2014 foreword added 2018
Is it possible to write a general tutorial on painting pretty pictures or is this art-thing too complex and magical to ever be understood, a thing for human intuition only?
No human is alike, but we live in the same environment and train our brains together. I believe this creates certain shared structures which can be analysed and understood, at least on a sufficiently useful superficial level. I think anyone who practices a trade a lot will begin to formalise the relevant structures, many of them subconsciously. A musician becomes aware of the structures of a song. A doctor might spot health problems based on very subtle common hints. A movie director will know why a scene doesn't work, when the same scene just makes a regular viewer a little confused. To be a painter is to constantly train the brain to spot faults and inefficiencies in order to improve. My personal preferences and limited human cognitive and intellectual abilities creates a sort of "good-enough" frontier where I'm no longer able to notice or care about imperfections. Below this threshold lies an ever expanding grumpiness about the state of things. When I was young I often thought that there were rules and simple tricks which, if I only knew them, would unlock the awesome art powers which I saw and I lusted for. Rules to art? Blasphemy! But maybe not. Perhaps it's not hard simple rules, but a myriad of assorted rules which lay madly intertwined somewhere in the mists of our minds. Some rules play against each other or cancel each other out. Some become invalidated because of a stylistic approach. Some are shared, some are not. What I will elaborate on in this tutorial is just the very broad basic rules (or "tricks" if you like) which can be communicated. I see them as discoveries common and useful to us Space Marine vs Hell Beast painters. I was quite active on art forums between 1999 and 2005. It was during this period that I noticed the aforementioned structures in the form of reoccurring critique. Eventually I begun to compile my observations into a tutorial which I published on my site. This PDF version is a cleaned up version of the old HTML page and it should print much better. I hope you find it useful! I should make a 2 column PDF version of this page.
Getting started from nowhere
It has been claimed that you need to put ten thousand hours into a skill to "master" it, whatever that means. I generally notice visible improvement every few months if I'm active producing art on a daily basis. Or, perhaps what I actually notice is an increasing dissatisfaction with my older works. The better you get, the larger this... mountain of shame, dissatisfaction and grumpiness.
So, anyways. Just drawing a lot will propel you forward, but I can think of several things which will slow learning down. 1.) Fear of failure. Every time we figure out why we failed (or unexpectedly succeeded) we learn something. When doing studies, put no value into what ends up on the paper. It's what goes into your brain that matters. Throw your studies into the trash or the bottom of a drawer. 2.) Trying to be fancy. If you try to produce works too far beyond your ability, everything will be a time consuming mystery from which you'll learn nothing. Learning is best done halfway into your comfort zone, I believe. 3.) Nonchalant sloppy drawing with no external input. If you don't think about what you're doing or challenge yourself, there will be very few epiphanies. I think one use of this tutorial can be to provide you with a sort of sort of echo in your mind which can provide considerations as you're working. Like a ghostly Patrick Swayze helping to draw a pot... wait, he just messed it up didn't he? It will no doubt take you a while to digest everything I've written here, so you might need to come back to the text with a few months separation and perhaps understand that thing which was so confusing the first time you read it.
Sometimes, often, an image just feels wrong somehow. Here are a few of my go-to self-critiques.
Fix that thing! Sometimes a few strokes are all that's needed. If I've spent 2hrs on the design, taking 10 minutes to correct the mistakes / little things which still bother me is totally worth it. It will feel good knowing that the wonky looking arm or whatever isn't out there in the wild shaming me.Kill your darling! Sometimes I have a good detail, or a detail I think is good, but it still bothers me. Perhaps it just doesn't work in the context with the other stuff that's going on. Also, just because I've spent time on something doesn't mean it's somehow increasing the quality. Ultimately, I'm just a mammal making marks on a board. Think about the greater good! If I am designing a set of characters, they need a certain coherence. This means that I sometimes have to change a great design to something less great just to make it fit in. I always try to remember that I'm designing a set, not individual characters. A group of characters can be thought of as a single piece of work when going for unity. In some cases, I find it useful to imagine an ancestor/archetype, then I evolve that into several sets of monsters. The same can be done with armour, weapons, etc. Remove the unnecessary! If a detail or line does not contribute it must die or change into something meaningful or quiet. It is easy to add but difficult to take away. People will look at the details and structures and try to make sense of them. I try to go over my designs and figure out which details are just meaningless fillers. If I have a focus area with a meaningless detail taking up space, I try to change it into something functional or figurative. Areas which are not in focus should not attract attention with loud nonsense details. A good way to quiet them down is to just brutally remove stuff, or make sure the details are quiet and pleasingly decorative, flowing with the rest of the design. If they don't flow, they'll stick out and attract too much attention. Often when designing stuff like spaceships which have 'greebles' (random tech stuff) I start doing too much meaningless noise, and then I have to back away and figure out how to get a good silhouette and suggest a meaningful structure with the tech details. An analogy could be music where you have a theme, certain instruments, an overall shape, and also rhythm in the detailed parts. BAM BAM BAM ... dirudiru-lutt. Whatever. Random noise isn't memorable. Iteration! Throw away the first sketches and iterate the design. Sometimes a design comes out pretty good right away, but I've found that redrawing it makes it much more fluid because I've learned how to draw the essentials by heart. It doesn't take much time to make another rough. I won't know if it's actually strong until I've let it do battle with a few other variants. Flip it! Every time I flip (mirror) the canvas, the problems that I thought weren't there this time show themselves. I think my eyes are a little skew. Respect the source material! I try to stay reasonably faithful to the source material when doing designs based off old game characters. It is possible to stay faithful to the source AND do a great design. What's the point of basing a design off something if it ends up looking nothing like the original? I try not to make arbitrary details when I just as easily can use something from the source and increase likeness. If I can't manage to do that then I'm incompetent and only capable of simple, blind, low vocabulary reflex drawing.
