Power to the (pixel) people

another dumbass who can't draw...


Well, it seems here I am at the start of another graphics article and like a 99% complete flood-fill I find myself surrounded on all sides by strange pixels. :) In this article I'll jump into the world of layers and offer a few tips to newbies and pixellers who want to move into a pan-dimensional-layered world. ;)

The old days of pixelling are sadly almost extinct, gone like the do-do to the spirit in the sky (and no, I'm not talking about the duty-free drink on an British airways flight ;)). Now in place are super complex, bulky programs which often chew up more CDs than an ELSPA anti-piracy crushing machine. The price tags often require a degree in higher mathematics to help you understand that you can't afford them, well, not without selling both kidneys. The range of functions can be very daunting to newbie and intermediate artists, in fact some oldbies might find it difficult to get started too. I admit that I found it hard to make the leap from D-paint pixelling programs into the realms of layers, objects and filters that the likes of PhotoShop, PaintShop Pro and my personal favourite Picture-Publisher v8 all have.

Upgrade your machine now!

This may sound strange, but artists probably need more powerful machines than anyone else. The amount of disk, memory and cpu resources your jazzy new paint-program chews up is really shocking. With anything less than 64 megs and a Pentium II you will soon find it difficult and frustrating. Why? Well, in one word: "layers!". These are without question THE most important feature and will change your mental approach to drawing a piccy. Take the low resolution of 640 x 480 x 24, that's 900k, now imagine just 10 layers (yes, 'just' 10 layers), that's 9000k. But wait, each layer needs a mask or special data channel to enable the paint program to handle transparent or semi-transparent areas in each layer. This is normally 8 bits, so 640 x 480 x 32 for a layer means 1200k and 12000k for ONE image (ouch).

In order to display one image your CPU must perform lots of overlay/merge/mix operations (in the above example 9 of them). This chews up CPU time and we all know how lousy using a sluggish program is. When drawing graphics you need almost fluid feedback. Can you imagine using a pencil where the black marks on the paper lagged 5 seconds behind the pencil tip?


Layers = more work!

This is true for both your machine and yourself. As you probably know layers (or 'objects' in some programs) can save you lots of time because you can easily rearrange your picture, change the composition, shade, lighten, merge and duplicate parts of it. In fact you can have multiple pictures (or versions of it) within the same image by using layers. Want to add a reflection? then create a new semi-transparent layer and draw a suitable highlite. You can visually mix multiple layers to create beautiful effects without having to modify a single pixel. Thus saving hours of drawing.

The most obvious task for layers is for splitting the foreground, mid-ground, background and sky of a landscape (something I've done a lot recently). You can arrange these components and mix them until you achieve something you like. In a traditional pixelling program you would have to cut, paste and then fix every pixel in the hole left behind.

But wait, you said "Layers = more work!"

Yes, I did. When you first jump from pixelling to layers you may be tempted to only draw the visible parts of the final image, ignoring those parts on the background layer hidden by the foreground. This is a mistake which can make you more work in the future. Now suppose you want to re-arrange the composition and move that huge foreground object (a rock for example) to the opposite side of the screen... Oh, no. there is a large hole in the underneath layer where you didn't bother to draw behind the object. Trying to patch up a hole can lead to horrible results because you have forgotten all the countless steps you took in creating the surrounding pixels.

So my advice is to draw "complete" objects where possible. Even if the object will be half hidden by the edge of the screen, you should try to draw more of the object then you think you will need. Call it a 'recomposition safety margin'.


There's no place like Clone

One life saving feature that any artist will love is the 'Clone-Brush' or 'Clone-Paint' function. This allows you to copy from one part of a layer to another using two cursors (one acts as the Source area and one as the Destination). Picture Publisher (gee, this is starting to sound like a commerical or something.. I don't work for Micrografx.. Honest!! I just like their proggy) allows both transparency % and Merge-mode to be used with the clone function. Using transparency on a previously black area can lead to a simple shading effect and likewise on a white area can give a highliting effect.

Try 50% transparency and set the Source and Destination cursors close to each other. Move the mouse away from the Source and you should see that it gives a shaded sweep effect. It's kinda like a feedback thing where the transparent Destination is fed back into the Source.

