This was originally just going to be a small post in my sketchbook to answer a few questions people had about how I create my images. As I typed more and more I realized that it was going to be too big to post in the middle of my sketchbook so I guess I will put it here.
This is just some information on how I paint.
Last edited by m.c.miller; December 30th, 2008 at 09:26 PM.
First, I want to say that I never intended for this to become so long. I just wanted to put something together to provide those interested with some information on how I make my images. The more I typed the more I realized how much more I needed to explain and…well this could have gone on forever. Also I do not at all claim to have invented this process…this is a way that I have seen many others work to a greater or lesser degree and is something that I have learned from them just as all artists learn.
I would have never dreamed of painting this way when I first started painting digitally. I used to paint on a single layer with the simplest brush, never touching adjustments or layer blending modes. At a certain point though I realized that as long as the computer wasn’t perfectly simulating paint there was no reason for me hold myself to those rules for which much of the strength of the medium is missing digitally.
I cannot stand choosing colors in photoshop, slowing down to drag sliders or select swatches is a pain. Also, this way of working with color is nothing like real paint mixing in which colors harmonize and pull together creating atmosphere and a rich surface. Applying color with photoshop brushes can be like painting with sidewalk chalk. There are a lot of great digital artists who make it work but it takes a lot of time and effort, and is not the way I want to paint. I want to paint like an impressionist who attacks the canvas with color. With real paint the strokes blend and sit nicely next to each other even if their colors differ quite a bit. In photoshop colors tend to sit on each other dry and flatly. The technique that I use addresses this and attempts to create a nice color harmony with which to paint as a continuous part of the painting process.
I know that other software addresses some of these issues, but the main issue I have with all software with a more robust brush engine is speed. I need to be able to use brushes nearly half of the canvas size or more with lightning fast speed. Also, no other software can match the photoshop interface and all other non painting functions in my opinion.
I realize that this way of working is not for everyone and is very different from a traditional painting process. A lot is done that ignores or postpones many fundamental rules of painting and this is not for anyone who wants to be in total control from the beginning. This process is for those who want to explore, sling color around, enjoy the journey and be surprised at the destination.
That said, you will learn a lot about color and value by painting this way and all of this information can be used to inform any more traditional painting that you may wish to do.
I always use a brush that has a bit of texture and randomness to it. This keeps me from trying to keep things too perfect and clean and also prompts imagination and texture in much the same way that you can see shapes in the clouds. You can see a description of the brush that I most often use earlier in my sketchbook. All brushes that I use have opacity control set to pen pressure.
I always start with a somewhat mid value gray base that I paint in with a huge brush. In the beginning all I am looking for are basic value relationships, so it is easier to work from a mid value.
Next I paint using lighter and darker values to pull out and separate objects. I am not concerned with specific values, I am just putting down relative information: 'this is darker than that but lighter than this' etc...
It is very important in the beginning to create a pleasing composition of values on which to build your image. You want a nice balance of light and dark areas as well as areas of activity balanced by areas of rest. This is a rule that I often find myself breaking, but try to remain aware of. Note that my composition for this image is pretty terrible. It gets somewhat better later on, but generally you do not want such a static and centered design. Make sure your shadow areas stay dark and your light areas stay light. Try to create a pleasing composition out of these values even before you start defining objects.
Once I have the image mapped out, I add a levels adjustment layer and adjust the levels to a more appropriate range. I am careful not to put too much contrast in though, as that will happen in later steps and I want to leave some room to work with.
Now I add a new layer above the levels layer and paint using values already in the image. I am constantly using Alt to select colors from the image. I only select a new color manually if I absolutely have to. This is actually my method for the entire painting process. I paint for a while on normal layers and then adjust the colors and values with adjustment layers and then paint on top with normal layers again.
I paint almost exclusively in terms of shape. This may be hard to do at first if you are used to starting everything with a line drawing. You have to be able to think in terms of large shapes and masses as opposed to outlines. So I am creating shapes with as big a brush as possible. Every object starts out as a blob and is refined by ‘carving’ back into it with surrounding values. So in a basic example of a black shape on a white background I am constantly selecting black and white with Alt (eyedropper) and working back and forth to create the shape. I do not use the eraser or separate objects on layers. Each layer is treated as a flat painting surface.