I seldom do studies, but when I do I notice improvement very quickly. However, relying on reference will only get you so far. In order to be fluid and fast you need to be able to draw without having to think about the basics (such how humans look and act). Struggling the fundamentals will prevent you from seeing and focusing on the complete picture, its design intricacies and flow. That's where the fun lies.
To see the complete picture faster, and to give you something to hold on to, you can put down blobs and reference marks. You can then gradually add details without getting lost aimless pen wandering. However, at some point you'll have to grab the bull by the horns and do the final detail pass. Some artists can do this pretty early, perhaps because they are able to sketch in their mind and put ink on top of that. When doing studies, it's very rewarding to practice this... internalisation skill. You look at the subject, memorise the contours, volumes, relations, negative space, and put down the figure directly with a ball point pen. It's faster and creates more lively lines than those which have been "accurately" rubbed into place. By working this way you are able to do a higher volume of studies, and by doing the lines and curves directly, perhaps you will remember them more directly too? It is said we learn from our mistakes, but we need to be able to identify what went wrong (or right). Try to learn one thing at a time. It's hard to practice juggling while also trying out 400m hurdles. When learning new tools, pick a subject you know well, and vice versa.
- 25: Changed foreword (+ stuff above) to something I wrote mostly in 2014. Edited terminology.
- 02: Clarified stuff in the Terminology section. And minor changes to Seeing. Perhaps I'll work my way through the entire thing eventually.
- 06: Fleshed out the licence a bit.
- 04: Rewrote parts of the text. Added some more (old) illustrations. CSS layout & markup. Anchors. RSS.
- 27: New value/noise scibble. Spellchecked (tons of typos). Fixed erratic formulations.
26: Four new illos.
24-25: Wrote/compiled the tutorial.
Table of contents
This tutorial is, in its current form, free to translate and 'mirror' in that form. It is also free to distribute for educational purposes (as a hand out to students or whatever). It must remain free.
Because I may be updating it and new versions are generally better, I'd rather not have it mirrored too much. I can not update mirrors I do not have control over. Presenting it with ads or with intent to gain website traffic is for poopieheads. This is the internet and a link to this page should suffice if you want to bring this page to someone's attention.
I guess this licence comes pretty close:
If you have translated it, or have seen a transated version, please let me know so I can link it here.
By Arne Niklas Jansson - 2005 & 2007
Contact: (Hint: It's diglett at google's well known email service.)
Terminology (Upd. 2012)
A small glossary.
- HSV - The three dimensions of visible light, Hue, Saturation and Value. The structure can be viewed as a cylinder with a circumference, radius and height. Because we use 2D displays, this 3D structure is often unfolded into something like a hue circle/strip coupled with a value & saturation square/triangle.
- RGB - Computers generally display colours using a 24-bit colour data structure, divided into the Red, Green and Blue channels. I prefer to paint using RGB sliders because it allows me to easily add or subtract colors. If I have a gray, 120, 120, 120, and want to add some orange I might adjust like 120+30, 120+16, 120+2.
