Black and white from Colour
Converting colour images to black and white
Collected techniques for monochrome conversion
The images below represent some different ways of converting a colour image into black and white.
Some of the differences in the monochrome images are quite subtle and will be more obvious with different source images,
This article is one of the older ones (2004) on our site, and whilst we’ve added updates, we felt that the basic techniques described were still valid and worth discussing.
Apr. 2015: – even though I use Nik Silver Efex Pro 2 for and MacPhun Tonality Pro many of my conversions, I do still like to try some of these simpler conversion techniques. I still teach them in classes, since I believe they help understand more about what can make a B&W image special.
Sept.2009: A particularly useful B/W conversion plugin is currently one of Keith’s favourites – Nik Silver Efex Pro.
Some of the photoshop techniques may not be too obvious at first. One such method is using both a channel mixer layer and a hue/saturation layer to give fine control over the conversion, although you should be wary of pushing too far — unless you are specifically looking to show JPEG artefacts and sensor noise.
Jan. 2009: Photoshop/Elements now includes specific B/W conversion adjustments – These are quite flexible and give a whole range of conversion effects, so not really suited to the collection of example shots further down the page.
Variations on the Old John image
Conversion of digital colour to black and white is quite easy and these examples should help with your experiments.
Digital black and white does not need a special camera… and even if your camera supports black and white, you invariably get better results shooting in colour and doing the conversion yourself.
The images are all about 120k in size. They are JPEG files and are produced from the colour picture using Adobe Photoshop.
Do try some of these techniques on your own images to see the variations.
Original colour image of Old John, Bradgate Park, Leicester. Taken with Olympus E-20 (with wide angle adapter). Image was adjusted for white balance/saturation and resampled down to 640×480. It has been sharpened quite a bit for web use and uses sRGB for its colour space. The original raw file was converted to Adobe98 (a wider gamut space than sRGB) at 48 bit colour (’16 bit’) before adjustment.
Channel mixer - Red
Using just the red channel (100% red in the channel mixer).
Channel Mixer - Green
This is using the channel mixer at 100% green — or just the green channel
Channel Mixer - Blue
Just the blue channel (100% blue in the channel mixer)
Photoshop - Desaturate
This is simply desaturating the image in Photoshop
Plugin - 'Yellow Filter'
Plugin - 'Red Filter'
Channel Mixer - 66/24/10
This is using the channel mixer in Photoshop to do the conversion
- 66% red channel
- 24% green
- 10% blue
Do be careful if pushing any particular channel too far – this image starts to show noise in the sky from the red channel (which has little red light from a blue sky)
Plugin - FP4 'look'
Plugin - Tri-X 'look'
PS - Convert to grey
Just use the Photoshop Image>Mode>Grayscale function
Quick method that is not the same as the desaturate function. Similar to a channel mix, with preset values
PS - Channel Mix and H/S layer
A more complex Photoshop technique that offers considerable scope for experimentation.
First add a hue/saturation adjustment layer to the image
Next add a channel mixer layer (remembering to select the ‘monochrome’ option)
You can adjust the channel mixer to get differing effects and then tweak the H/S layer settings to ‘fine tune’ adjustments.
Note how using an agressive expansion of the saturation with the Hue/Saturation layer and subsequent channel mixing has
emphasised some of the noise (mostly blue channel) from the original digital image. Pushing effects like this can give an Infra Red like look to the images.
Channel mixer layer was at R+60 G+30 B+10 (monochrome), Hue/sat layer was Hue+40 Sat +60
Convert to Lab, remove 'ab' channels
This image was first converted to Lab mode from RGB in Photoshop.
The a and b channels were then deleted leaving just the Luminance Channel.
This was then converted back to greyscale (although this JPEG is RGB with an sRGB profile for web display consistency).
Note the extra detail in the dry stone wall compared with simply desaturating the image or converting to greyscale.
