back to main

 
Well, I've finally gone and done it... I have finally decided to make a tutorial on doing cg, something I've been forgetting to do for a while now. There is no right or wrong way to do cg, and I personally have more than one way of doing it, but hopefully some of what follows will be helpful.

CG Start to Finish

Part One: Outline in Photoshop

The first few steps cover creating the lineart. If you already have lineart to color, you may skip ahead.

Step 1: the Sketch

To begin, I open up Photoshop and make a new canvas. For the sketch, I want my whole canvas to be visible onscreen at 100%. This is so it doesn't take much time for the computer to render my strokes, and also so I can see everything. The sketch is quick and messy and I don't want to sit there with a giant canvas squinting at my sketch lines and waiting for Photoshop to finish rendering my last stroke.

For this image, I started off with a basic idea of what I wanted to do: one character would have his arms over another character next to him. It's always a good idea to have some idea of what you're doing. Since I knew the basic layout in my head, I could approximate the size, and began with a blank 800x500 canvas.

Next I made a new layer for my sketch. I set my brush to 3-pixel, hard, without any tablet sensitivity. No shape or opacity variations. You can do your sketch with any brush, really, but I find a solid 3-pixel one suits me best. Also, my computer has been having memory issues recently (need to clean up the HD) so Photoshop can't figure out the tablet pressure sensitivity thing. XD

Anyway! I start by drawing a very rough approximation of where the heads are. I decide the heads are a bit too far apart, so I lasso one and drag it closer to its friend. I also decide one of the heads is a tad too big, so chop it down a bit. Don't feel bad if you have to move, resize, or redraw things at this point of the process. It means much less work later. You shouldn't be redrawing too much, though. This is a sketch: something really rough and unfinished. It doesn't have to look perfect. In fact, don't obsess over the details! Just focus on the basic things, like layout and flow and shape.

Here's my finished rough sketch. It took less than ten minutes. (I've been typing this up while working on it.)



Now I decide there's a bit too much space on the sides and not enough on the top and bottom, so I crop the image some and add pixels to the canvas. I also tweak the shoulder on the left a tad.

Step 2: Refining the Sketch

So now I have this rough sketch, but it's fairly small, roughly 800 x 500, and when working on a computer, you should always work nice and big! So I resize the whole image 300%. It could really be bigger, but this really isn't a major project. Next I make my rough sketch grey by changing its layer opacity to 30% (or thereabouts; this isn't an exact thing).

Now I make a new layer and start drawing again. I do this drawing at 100% using my 3-pixel brush. The rough sketch becomes important, because without it to use as a guideline, working at 100% would be impossible. It's too big for me to see well! I could draw it at 50%, but then it would be a bit rougher. It's your choice. I do zoom out to 50% often to check my work and for drawing longer lines.

I also make a second layer for difficult overlapping areas, in this case the hand draped over the shoulder. This let's me work out the specifics of the hand without messing up the line of the shoulder. Normally I would also do a separate layer for the hair. After about an hour, I'm finally done. I erase the lines that won't be seen due to overlap.



Here's the partial lineart over the sketch. (The sketch has been colored blue to make it easier to see.) Not a lot has changed. The positioning of the eyes on the left is the most noticeable thing.



Flipping the canvas horizontally, I can see my characters have a rather annoying bit of tilt to them.



This is a frequent problem, and actually, it's not such a terrible thing. Art is made to be viewed in a certain position. Personally, I have double the trouble when it comes to skewed images because of my astigmatism. You absolutely don't have to worry about flipped things. Look at some of your favorite artists, flip their work, and you'll find quite a few don't pass the flip test, but still create beautiful art. So really, you don't have to worry about this. In this particular instance, though, I felt like bothering to flip and tweaking my lineart a bit. It's not hard to do, after all. I just lasso up the areas that need fixing and skew them a bit. Flipping also helps me notice a few details I might not have normally missed, like the pupils on the right (normally left) character were pretty badly misshapen. Also, the lower character's hair was uneven. A few quick tweaks, and I'm happy.



Flip it back, notice a few more little tweaks, and all looks good.




Step 3: Finishing the Lineart

So I've got the start of my lineart, but it's more like a refines sketch, because the characters are still missing something: clothes! So making yet another new layer for drawing, I draw on the clothes.

Obviously, you could include clothes from the rough sketch stage, but I like to work out the underlying framework of the body more before adding the clothes. It's a personal choice thing. In any event, I don't bother sketching out anything for the clothes. Having already finished off the characters, I don't really need to. I do turn the character lineart down to 50% opacity so I can see the different layers.

I also take the opportunity to do some more tweaking to the character lineart. XD



Now, since I'm about to erase portions of the character lineart, I re-save as a second psd. Really, I could have just duplicated the character lineart layer to preserve the original, but I like making separate files because I'm also going to delete my rough sketch layer in this new copy. Wave bye-bye to the rough sketch, among other things.



Now I place my lineart layers in one happy group, minimize it, and finally start on the coloring!


Step 4: Mapping Out the Color

The first thing I'm going to do now is map out all the giant areas of color on different layers. First the skin. Since the areas of skin on the character don't overlap, I can place all the skin on one layer.



When mapping, don't worry about areas that will be covered by other colors. Notice how my mapping of the skin is only exact where the skin goes over the background. In other areas, like under the hair and shirt, it's sloppy. Those sloppy areas will be covered by the exact mapping of the layers above them. I'm sorry this sounds so complex... it's really very simple. You actually could map out all the areas exactly, but not doing so saves a lot of time, really. Since the hair doesn't overlap, I put both the characters hair on the same layer, too. I also put the band on the arm things on the hair layer. This is going to end up being my top layer.



Now I place the left character's clothes and the right character's undershirt on the same layer, because none of those things overlap. This layer will go under the hair, but above the skin.



But Em, you ask, couldn't you put all of these things on separate layers? In short, yes, I very well could. But my computer is having memory issues, so by placing multiple things on the same layer, I save my computer a bit of hassle! And really, putting everything on separate layers is more complicated than what I'm doing, especially if you forget to label the layers. Now the only major layer I have left to do is the lower character's jacket, which touches everything and has to go on a layer of its own. Clean up a few edges and voila, my image is color-mapped.

On to Part Two!

Back to Tutorials