- Exposure - Just like the eye, a camera is a device which catches photons. When it saves an image it sort of saves a rough photon count per 'pixel'. When an image is converted for display on some sort of canvas, we often lose detail in the dark and light ends. You've probably taken a picture a sunny day where the sky looks washed out white and the shadows totally black. However, cameras can actually save images in a format called RAW which does not crop values as brutally. These images can be viewed on a 'High Dynamic Range' monitor which supposedly is capable of displaying the extra intensities. RAW images can also be used during editing to bring out details in the sky or shadows, but it can look a bit flat and fake when displayed on a regular monitor. When we paint, the inability to portray the natural full contrast, we're forced to resort to various tricks that somewhat emulate brightness. Painters like Anders Zorn were excellent at this.
- Hue - Variations in colour. Skin has some hue variation, such as cheeks being rosy. Even if an object is a single flat colour, it can still vary in hue because it's subject to coloured light and shadow, reflections and other effects.
- Saturation - One might say that saturation is much colour there is. Grey is the absence of saturation whilst neon colours are very saturated. A computer program might quickly measure saturation by looking at the difference between the two most extreme RGB values. On a computer monitor, dark and bright colours look less saturated because it's not possible to bring the sliders below 0 or over 255. Things like lightsabers often has a bright pastel or white core, with saturation around the edges to portray overall saturation.
- Value, Brightness, Darkness - There are several ways to calculate value. In the RGB colour model, Green is the brightest, then Red, then Blue (i.e. 0, 255, 0 is brighter than 0, 0, 255). In a program I wrote once, I used the following common formula to calculate value from RGB: Value= (0.299 Red) + (0.587 Green) + (0.114 Blue). If an artist praises a painting for its "great values", he probably means that, colour choices disregarded, the painting is effective in the way it uses contrast to convey forms and make things read well.
- Read, Readable - You can tell what something is supposed to be, what is described. If a painting is messy, it's difficult to read.
- Radiosity, Ambience and Reflected light - When light is reflected of a surface and then hits something else. This happens all the time. In a green room everything will get a green hue for example. Radiosity is a 3D rendering term, iirc. I use 'reflected light' for sources I can trace, and 'ambience' when the colour is some sort of average of lots of things in the surrounding environment.
- Specular highlights - Dots of light that appear on glossy or wet surfaces. It's really a mirror image of the light source that you're seeing. On more dull materials, light is perhaps broken up by very small bumps, so you won't see the specular dot moving as you shift your position. The lighting on the surface will look almost the same from different directions. A mirror is so smooth and reflective that light comes back almost undistorted from the reflective surface.
- Flatten - Removing unnecessary texture and values. 2018 note: Huh? Simplifying?
- Texture, Noise - Tiny details, sometimes repeating patterns, which sometimes used to show the material of surfaces, or just to provide interesting contrast and flavour. Photo models can look strange if they've been airbrushed so heavily that there's no skin texture. You can do neat things with texture, such as alternating a warm and dark color. This makes the average colour of the texture appear more... dimensional and rich than simple monotone colour would. Texture is a tricky beast to wrestle with, because textures can make a surface look too noisy, such as a tree crown, rubble, grass, or a brick wall. In such cases it's common to instead partly suggest the texture here and there.
- Form, Modeling, Sculpting - The 3D shape of something. Values (shade/light) are used to describe form? However, texture/patterns can also be used to to indicate planar angles though it's a lot of work and often you'll see people applying a flat texture for cloth, chainmail, stockings and such, especially in simplified (already flat) styles. Hatching lines can also be used, sometimes wrapping around curvatures (I do this a lot as of 2018 - hand-drawn lines livens things up I think). Hue alone can be used in some circumstances as an indicator, when there's a cold light from above and warm fill light from below and not much contrast in value.
- Plane - Can be used when talking about the Foreground, Middleground and Background. If your planes are too similar in value and detail level (in particular), they will be hard for the eye to separate. Often, you want the viewer to focus on the plane with e.g. the character/thingy. Plane can also be used when talking about a flat sub-surface of an object. Even though the human face has no flat planes, it helps to imagine it as a simplified polygonal structure when approximating lighting and foreshortening.
- Atmospheric perspective - Air is not fully transparent. As you may have noticed, things far away tend to fade away and sometimes shift towards a sky blue/grey colour. Atmospheric perspective can be used to visually separate overlapping parts at different distances (or planes). It can also be exaggerated. Some artists (including me) use atmospheric perspective on character pinups to help separate limbs crossing each other (and also to integrate the figure with a paper-white background).
- Local colour - I think this refers to the colour an object has when viewed under neutral lighting conditions. In reality, things rarely show their local colour because there's so much coloured light bouncing going on. Kid drawings and icons often feature overly simplified local colours, like green grass and brown tree trunks. You know that the paper you're drawing on is white, so you see the expected white regardless of which environment you're in. You don't see the cold grey gradient across it when you sit near a window, or it's beige tone at the library.