Using 2 H/S layers in Photoshop
Two hue/saturation layers are added one of which does the black and white conversion, whilst the other alters the picture that is being converted — easier to show than explain :-)
For a description of doing this in Photoshop Elements see the convert colour pictures to black and white tutorial on this site (works the same in Photoshop see below)
Using two Hue/Saturation layers
Create a H/S Adjustment layer with no adjustment.
Change the layer blend mode to Colour
Create another H/S Adjustment layer with Sat at -100.
Return to first H/S layer and tune the black and white effect using the Saturation and Hue sliders. Do not use lighten as it will change any end point already set — (method first suggested by Clay White – Thanks)
The image above used Hue -42 and sat +70
This is a very powerful and subtle technique.
The variable H/S layer can have individual colours adjusted to alter the intensity of different colours.
BW PS Action
Using a free Photoshop Action from Inkjetgoodies – unfortunately their site has long since vanished. However, the basic steps are here for one-off use or making into your own action.
The process is a combination of several techniques and does the following:
- Create New Hue/Saturation Adjustment Layer with the
- Convert LAB Color. Choose the Flatten Option if given.
- Open the Channels Window and delete Channels ‘A’ and ‘B’
- Convert to Grayscale.
DR Greyscale techniques
This very powerful technique was suggested by David Riecks and although it may seem a bit involved to with the various steps, it is quite straightforward and well worth a try. The image above was created using the five steps below. At the bottom of the page are several variations with differing opacities and switching the Alpha 2 and 3 layers in the final blend.
I quote David’s explanation… (see also his info at http://www.riecks.com/forphotogs/rgb_grayscale.html
The “DR Custom grayscale” technique:
Preamble. The range of values I give below work well with images that start as full scale (RGB 3 to 5 for blacks, and 247 to 252 for whites) Colormatch
RGB images. If you are starting with images that are in Adobe RGB, or full out 0 to 255 values you may need to “extrapolate” and experiment.
1. Change from RGB colorspace to LAB color space.
2. Choose the lightness layer (from the Channels palette), from the IMAGE menu choose CALCULATIONS. A dialog box will appear with a bewildering
number of options. There are essentially four main areas, Source 1, Source 2, Blending, and Result:
I’m going to tell you what to change, the default will be for Source 1 and Source 2 to be the name of the current file. Leave these as they are, also leave the “Layer” selection at “Background.” Provided that you have chosen the “Lightness” channel, “Lightness” should be the default selection for “Channel”.
Change the “Blending” mode to “Multiply”
Set Opacity to a value between 40 and 60 percent. This will affect the shadow areas disproportionally more than the highlight values. Aim for solid blacks in the darkest parts of your image, with the detailed shadow areas just barely visible.
The “Results” area will be set to “New Channel” as default, leave this as is.
Choose OK, and the result of the calculation will appear as “Alpha 1” in the Channels Palette.
3. Choose the Lightness channel a second time from the Channels Palette.
Choose Calculations from the Image Menu a second time.
We will repeat the operation outlined above EXCEPT with a change in the “Opacity” field (in the blending area) to 5 percent. The result will be a slighly denser lightness channel, but will appear washed out.
Click OK, and the result of the calculation will appear as “Alpha 2” in the Channels Palette.
4. Leave Alpha 2 selected in Channels Palette.
Choose Calculations from the Image Menu a third time.
This time change the Source 1 “channel” to “Alpha 1” and change the Blending mode to “Hardlight”
Set Opacity to between 40 and 50 percent. We are looking for an effect similar to a grade 3 paper as in days of old.
Click OK, and the result of the calculation will appear as “Alpha 3” in the Channels Palette.
For some this may be a result they wish to keep. I usually find it necessary to “temper” the result a bit, so I add one more step.
5. Leave Alpha 3 selected in the Channels Palette.
Choose Calculations from the Image Menu a fourth time.
