Thursday, October 21, 2010

Grayscale conversion of color images by Paul Baldassini

Please contact me via email:

Conversion to grayscale

In keeping with recent Studio Notes, I am republishing this post with some minor edits and as a stand-alone Photoshop Studio Note.  This Studio Note and all previous ones will soon be available as downloadable .pdfs on my next website update which include a new page RESOURCES.

After over 30 years as professional graphic artist, I have grown quite used to using image editing software -- mainly Adobe Photoshop -- to develop my compositions.  I am very comfortable with this way of working even though I sketch and draw often, usually at cafe’s. My compositions are initially developed using the viewscreen on the back of my digital camera (I use a Canon G10, an incredible little machine for under $400) and then the images are assessed for usability in Photoshop.  I use both Auto and Program mode, Auto for a moving subject like a farmer on a tractor for example and Program mode for stationery objects, like hayrolls or a tractor idle in a field. This requires a VERY steady hand but produces fantastic source material for paintings.  I will elaborate on this technique in a future Blog posting.

Nothing beats a grayscale image for checking values so I always create a grayscale version of my completed composition as visual reference to check values while painting both the underpainting and overpainting. There are many ways to do things in Photoshop, but the easy way is often not the best way.  Going to MODE >Grayscale and converting from RGB to Grayscale almost always produces a muddy, flat and dead looking image.  And this may be fine for most people, but I use my grayscale conversions for a lot more than just checking values.

First, I can use the data stored in each of the 3 channels (assuming RGB color mode) to significantly improve the tonality and contrast of my original image by creating new alpha channels that have been blended from the Red, Green and Blue channels. If you click on each of the 3 channels separately you will notice that each channel is, firstly and most importantly, a grayscale image already.  From the top (Red channel) down each channel gets progressively DARKER. The Blue channel contains all of the bad stuff -- in a portrait for example, the Blue channel contains the pimples, wrinkles, 2-day beard growth, etc. along with all of the noise (grain) present from the original camera settings.  The GREEN channel, however, contains all of the good stuff this channel BY ITSELF, again in the case of a portrait, will produce a MUCH better grayscale image than simply converting the whole image to Grayscale Mode.

There are several ways to do this.  One is to simply click and drag the the Green channel into the little dog-ear icon at bottom left of the Channel palette right next to the little trash can.  This will DUPLICATE the Green channel into a NEW alpha channel named “Green copy”.  You can easily duplicate Layers the same way.  Then go Image >Mode and select >Grayscale.  You will get a dialog “Discard other channels?”  Click OK or hit the ENTER key and you now have a single channel Grayscale image. Go to >File >Save As and name and save your new image.

A better, and more advanced, way if to use one of the two most powerful and overlooked functions in Photoshop -- Calculations and Apply Image.  They live in the >Image drop down as shown in the screen grab here.  Calculations is used for generating new grayscale Alpha channels only and Apply Image is used for effecting real-time changes to the original color image only.  This post will only deal with the Calculations functions.

For this demonstration I will use a portrait as they best illustrate the technique of grayscale channel blending.  The image is of my daughter. The original image is an RGB file as seen here and next to that is a standard grayscale conversion.  Very dull and flat.  After examining the channels it can be clearly seen that although the Red channel image is too light in the facial tones, the Blue sweater has much more detail than in the Green channel.  The facial tones look much better in the Green channel but the sweater is very dark.  I could use the Green only -- I think its better than the converted Grayscale image -- but it can easily be improved by blending, or combining, the best data of both the Red and Green channels.

Left: original RG image; right: standard grayscale conversion

Below is screen grab of the pane and here’s how it works: go to >Image >Calculations and you’ll get a new pane with options to set and change things.  The first one is Source 1 which is the name of your open file. Unless you have another image open the same physical size, this will be you only option here.  Source 1 has two sub-menus: Layer and Channel.  Unless your file has more than one layer, Background will be your only Layer option here.  If you have more than one layer, be sure to set to the layer you want to blend channels from. The second sub-menu is Channel and here is where you choose the first channel you want use to blend.  Choose Red.  Everything repeats again for the next set of options, Source 2 and your only choice here is your open file so leave it alone.  Set Layer to Background or leave it alone also, and then for Channel choose Green.

Calculations drop-down menu and work pane

Next is where you can really make some radical changes to your grayscale image depending on which Blending mode you choose.  Blending modes are the real core of Photoshop’s ability to manipulate images and If you have never worked with Blending Modes before they may seem daunting.  But please try some out here to see what happens and then you can try them out later on your layered files in color, in real time.

For this demonstration I want to combine or ADD some of the Green channel to the Red channel, so from the drop-down Blending menu I cho-se Add, >Opacity 100%, >Offset -20, >Scale 2.

Finally at the bottom, you have another drop-down menu where you can choose what to do with the new channel you just created.  The default is New Channel and most often this is what you should do. The default option adds a new alpha channel to your file and you can do with it what you will.  So just click OK, or hit the Enter Key and you will be now see a NEW 4th channel in the Channel palette named Alpha 1.  Double-click on the type only and name the new channel.  I chose “Red.Green.add.-20%” so I could remember what I did.

To make a NEW standalone file click on the new channel highlighting (a blue tint), then go Image >Mode and select >Grayscale.  You will get a dialog “Discard other channels?”  Click OK or hit the ENTER key and you now have a single channel Grayscale image. Go to >File >Save As and name and save your new image.

The other option is to select >New Document. Click OK, or hit the Enter Key and you will now have a new file automatically in Grayscale mode.  Just save and name the new file.  If you compare the standard Grayscale conversion to the Calculations generated function version you can see the difference, even though that differernce may not be that apparent on your monitor and/or at this resolution.

Left: Red channel; center: Green channel; right: blended Red & Green channel

Below is the same treatment applied to a source image for one next paintings. At top left is the standard Grayscale conversion. The others were generated the same way as in the preceding demonstration.  The main way that I utilize my Grayscale images is to produce a large paper proof same size as my prepared panel.  I use this paper to produce an outline drawing of my composition onto my panel by sandwiching a piece of Sally transfer paper between the proof and the panel. The standard Grayscale conversion would have produced too dark an image for me to see the traceable detail, but the Red channel combined with the Green channel kept all important detail light enough to see and easily trace onto my panel.

Top left: Standard Grayscale conversion; top right Red channel; bottom left:  Green channel; bottom right: blended Red & Green channel

Sounds like a lot of work but its actually quite fast, the variations are unlimited and I find it well worth the effort.  It also opens you up to new creative levels of understanding in Photoshop which can only improve your design and composition choices. I urge you to try many of the Blending Modes, especially on layered images, and change the opacity on some to produce a very different range of possibilities from the original image.

That’s it until my next post.  Thanks for visiting.


No comments:

Post a Comment