Seeing (Upd. 2012)
Apparently the conscious part of the human brain can't handle much information. It's more efficient to develop simple reflexes for common and repetitive tasks. There's a part of our brain which processes the image from the eye. After having processed it, very little information is actually sent to our awareness (some autistic people have an error in this process and they see 'too much' instead). Unfortunately for us artists, our brain also tweak our perception of the image so we don't really see what we actually see. To some extent, we must learn to reverse engineer our perception.
- Unwrap - We 'see' surfaces from the front, or several sides at the same time. The side of a box doesn't look skewed to us for example.
- Exposure - Often, an object in the shadow looks the same to the eye as one in the light, but when we take a photo of something we're often surprised by the difference in light levels.
- White balance - A white paper always looks white to the eye, regardless if we're indoors or outdoors. Cameras on the other hand are bad at white balancing automatically. If the photographer is careless, indoor photos may come out very orange (because of lightbulb light) whilst outdoor environments can appear too cold (because of the blue sky).
- Distance scaling - A person far away will still look human sized because the brain understands distance intuitively. However, I heard a story about a man who had grown up in a dense jungle, and thus never had seen anything at a distance. A scientist took him out on the savannah where a rhino was grazing on the horizon. The man thought it was an insect. As they came closer, the man thought the 'insect' was growing.
When drawing from flat images it's easier for us to see proportions, alignments and perspective distortions. In a way, a part of the job has already been done (though perhaps not in an desirable artistic way). It's a lot harder to make studies from life, but currently pictures from cameras don't give us the whole picture. They alter colours, crop out values and ignore subtle hues... not to mention that the lack of stereo and depth.
The onion (Thinking in layers)
Before laying down a stroke, there's a number of things you need to think about. Well, actually you shouldn't have to think about them, it should just go automatically.
- Feel volume and angle of the form.
- Where is the light coming from?
- Try to figure out if there are any shadows that might be falling on the surface.
- Is there any reflected light (radiosity) that hits the surface?
- What is the ambient color of the scene? (sorta like global reflected light.)
- Any speculars. Is the surface gloss/wet and also angled so it reflects a light source, such as the sky?
- The exposure level. Perhaps it's so heavily lit that it becomes more than white? Perhaps it's so dark that even the brightest spot is hidden in darkness.
- Is there any fog in the way?
- The texture of the surface.
Note that this mainly goes for realistic styles. A brushstroke should also look efficient and consistent with the rest of the painting and your color scheme choice. You might also have an idea or style which disallows certain colors or textures and puts priority on other things. However, even in a powerpuff girls illustration there's simplified elements of realistic rendering. Don't hide behind "it's not apart of my style so I'm not gonna learn it".
There's really just one kind of light. It bounces. You can only see the light (photon) if it enters your eye. Light does two important things when it hits a surface. First, a part of it is absorbed. This is how colors are made. A red apple reflects mostly red wavelengths, the rest are absorbed and turned into heat or something. That's why black stuff get so hot in the sun. Anyways, the reflected light bounce away differently depending on the surface. If the surface is bumpy it will bounce away sort of randomly, like a tennis ball that hits rocky terrain. If the surface is smooth it will bounce away in a predictable path. A mirror is very smooth so the light comes back undistorted, so we can see our reflection.
Note that all surfaces have speculars, because speculars is just reflected light. It's just more broken up/diluted on dull surfaces.
Depending on where the eye/beholder is, it'll see different light and different specular spots on a curved surface such as this. A puddle isn't curved (other than the edges because of surface tension) so you'll only get a shiny reflection from a certain point of view. Point speculars can only appear in an environment where there's a point light source, like a sun, lightbulb or small window.
Photo - Speculars do exist on cloth, diluted and subtle. I stretched out my shirt sleeve with two fingers to get a flat surface between the two marked dots (I moved the camera and not the sleeve).
Here on earth we have lots of stuff around us that the light can bounce off, so things here are more or less lit from all angles. For example we have the sky which is like a dome shaped blue light source. Then theres the ground, walls and other surfaces. In space there's basically just one light source, the sun. This is why the moon just has a lit and shadowed side, and looks kind of flat. If you looks carefully however, you can see earthlight on the shadow side of the moon, but it's very weak. Then there's starlight, which I guess is even weaker.
When light hits a surface and bounces, it also change color. If it hits another surface of the same color it bounced off, it will make that surface look even more saturated.