Do not create edges of shapes with small brushes and then fill them in. Create everything from the largest mass possible and then paint back into it to refine. Always work back and forth. Paint edges into each other and not up to each other.
Always choose a single value for each object in the beginning. In order to establish a value structure you need to understand that objects are usually seen as a single mass more than a combination of parts. So a dark object has lighter and darker areas that define it but the object as a whole is seen as a single dark object. Don’t break apart your objects too much by overstating the shadow and highlight values. In a complex scene objects are mostly seen as flat in relation to each other.
3. Initial color
I usually move on to color when I feel like it will help clarify the image. While it is true that a good image should work in black and white, it is also true that color can add clarity by differentiating areas of similar value with hue and saturation. And since I do not design the entire image in black and white, it helps to add color as soon as I have an idea of where it should go.
I add the color on a new layer set to the ‘color’ blend mode. I always use a large round soft brush when painting on a layer set to anything other than normal blend mode. This helps maintain clarity and keeps things simple. I never add texture on these layers, I only use them to add color or lighting.
I put the color on very simply, first by covering the entire image with a single color. This color is the basic color of the environment. This image is in the forest/jungle so I put down a yellow green color. Next I add some color to the space ship and water and put some light orange in the trees to suggest light and atmosphere. That is enough for now.
Notice that I am using very dull, pale colors at this point and that I am not at all concerned with staying inside of specific shapes. This stage is just like the early black and white stages in that I am only using enough color to differentiate areas from each other. As long as we have a small bit of this information stated, it will carry on through the later stages when we increase contrast and saturation. Also notice that I am fully aware that this stage looks terrible. This is something that you will see throughout the process. Sometimes the image will look worse before it looks better. This is because it is easier to add information and then orchestrate it later on than it is to try to do everything at once, especially when you are exploring different ideas.
Now I pull the image back together with some adjustment layers. It is important to understand that these adjustment stages that I describe are totally dependent upon whatever image is being created. There are no rules here.
The important thing is to understand what each of these adjustments does and why.
Photo filter: I use this to unify the colors in an image. All colors in a scene are affected by the environment that they exist in. This includes the sun and all objects that the sun bounces off of including the sky. This has several consequences, but the major thing to keep in mind is that all colors are under the influence of the same environmental light. This is why it is said with regards to perception that color is relative. You may see green red and blue in a scene lit by a strong orange light, but the orange has a dominating presence over the entire scene. So you are just seeing things that look greener redder and bluer than each other. You can have a lot of colors appear in what is essentially a small range of hue and saturation.
So I can put down colors in their literal form on the color layer and then come back and unify them with a photo filter to describe the environment that they exist in. I would use a warm color for a sunlit scene and I might use a cool color for an underwater or night scene. Make sure that the little check box next to preserve luminosity is turned on.
This would be a pain to do by hand dragging sliders. You could start with a warm base and add local color to it for each color, but this is slow and painful. In real paint this is much easier to achieve since you can just mix each color with a base of yellow or paint on top of an under painting which would harmonize the color directly on the canvas.
You can also use a photo filter to do other things later on in the process like pull back saturation if things get a little out of control. Using a cool photo filter on top of an oversaturated warm image can sometimes produce better results that simply using hue/saturation.
Color balance: this is my favorite adjustment tool. The important thing here is that you can adjust the colors of the highlights midtones and shadows. On a sunny day highlights are usually warm and the shadows are usually cool…both because of reflected light from the blue sky and because they are simply less warm than the sunlit areas and thus appear cool. The opposite is true for scenes lit by cool light -the shadows appear warmer.
For a sunny day I usually add yellow and a bit of red to the highlights. For the midtones I usually add blue or red…whatever looks best. For the shadows I sometimes find that defying conventional wisdom works best and I might add some red or yellow (but I use blue just as often).
You may be thinking here that there are a lot of flaws with using color balance this way since you can have a dark object that is in the direct sunlight and a light object that is in shadow. While this is true, it is also generally true that light areas of the image are hit by the sun and dark areas are in shadow. In my experience it works out just fine and these exceptions are rarely an issue in relation to the overall effect.