Change Source 1 to Alpha 2 (the weak washed out image)
Change Blending mode to “Normal”
Set Opacity to “taste” (typically between 40 and 60 percent.
If you are happy with the result, click OK, and the result will appear as “Alpha 4” in the Channel Pallette.
Sometimes you may be getting too dark of a shadow area. If this is the case try reversing the sources, by setting Source 1 to “Alpha 3” and Source 2 to “Alpha2” and play with the opacity setting.
When you are satisfied, convert the result to Grayscale (discard color), and continue on from there.
If you feel that the blue sky should be a little deeper, or what was an orange colored object is too close in value to a light blue object, then try converting to RGB (the “Alpha 4” channel will remain selected).
Inspect the R, G, and B channels. If you are wanting to make the sky darker, you’ll find that the R (red) channel will have a much darker result. Lets assume that’s what we want for this example.
If so, set Source 1 to the “R”(red) channel
The “Normal” blending mode set at 50 percent opacity may look OK, but you may want to experiment and see what happens with it set to “Darken” or “Overlay” as well.
Sometimes you may wish to lighten a particular color, and that’s possible as well. Usually the “lighten” or “screen” mode will be needed in these cases.
Set the Source 1 to the color you wish to lighten and set the blending mode to “lighten” or “screen” and adjust to taste.
As above, when you are satisfied, convert the result to Grayscale (discard color) and save in your choice of file format.
In some very drastic situations (usually when handling imaging chores for a client that supplies their own scans or prints) I’ve had to resort to doing a separate calculation for one color or tonal value. I then duplicate the image, convert the best channels to grayscale, and then drag the “correcting” layer on top of the other (holding the shift key down to align). I apply a “layer mask” typically the “Hide All” variety, and “paint on the mask with a white brush to change the tonal value of that specific area. It’s easier than burning and dodging, as the results are completely reversable, even after saving the image (provided you saved it as a PSD file with alpha channels intact.
IR film look
Infra-red film effect
Two hue/saturation layers are added one of which does the black and white conversion, whilst the other alters the picture that is being converted — easier to show than explain
This method is a minor variation on the 2 Hue/Saturation method that I found by accident (I got the blending modes the wrong way round)
Using two Hue/Saturation layers — Other method
Create a H/S Adjustment layer with no adjustment.
Create another H/S Adjustment layer with Sat at -100 Change the layer blend mode to Colour
Return to first H/S layer and tune the black and white effect using the Saturation and Hue sliders.
The image above used Hue +27 and Sat +73
Curve and H/S layer
Curves and H/S layer
A variation on using a H/S layer to do the conversion. First suggested to me by Brian Price.
- Hue/Saturation layer, set saturation to zero
- Curves layer, adjust to approximate density/contrast
- Go back into Hue/Saturation layer. If you want to make, say, the sky darker set the drop-down menu to ‘Blues’ and reduce theLightness by the required amount. If you want to make foliage lighter, choose ‘Greens’ and increase the lightness, etc.
- When you are in one of the channels the cursor becomes an eyedropper and you can fine-tune the colour range you are adjusting.
- Finally, tweak on the Curves layer if necessary.
Gives good direct control over the tones. In the example the lightness of the ‘Blues’ was reduced to darken the sky, and the ‘Greens’ increased to lighten the foreground. The curve layer is used to set the overall contrast of the image, and would probably be included with most conversion techniques.
Blend black layer in Colour mode
Blend black layer in ‘color’ mode
Create a new layer and fill it with black.
Change the blending mode to ‘Color’ and the image goes to black and white
Flatten when done to get your greyscale image.
As with many of the methods here, the variations may be quite subtle, so try it with images of your own to spot the differences.
Alternative techniques - suggestions always welcome!
This one is from Michael Friedman in the US — It’s a much more complex and flexible version of the channel mixer version
- Create a copy of the original image, select it and then delete, leaving only the window. Alternately, you can simply create a new, blank file, but copying and deleting the original ensures that both files will be exactly the same size and your registration will be correct.