(Too orange to be some sort of skintone anyways.)
The sunlight is much stronger than the skylight, which is in turn much stronger than indoor light. Our eyes adapt automatically after a while, and we can also adapt by squinting or just focusing on an object. Because we do this without thinking about it, it's hard to understand that our eyes are actually kind of limited. This limitation becomes even more obvious with cameras. If you take a picture indoors, the windows will become overexposed (bright). You might try to adjust the exposure levels to the window light, but then the indoor environment will become underexposed. This can be used to your advantage. By for example putting a character or object in the foreground where it's darker, you can make the silhouette read well against a well lit room.
The exposure to light can also make parts of a body look very bright or dark, not skin tone color at all. When the shadow is dark and the lit side is overexposed, the only place for the color to go is on the edge between them.
Here's an example of various materials and how i render them.
- Cloth - Hardly any speculars, just shadow and light. Sometimes strong light can penetrate thin cloth and cause some sort of subsurface scattering.
- Leather - Can be a little gloss and thus have a few speculars. Also, don't make it too saturated.
- Trees and wood - Dull. Not very saturated either (sort of grey-brown-sienna).
- Stone - A bit like cloth. The surface is often to rugged (both at micro and macro levels) to have any serious speculars.
- Plastic - It seems like the speculars and reflections are colorized in the color of the plastic. Plastic can also be a bit transparent.
- Gold - Gold isn't orange. I use black - desaturated orange sometims with hints of green, then up to yellow and white.
- Silver - More or less like a mirror.
- Metal - In the case of armours I often push the values a little more, not as much midtones.
- Brushed metal - It's sort of like an inbetween of a grey surface and a silver surface.
- Glass - Often just transparent with distortion. The speculars come suddenly and are often white. In the case of car windows you might have noticed that it's easier ot see what's behind if there's a shadow over the window. The brighter reflections obscure.
- Wet stuff - more speculars, can become transparent in the case of cloth, and stones get more saturated and pronounced details.
Shadows are quite flat and generally less saturated than the lit side. It's easier to notice ambient light in the shadows. Shadows get blury over distance, this is called diffraction.
(Shadows don't add (multiply) with just ONE lightsource that is...)
Consider the environment. The light is stronger outside, and the skin color tend to be less saturated due to the sky blue ambient light and sky blue speculars. Sometimes the skin color become shifted towards purple because of the sky blue being mixed in. This is especially true if the subject is standing in a shadow.
Indoors (no windows, only light bulbs) the light is warmer and allows skin saturation to be amped up to oranges and reds.
The shadow color of the skin can sometimes wander off to greens, especially if the room have green components, like wallpaper, plants, furniture.
In a white room or a bathroom the skin tones would be quite pale, closer to local colors and less contrasted (shadow/light) due to lots of ambience.
A room with a single strong light source will probably result in near black shadows.
...so, the type of environment your character is placed in very much affects how you should render it.
The human body has a lot of different hues. Parts covered by cloth gets less tan. Mons pubis, the hip bone area and the chest is quite pale. The shoulders and lower arms gets a little extra tan. The inside of the lower arm is often pale however. The kneecaps and elbows have a little darker skin. The face also has a lot of hues, such as rosy cheeks, males might have grey or almost green jaws because of stubble. The best way to learn the hues of the human body is to make studies of course. Don't forget that animals, monsters and objects also have hues. If you paint everything with the same hue and saturation it will look boring.
Some hues are due to ambient and reflected light. The shoulders and surfaces pointing up can get a blue hue because of the sky reflecting.
Saturated gradients - The gradients between the shadow and light is not just an in-between color of the shadow color and light color. If the shadow and light is just blended, it will look very lifeless. If you look at pictures you will see that the gradients is saturated. It's especially easy to notice if you remove that saturation.
Sub-surface scattering - Strong light can penetrate the surface of some materials and bounce around, then exit again. This will increase the saturation and make the surface look illuminated from the inside. In the case with human skin, we sometimes see it on hard edges between light and shadow.
- Leafs are gloss on the top side which means there can sometimes be a sky blue specular here. The light shining thru the leaf makes the bottom side more saturated, this is also true for ears and fingers, which can turn super red when heavily backlit.
- Sub-surface scattering on the fingertips. The light on the left side of the thumb is probably light reflected off the index finger.
- Note that the edge only appear if the light is overexposed. It does not appear as pronounced on the thumb.
Colors and values are relative. By using various tricks it's possible to trick the viewer into thinking a color is really another color, or a value is darker than it is. Unfortunately, the artist is also tricked into using too much colors or values than is needed.
