. Lets assume for now you have one flat layer for the monochrome image. This will work on layered images, but it's more complex to explain. Ensure Source image #1 is your current image. Change the channel to Grey
2. For source #2 change the channel to Grey again (this is equivalent to Control clicking on the RGB channel in the other 2 methods)
3. Change blending mode to "Multiply" and Result to "New Selection"
4. Rather than inverting the selection by hand, click the "Invert" button beside each source option
5. Click OK

You now have a selection where most of the highlight detail is masked out and the shadow is clear - no hard edges and a nice transition in between. This selection can be used for loads of different reasons, including blending images where two exposures have been made (one for shadow/one for highlight). It can also be used for the sepia photo filter adjustment layer or solid colour adjustment layer

The output of each toning method is shown below

1. Image top left is the original.
2. Top right is converted via the colour balance layer, set to only affect shadow detail
3. Middle left is via the sepia photo filter, with a normal greyscale selection for the shadow detail
4. Middle right is via the sepia photo filter with calculations used to generate a more contrasty shadow detail mask
5. Bottom left is via the solid colour layer using normal greyscale selection for the shadow detail mask
6. Bottom right is via the solid colour layer with calculations used to generate a more contrasty shadow detail mask

All adjustments can be toned down by varying the opacity of the adjustment layer



Original image of buildings above was taken by MikeAlex (http://www.mikealex.ca)

[Rense]

I take very often an additional step after the L-channel to gray image conversion:
duplicate the layer, st blend mode to multiply and change the opacity to your own likenings.

Thanks for the writeup.

[JimAuburn]

This is a fantastic tutorial. I've bookmarked it and will look out for anything else you produce. Thanks!

[ciaran]

That's a nice addition to the conversion! I normally burn in the edges of a print using this method, but used sparingly I can see how it could also be applied to the rest of the image.

Originally Posted by Rense

I take very often an additional step after the L-channel to gray image conversion:
duplicate the layer, st blend mode to multiply and change the opacity to your own likenings.

Thanks for the writeup.

[Danno44]

Thanks for the great tutorial!