- Go back to the original image and view the three channels in the Channels pallet. It is usually the case that the information in one of the three is noticeably better than the others. (Note: which channel contains the best information depends on the image).
- Select and copy the best-looking channel into the duplicate (and now blank) image window. It becomes a layer.
- Go back to the original image and select a different channel, then copy and paste it into the duplicate image. It will become a 2nd layer on top of the first.
- Optionally, you can go back and create a layer from the third channel in the same way. I usually find, though, that two are all I need for this method. You will now have an image that has a BW layer for each channel you selected and copied.
- At this point you have many different options. You can blend the layers by changing the opacity settings of the upper layers.
- You can try the different blending methods (all will look different)
- If each layer contains areas that are superior, you can create a layer mask out of the top layer and paint in the better values from the layer below. (This is a very powerful method of getting great tonal values in all areas of the photo).
- This is where I will usually merge layers or flatten the image.
- Once you’re satisfied with the results of the above procedures, you can use curves and levels adjustment layers to fine-tune the results or generate more contrast.
- If you need to, you can now do one of the Photoshop procedures for additional shadow recovery.
- If you wish to colorize, or warm/cool the image, this is a good time to do it.
- Sharpen to taste.
One adjustment you might like to try after your conversion is to run Unsharp Mask at around 5-20% intensity, with a radius of 50-150 (threshold = 0) This does not sharpen in the way you might ordinarily do with USM, it provides a subtle contrast enhancement and for some images it is very effective, particularly where you have a big dynamic range in the picture — for example between the shadowed wall and cloud. I often run it after using the Lab conversion technique. Give it a try in association with one of the methods on this page. It’s also known by the acronym HIRALOAM (HI RAdius LOw AMount – Apply an unsharp mask filter with high radius and low amount settings).
The view above is of “Old John” a folly built on a high hill overlooking Leicester (a few miles to the other side of the hill). It is in Bradgate Park, a popular place to go for a walk for the people of the region. There is more information here[WP]. Keith Cooper first visited the Bradgate area as an undergraduate, doing his Geology degree at the University of Leicester. It is where the oldest fossils in the UK (Charnia) were first discovered in 1957 and provides a glimpse of some of the earliest complex lifeforms, over 500 million years ago.
I’m always happy to hear of any alternative methods available, so please let me know.
We also have an article on why it may be better to convert your 8 bit RGB images to 16 bit before converting to black and white.
The second set shows similar B/W conversion effects on completely different source material
The colour original was taken in a brightly lit fast food outlet at about 2am.
The mix of fluorescent and neon lighting makes it somewhat difficult to get a good white balance (an 18% grey card not being at hand).
Taken with an Olympus E-20.
Channel mixer - Red
Just the Red channel of the original RGB image
Channel mixer - Green
Just the green channel of the RGB image
Channel mixer - Blue
Just the blue channel of the RGB image
This is a basic conversion Using the Photoshop Desaturate function
Photoshop - Convert to greyscale
Using Photoshop (or Elements) to convert to grayscale
Convert to Lab and use 'L' channel
Convert to lab colour space and delete the ‘a’ and ‘b’ channels.
Then change the mode to greyscale.
Channel mixer (bright)
In this conversion, the values of the channel mixer (54R 44G 12B) add up to more than 100% giving a brighter image.
There is no reason that your settings must add up to 100% it’s just a useful starting point.
DR custom - V2
This conversion uses the method suggested by David Riecks – see the Old John version for a full explanation.
Never miss a new article or review - Sign up for our occasional (ad-free) Newsletter
Enjoyed this article?
All the latest articles/reviews and photo news items appear on Keith's Photo blog
We've a whole section of the site devoted to Digital Black and White photography and printing. It covers all of Keith's specialist articles and reviews.
For All about using tilt and shift - articles/reviews about tilt/shift lenses
Articles below by Keith (Google's picks for matching this page)