A hard edge between two values will be much more obvious than a soft one. You'll have to know when to use which. Sometimes your choice of values is very limited, such as when you're working in the shadow. By using hard edges you can describe a lot more detail with less values available. However, using gradients is very useful for changing value without the viewer noticing. The 'fake flat' illustration looks flat, but is actually a gradient. The square is the same color as the left side of the 'flat' rectangle.
Colors with the same value are relative in terms of hue instead. A common mistake is to draw one detail too saturated, then something else nearby looks grey, so to compensate you increase saturation on that detail too, and as a result the whole painting end up too saturated.
It's easy to get carried away and go over the top with highlights. This makes it hard to see what color the subject is. Instead you should use shadows to describe the volume of the subject.
Flatten and simplify
Work with larger brushes and remove unnecessary brushstrokes. See the bad and better example below. I really didn't do much on the second one. It's actually simplified. It's surprising how much a little flattening here and there can do. I did spend some extra time on the face though. A bad face can ruin everything. Image is from reference.
A painting is a hierarchy of important and less important details. If you're doing a pin-up the main figure and silhouette is the most important. In comics they often use a fat outline around the silhouette, whilst the less and less important details get thinner and thinner lines. When painting you do the same thing, but with brushstrokes instead! You use differences in hue, saturation, value, edges, sharpness, detail and composition to lead the viewer's eye towards the focus point of the painting.
If you use the same rendering everywhere on the painting it will look flat. You can lead the eye toward important spots, but once the eye is there it needs something interesting to keep it there, like proper details. The amount of details on a spot should be proportional to the amount of time the eye stays there.
Attempt to isolate some of the techniques you can use to attract the eye.
It can be very dangerous to get excited about rendering details, especially at an early stage. You can not render details the same way in the shadow as in the light. On the second one (B) I just rendered all the details to demonstrate how it can look if you just scribble down all the detail without thinking about the important forms (A). (C) is still a bit confusing but that's more of a construction issue. Side views can only get you so far, and the anatomy is pretty odd which makes it harder to read.
A is the form without detail, B is the detail and C is both. Be careful not to do too much B, the form has to read!
Perspective and construction
I'm not going to have a long perspective tutorial here. Instead I'm just going to mention how much easier it gets if you make a few guide lines to align the figure/environment after as you draw.
If you're reluctant to do environments just because perspective is tedious to set up, you can use a simple 2 point perspective and just guess in the details. It's surprising how much easier it is to align things correctly with a few guide lines (for me anyways). It doesn't have to be perfect to appear correct. There's a risk that your designs will suffer if you can't improvise quickly and have to consult the ruler every other stroke. First draw a horizon line, then radiate lines from both ends in a random manner, then just crop and sketch away.
Here's a few quick freehand lines helping me to align the shoulders and stuff. When drawing people, it can help to align stuff after a 'spine'.
Exaggerate - One great thing about art is that you can exaggerate things, like hips and boobs. Haha, no, actually I'm serious. It's a good thing if important curves are more pronounced.
Simplify - The advantage art holds over photos is simplification. In a photo you'll get distracting details. When drawing, you can remove objects that aren't relevant to the scene. Wrinkles and minor protrusions can be removed to get a better line flow. A common mistake I see is when someone has drawn all the abs (belly muscles) with an overly amount of crosshatching. It's better to leave out lines, especially if you're going to color, because then the contrast between different color fields can work as lines.
Harmonize - Another word for this is 'swooshyness'. Unlike the above things it has to do with the relation between details and how lines intersect and take over ofter another. Try having a few swooshy lines that you align several parts after.
Stylize - When going for a style it's important to be consistent. You can turn curves into hard edges, or you can go for sweeping sinus lines. I prefer a combo where I turn a curve into a hard edge at a certain threshold.
Line weight - With a few exceptions, I'm not a big fan of fixed line width. Here I'll attempt to devise some general guidelines for when and how to vary line width.
- Lines are thicker on the shadowed side, thinner on the lit side.
- Lines are thicker near the viewer, thinner further away.
- Silhouette lines are thicker. Inner details get thinner lines.
- Lighter materials get thinner lines.
- Thin lines works good with detailed motifs.
- Thicker lines work well with simple figures.
- Fat line art works well with flat colors (cel) .
- Thin line art works good with realistic rendering and pronounced volumes.
Also, you do not always need to draw a line. Sometimes it you can just hint the ends of it, and the eye will fill the rest in. Examples are places where skin is pushed together, like the mouth, buttocks, pushup boobs etc. It's good to make the line a little thicker where there's a gap. Examples are places where clothing stretch over gaps between muscles or... cleavages.