The 'Merge-Mode' determines how pixels are modified (like a 'Write-mode' in programming terms). These can be very effective in combining two areas, shading, colouring and creating reflections (although using a separate layer is a much more flexible method for relections). There are modes for Additive, Subtractive, If lighter, If Darker, Filter, Multiply, Difference, Texturize, Color, Hue, Saturation, Luminance, Red, Green, Blue, Invert and Screen. As you can see there is a wide choice and almost every mode is useful in different ways. For example, want to keep some colours but want to change the texture? then try using the Texturize, Saturation or Luminance modes. The results depend on your image, but don't be afraid of experimenting, you will certainly find some interesting results.


Shading without changing.

It's often easy to slip back into your own pixelling ways of thinking when it comes to shading and highliting. If a certain object/layer is too bright then you may immediately jump to the Effects menu and reduce the constant and/or brightness, but you may regret this later on because those pixels are now modified forever. Instead you can use another layer to shade the current one. Something like Subtractive, Multiply or the Luminance Merge-Mode can produce the same effects without modifying a single pixel.

Using layers mean you can separate out colour, texture and brightness (not to mention the R,G,B, components). Layers give you more opportunities for experimentation and UNDOing earlier mistakes. Although it may be confusing at the start (because you have to work with 10..50+ 'pictures' instead of just one) after a while you will never want to go back.

Want a simple 'ground mist' effect on your mid-ground objects?

Then create a cloud layer, combine this with a gradient-sweep from white up to black then overlay this layer as Additive with a high degree of transparency. After a little bit of experimentation you should have a misty layer which look realistic and can be instantly moved, modified or deleted.


Stop it, that tickles!

The Feather-edge operation can be useful for creating smooth edged objects. This basically affects the transparency of an object/layer mask edge. Instead of a 100% outside 0% inside edge mask (which gives a very sharp edge) the feathering smooths the transparency mask over a distance of N pixels, for a small value this is hardly noticable but can be good for helping two contrasting layers/objects to be more seaminglessly merged. Using a high value gives an extremely out-of-focus effect, which is nice for web-page buttons or designer style background logos.


Smothered by Smoothing

This is something I was guilty of in the past (at least I hope it's in the past ;)). Using lots of filters, trying to undo changes and scaling up small graphics, these are all pixel crimes which can lead to horrible results, also not using layers correctly (if at all) will give blocky poor quality images. There were some comments about the inferior graphics in Hugi #19, yes I admit they weren't the best and were due to my newbie-ness (hmm.. you won't find that in your dictionary I guess.. hehe) with my now favourite paint proggy. Since then I have repented, learnt my lesson and finally finished replying to all those death-threats from PC artists.

So please, take some friendly advice, use smoothing functions with great care and only in small doses. Also, change your desktop resolution to be same/close to the image resolution so that you aren't working with a completely different pixel-aspect-ratio or trying to edit graphics which appear half size on your vast 1024x768 desktop.


Big is beautiful

Scaling up small graphics can have truly horrible results. So plan your composition carefully and try drawing bigger items than you think you need, you can always scale down and lose some detail, but scaling up even with bi-cubic-interpolation or pan-dimensional-image-coloring-filters can not create more detail, it only creates blurry blocks.

As most pictures have lots of similar areas (self similarity) you can use scaled down versions of the same object, or parts taken from a big object to be used as smaller ones. For scenes like forests, deserts and the like this can save lots of time and means you can more spend more effort with the finer details later on or experimenting with new ideas.



Doing the details

This is, or should really be, the last stage of drawing. I know its all too tempting when you're in the middle of drawing one specific object to add lots of detail, spend ages shading etc.. but this can lead the problems later on. It's a trap I've often fallen into in the past. You end up with one fantastic 32x32 object surrounded by a 95% blank screen. Trying to re-create a texture, a certain look or style can waste hours. So instead take a 'zoom-in' approach. Draw a rough outline and see if you're happy with the composition, fill each area with a base colour, possibly added some gradient fills to try out the overall lighting/shading, then move onto textures and finally when you're happy zoom in and start adding detail (pixel by pixel if you wish).

Keep it simple. Of course everyone likes highly detailed images rich in a wide range of colour, shape and texture, but try to break up the picture into areas, concentrate on those areas which are the key focal-points and spend more time adding detail to them. As a general rule the brighter the area, the more detail it needs (dark areas can often make do with 'hints' of the object rather than pixel-perfect renderings).



Tempting Textures

The use of textures can really help to bring a picture to life. They can add terrific amounts of detail with very little effort. When you first start experimenting you will probably just fill an area with a texture, fine, but this is only the first stage.