You do have to be careful with the highlights and shadows though. Adding too much yellow to the highlights can blow out your highlight detail and too much red or blue in the shadows can over darken the image. Here, as with the photo filter you want the preserve luminosity box checked for more attractive results.
As with photo filter you can do other things with color balance like adding blue to the highlights at the end of a painting, which sometimes looks good on overly warm images. The best thing to do is just experiment…sometimes doing the opposite of what you think works best.
Levels: I do not use levels that much by itself after the black and white stages. I usually use it in combination with a photo filter or color balance that has been set to multiply. For this image I have a color balance adjustment layer set to multiply which produced way too much dark, so I used a levels layer below it with the shadow output level set to 12 (whatever looks good). This brings the shadow detail back a bit by compensating for the multiply.
Another neat thing you can do is set a levels adjustment layer to any blend mode and then adjust the levels to change the strength of the effect at different values.
So the idea with these adjustment layers is to unify color, describe environment and refine values. I paint for a while and then use these layers to pull the image together.
Once I create and adjust an adjustment layer I often scroll through the layer blend modes to see if anything jumps out at me. The power of these layers is really only found when you start experimenting and combining them with different blend modes and opacities. When you work this way you are just looking for something to work from…you are not looking for a final fix or polished look.
Don’t forget about layer masks when using adjustment layers. If an effect looks good in some areas but not in others…go ahead and paint them out with black in the layer mask with a big soft brush.
It is important to realize that these layers can be very effective but are not magic. Color balance for example can be very effective, but sometimes you have multiple light sources with different temperatures that complicate things. Even these situations take care of themselves to a degree though since there is still a dominant light in most scenes. The point is that these tools are great but not perfect and you have to know what you are doing and be able to plan ahead.
You do not want these adjustments to remain untouched after you make them. The idea is to slowly develop your image in terms of shape, composition, color and value throughout the painting process. So in theory, as the image progresses and you get closer and closer to the colors you want, your adjustments get smaller and smaller. As you can see that does not happen in my example…but that allows you to see how much painting is needed to make such a large adjustment work.
It is basically true that the larger and more extreme the adjustment that you make, the more painting you have to do to make it work. Smaller adjustments require less painting to be effective. At the very end of the image there is a final adjustment…but hopefully it is only to polish the image and not change it too much.
I treat these adjustments like color studies in a way. I am doing color studies throughout the painting process, painting for a while with each one until I understand (or destroy) the image enough to do another one.
Last edited by m.c.miller; October 19th, 2008 at 11:15 PM.
I use overlay and linear light layers to add and adjust lighting as I paint. As I said before this is always done with a large soft round brush. Overlay layers will light the paint as it exists in the image, while a linear light layer will add a more visible and opaque light (useful for depicting the glare of a very bright light). Use very light and pale colors when adding this light otherwise the effect will look ugly.
I usually go over the painting and use a warm color to light areas hit by sunlight. The goal is not to put highlights everywhere, but to add warmth and light. One exception to never using the eraser is with these layers. Sometimes I use a big soft eraser to refine the shape of the lighting that I have applied. This helps to create less predictably circular effects.
As with any adjustment, you want to be careful here not to overdo it. Early in the process you can be very harsh and go crazy since you are going to do a lot of painting back on top, but as the image progresses you need to be more and more careful.
If I can, I always use colors already in the image to light with. A light pale yellow or orange for example, but the color has to be very light…nearly white to work well.
In between adjustments I paint on normal layers using colors already in the image, with occasional exceptions where a bit of local color must be added. When I need to introduce a new color I always do it by choosing the pure color that I need and ‘mixing’ it on the canvas. I lightly brush it in where needed and only use just enough to be effective. You could also select the color of the area you need the color to go and add just enough of the color you want using sliders and then paint it in.
When painting with colors already in the image I try to think in terms of light. If I want to add a highlight to an object or to lighten something, I will select a color in the image that represents the color of the light source and ‘cast’ it onto the object by brushing it on using an appropriate level of pressure. The same goes for shadows, where I select a color in the image that represents the color of shadow areas and is darker than the area I want to darken. So if you want to lighten a part of a leaf on a sunny day, select a nice yellow from your image, if you want to darken it, select one of your cooler darker colors. So you are constantly bouncing around colors already in your image.