Finally, here are a few illustrations. The first one is Photoshop (5.5) + Wacom tablet and there is no line quality to speak of, but I hope you get the idea. Second one is an inked thing from a few years ago.
When painting, subractive edges which are remnants from the line art or construction stage can be harmful. You can separate shapes with other means than black lines, such as different value, or a bright line. Here's an illustration of a few possible solutions and what they convey.
A line art image might go several stages of refinenement where previous, rougher stages, are constantly faded and painted over. Final stage not completed here.
I didn't start making studies until just some years ago, and that I regret. You won't stumble upon the right lines with guesses and wild scribbling. Using reference will only get you so far. In order to be loose and fast you need to be able to draw without having to think about the general construction of things (such as basic anatomy). This way you can concentrate on the actual design and detailing.
Construct your drawings/paintings. Don't march around with the pen doing detail by detail. Always check the general proportions and don't get caught by details too early. Place marks where the important features are. These well be like landmarks you align after.
Try to learn one thing at a time. You can't learn juggling while also practicing 400m hurdles. If you want to learn how to handle the medium/tool, try making studies of easy stuff like fruits. If you want to learn human faces, use a medium you know so you won't have to struggle with that as well. If there's several things you can't handle then you won't see what it is you're doing wrong. Quantity is also important. I wouldn't recommend making anatomy studies with tedious woodcarving tools for example.
Lineart - Pencil studies doesn't have to be more than a few quick pencil thumbnails on a paper. I spend a few hours on studies when I do them, and I put about 10-30 on each sheet (A4). I only do a couple of studies a month, but I certainly notice improvement each time. Just imagine what would happen if you did them several times a week for years.
- - Some older study sheets made from a comic (poses) and photos (faces).
Painted - When I do digital studies I use Photoshop and mostly a picture from the net. I duplicate the window and clear the new one. Then I start placing the larger color masses on their approximate positions. After that I gradually increase the details and value/color accuracy as good as I can. When it looks close to the original, at a distance or with the eyes squinted, it's finished. I always work with the largest possible brush allowed to render a given detail. I don't color pick from the original, but I do keep the windows in the same size so I can see if I misplace anything. If you want to increase the difficulty (and reward) you can always try to draw in in a window with different size, and mirrored, or paint something from still life.
Statue study: | - First one is OC (Open Canvas), second is PS (PhotoShop).
| - Here I tried to be as economical as possible. Second one isn't a study but an example of a cleanup, which isn't necessary for studies.
| - These are made from reference. I took the liberty to add some line art and style.
- One of my first studies.
Subjects to study
Study everything! You need to build a large library of shapes and things in your head to be able to draw intuitively. This takes about a lifetime or more to do, so you better start now!
Human anatomy - One of the most important things you need to know. Even monsters have traces of human anatomy.
- The whole body. Use photos, anatomy books, statues or real people.
- The face is the thing we look at first. If you misplace a line just a bit the whole expression of the face will change. Make studies of photos, your friends or yourself.
- The hands are also important (and hard) to learn.
- The feet can be tricky too, not because of the shapes, but because you need to plant the character on the ground so it doesn't look like it's falling over or the ground is leaning.
- Daily clothing. It's important to learn how cloth wrinkles, how different types of cloth looks and fits.
Gestures & styles - You need to be diverse and get fresh ideas. Learning some different styles can be a good idea.
- Draw from life using your friends or people at a cafe, a bus or somewhere. How does a person pose when he opens a door, reaches for his keys, and looks intimidated by an artist?
- Marvel. How does the Marvel artists represent the human body with lines? What details are important and what is simplified?
- Modesty Blaise, or some fairly realistic comic style. Drawing gradients with just blacks and whites isn't easy.
- Manga or a style you like. Again, how does the artist convert the human anatomy into lines and color blobs? What parallels can you draw between the different styles?
Environments - Putting your character in an environment really brings it alive. This is something I definitely need to learn myself.
- Landscapes with fields, mountains or whatever.
- A dense forest or a jungle.
- An urban or industrial 'landscape'.
- An indoor setting, like a room with furniture. Boring, I know. To be honest I haven't done this yet.
Fetch an animal book - ...and draw some animals. A good way to design a monster is to morph different animals into one.
- All living things. Mother nature have spent millions of years perfecting the designs, so you better study them.
- Horses, Cats, Dogs, birds. These are especially important since they are more commonly seen.
Machinery - You also need to practice drawing machinery. It can be useful when designing robots and planet-smashing vengeance-crazed battle droids.