Almost every texture is a flat, 2d image with very little variation of colour or shading over its entire area. If you drew a picture using just textures then it would look like a 1995 (or before) texture-mapped game/demo. Each area edge would look sharp, un-natural like a low-detail polygon object. This is where Perspective, Distort and a few Effects can help blend these areas together in a more realistic manner. A 'Drag-Morph' or similar function can help to blend edges together, to re-model objects and produce sand, dirt, water and a whole heap of other effects. Our brains like detail and variation, so don't use the same texture for more than say 20% (or less) of the picture, try scaling up, morphing or combining multiple textures until you find something different. Adding shading and highlites is another technique to make a flat, dull texture become more interesting.

Remember, you can always create your own textures. The default ones in paint-packages are often past their use-by dates (let's not forget, a million other artists have probably used those same textures countless times before you). So experiment with obscure textures which you wouldn't first think of using for a particular task, merge it with other sources, invert it, flip it and generally mess around until you like the result and SAVE IT!! You will be amazed how often you find that an old, dusty texture on an unfinished picture will be useful later on.


Whoaa! Get your shades on!

The colour palettes, like textures, are often produced to look more like a rainbow explosion rather than a stuble, pastel, ready-to-use block of graphics. One of the hardest problems, for me at least, is getting into the habit of using softer, more natural looking colours and shading (my small brain cell seems to be locked into CGA style, "poke your eyes out with colour" RGB palette values). Of course while drawing objects those highly similar, natural colours will be difficult to tell apart from each other, so see if your paint-package has a Gamma-correct or Monitor-correct feature (if not simply turn up the colour controls on your monitor), this will reveal more detail and make editing the shading much easier.

You can start drawing using bright, rainbow colours then use a Colour-Saturation, Gamma-Correct or better still, layers to produce more pastel shades. This is, IMVHO, a good way to do things because you can gradually blend items together by losing detail (in this case the saturation of colour) and makes drawing/editing things easier. But this is just my opinion, I'm sure other (much better) artists will have their own methods and tricks.


Dry brush, wet white paint.

One of the later touches to a picture is to add small specs of highlites to various objects. This can add the sparkle to an image and make it come to life. By specs I mean the very thin edges or spots on an object edge where a light source hits, in the case of a cube this would be at each corner and perhaps a little along a few edges. For things like rocks a 1 pixel highlite can really make the edges look very sharp and more realistic.

As a rough guide, use highlites in the middle of an object to make things look smooth, and highlites along the edges to make objects look sharp.

Remember to use highlites very sparsely and try to vary how you apply them to objects. Don't highlite every rock in exactly the same way, add a little bit of randomness to the brightness and shape of them, this will add more detail without making the entire picture look like a tiled texture repeated across the entire screen. Our brains are good at spotting similar items.

Shadows on the other hand can be used more than you think. They add a little mystery to a picture. Just like a horror movie the audience's imagination will be working trying to think whats hiding in the dark. Dark areas also help to complement the light ones adding more interest, more variety and helping the viewer to focus on the key points of interest (i.e. the brighter ones).


Funky screen edges.

This is an often overlooked technique to add interest to a picture. A normal rectangular edge suits most images, but having something different can really change the atmosphere and mood. Try feathering the edges, add a rough, torn border or motion-blur an irregular shape around the image. A mist/cloud effect can give a magical, dream-like quality. For a film-like effect try using a letterbox shape with a large black top and bottom strip, so something like 640 x 380 will make an image look cinematic like a wide-screen movie.



Cheat, use a scanner!

This is something which I have only recently discovered and makes drawing far, far easier than using the mouse. Perhaps a decent quality electronic pen and pad would remove the need to use a scanner (maybe in the future I will look at buying one). For the time being I find that using a pencil and paper to draw a rough sketch, scan it in, then fill, texture and shade it using Picture-Publisher to give my best results. (Which, btw, you can hopefully seen in this and the next issue of Hugi.)

Like I said before in a previous article, scan using greyscale and then mask, create layers/objects etc.. to quickly get an idea on your screen. From there you can employ all the nice features, filters and flipping that your twisted imagination can think up to produce something unique art.


Closing words.

Well, that's enough of this. Don't be disappointed if your first offering isn't the masterpiece you thought it would be, just go with the flow, experiment with new ideas, look at other people's art and see how you would construct it yourself.


And above all else.... Keep on drawing !!!

TAD/Hugi