Working this way you retain harmony and color atmosphere as long as possible. Over time your colors will tend to lose saturation and harmony though so take this into consideration when painting and adjusting.
When I make adjustments I sometimes push them a little farther than looks good because I know that by painting on top of this I will settle it back down.
Those are the basic stages that I go through for each image. I probably go through about four or so rotations of adjustment and painting for each image. Usually each image reaches a point where everything clicks and I know what I am doing and have my color scheme set. In this image, however, you can see that I changed my mind rather late in the process and made a major color shift. There is nothing wrong with this but if you do it you have to do a lot of repainting to make it work, as you can see. You want your adjustments to look natural and not appear as if you just painted and then changed everything at the end. You can do this as much as you want during the painting process but not at the end. The end must display more of a sense of painting than adjustment.
Most of my paintings do have a glow to them that makes obvious the use of adjustments, but the point is that the colors should appear somewhat natural. You do not want your image to look strange like a photo negative or something from the cover of an early 90’s trapper keeper.
Feel free to flatten earlier layers once you are sure that you are not going back. You do have to be a bit careful with the adjustment layers though. In order to flatten an adjustment layer correctly you have to merge it all the way down to the background layer…you cannot merge them in the middle of other layers.
Last edited by m.c.miller; October 19th, 2008 at 11:21 PM.
When I paint, I paint fast. Very fast. That does not mean that these images are completed in a short amount of time though. You need to understand that when I am painting on a normal layer my hand is constantly flying around at the edge of control and I have no idea what I am doing. There is no way I can ‘see’ how the finished painting will look when I start.
It is a mistake to look at my finished paintings and think that I just directly paint that as if I could see it ahead of time. I said before that I like to use a brush that suggests shapes much like shapes can be seen in the clouds. This is what I do for the entire painting. I am always looking for something to happen that is better than what I already have. For every stage of this image that you see there has been a lot of trial and error painting and repainting. I look at an area that looks bad or lacking and I attack it with the brush pushing paint around trying different things. When I get something that looks better, I refine it and move on. What looks worse now? I work with that until it is better. Sometimes I work around the image and find that now what used to be the best part of the image is the worst and must be changed. Do not hesitate to paint right over the best part of your painting if you know it is not going to work in the long run.
Like everyone I have the tendency to get scared of ruining my image. All you have to do is make a new layer and paint on that. In a few minutes you can turn the layer off and see if it looks better or worse. There is a second exception to never using the eraser. If I decide that part of a new layer is better but another part is worse I will use the eraser to remove those areas.
I also use the history brush like a bookmark. I know that this is what snapshots are for but I prefer my method. If I reach a stage that I like that I want to come back to after trying something, I will just put the history brush marker next to that step in the history to preserve it. That way the step will never go away even if you surpass your maximum level of history states. You can also use this to try things before a step in the history without destroying it, or just to mark a step that was good.
My philosophy on painting is to try to paint each object in just a matter of seconds. The first time it will look terrible. So you paint on top of it in a matter of seconds and it looks a bit better. Each time you do this you refine the image because you know more about what you are creating. After doing this several times you end up with something that you never would have come up with had you just sat there slowly refining your initial concept. You need a foundation on which to build in order to create a more complex and exciting shape, object, image. Once you begin to build this foundation everything you produce is more and more informed even though you are painting it just as fast and freely. In this way it is no problem to repaint something that is not working because you are not destroying something that you have a lot invested in, but are instead quickly forming a better result.
Once your big objects are working you can move on to the smaller parts of objects and treat them the same way. Now you are painting and repainting details in a matter of seconds as they get better and better.
The downside to this way of working is obviously that it does not produce very polished and detailed results right away. I believe that this is something that is improving with time, however. If I look at the images that I made early on with this method, they appear very simple in comparison to what I am producing now with the same technique and in the same amount of time. Obviously, I could just relax and refine my images at the end, but I feel like I learn more by challenging myself to achieve these results as an integrated part of the painting process.
I think that covers most of the information specific to the process that I use to paint. I had to stop myself from trying to cover everything about painting that I have ever learned…but if anyone has any questions about how I handle these more basic aspects feel free to ask.
Last edited by m.c.miller; October 19th, 2008 at 11:13 PM.