- Cars of different models.
- Digging and working machinery.
- Military vehicles.
Classical still life objects - Or basically anything. Good for learning how to draw and paint in general, because of the simple shapes. You won't have too struggle much the shapes and can concentrate more on the materials.
- Flowers, fruit, skeletons, sculpts, chunks of wood, rusty metal parts.
Analyse what you're doing wrong. It's easy to get blind from staring at the image too much (which you must do to be able to work of course). Try flipping the images, look at it upside down, through a mirror (I use a CD), and zoom in and out (or back away). Don't sit and nibble too close to the paper or zoomed in. You can also make a 'New view' (PS).
You must also accept that just because you have worked on something for a while doesn't mean it's worth anything. You must be ready to sacrifice the time you spent on something if it looks wonky. Even if you're happy with the detail you might have to rework it (Kill your darling). Sometimes it's not the the detail you're concentrating on that's wrong, but something relative too it, like the value of the background or the perspective of another detail.
Orders of importance
Very generally speaking, certain apects of a painting are more important than other.
- Construction - What are you trying to paint? Your subject and composition should work on a fundamental level. If not, then no rendering in the world can save it. Don't have any illusions that you will be able to salvage the piece later. If a pose look wrong now, it will look stiff when finished too, even if Rembrandt himself painted it.
- Values - For a painting to work you need to use values to sculpt the forms. Values can do a lot of work grouping and separating shapes. - The first version here is obviously wrong. Each shape has just gotten the shadow and highlight treatment. Second one is better but there's just one value type. Third one has different values on different shapes. Maybe it fails at the construction step though; it's not a very interesting pinup. Here both value and color is used to separate the foreground and background, although I don't like this painting either, again it fails at construction.
- Color - You can be a little off with the colors (hue and saturation) and still get away with it. If you just can't make the colors work, it is probably the values that are wrong. On a side note, if the previous steps do work, it's easy to make fresh looking images with color balance tools. In my experience the original choice is often the best.
Critique and common mistakes
When I give critique on various forums I often end up typing the exact same things every time. Here's a list of the common mistakes beginner artists make.
Problem: To go shadow - midtone - highlight on all shapes, regardless of location and angle of the shape.
Solution: Try to zoom out, flip it, turn your head upside down. Don't render each detail individually one at a time. Equally lit minor shapes flattens the painting and makes it hard to make out the important major shapes.
Problem: To mix black into the shadow and white into the light, and then smudge the in-between colors.
Solution: This makes the painting look grey and dull. I've encountered a few people asking for 'shading tutorials'. There's no such thing. There's no shortcut saying you should start with a dark color and end with a bright color, then do that on every detail. You must learn how light works and almost render as if you were a 3D computer program. Some might knee-jerk about that, but for me it's true, I often get the best result when I let go of my bias and habits and just follow the 'render rules'.
Problem: To render and highlight details that will only distract the viewer. Dodge brush is real a sinner here.
Solution: You will have to sacrifice a lot of details that could have looked awesome fully rendered. What's important is the wholeness of the painting. So, no flares on the belt buckle on the unimportant little guy in the corner. The eye homes in on highlight, contrast and saturation. You should lead the eye to the important parts of the painting.
Problem: Flat and stiff poses, figures look fly-swatted.
Solution: I do this a lot. It's so easy to draw people from the side or front with arms stretching out. This will end up looking very boring and undynamic, unless you have a compositional idea, or you're doing an icon. It's tempting to try to show all the details of a character in one drawing, but if you're going for a dynamic pose that is rarely possible. Letting the body obscure an arm might be sacrifice you have to make in order to get a good pose. Also, when hiding a detail, you'll let the viewer's imagination decide how it looks, which can be a good thing! It can look better in the mind than it ever could on canvas. Anyways, Learn foreshortening and dynamic poses from doing studies of comics and real life models or photos. Fly-swatted characters will only get you so far.
Problem: Bad stroke economy, or attempt at artsy 'ooh, lookit I'm so spontaneous'-strokes.
Solution: Each stroke should be contributing to the piece. Some hope to get away with random artsy strokes. Random strokes makes random details. Random details distracts the viewer from the actual details of the painting or drawing. I know it seems like some artists paints a perfect scene with just a few swift well placed strokes, but I think they actually do a lot of cleaning up and optimising. Go over the entire painting piece by piece and remove/flatten strokes that aren't contributing. Make sure the motif gets the focus and not some artsy stroke. The major shapes and volumes are always the most important and many neat strokes will have to be sacrificed in order to make the whole thing work.