Capitol drawing


It’s been a long time after my last post. Now I am interested in drawing background image again.

This is the last time I made a cityscape background:
It took me total 230 hours, and I did now know much back then.

Now I am trying to practice a completely different workflow. I am following this tutorial:
and it is the kind of picture I am trying to achieve.

I will try to post each stage. Now I need some advice to keep me on track and remind me of what I am missing.


Hello Red and welcome back!

Sadly I can’t read any of the text of the tutorial but it looks nice. ;D
One question: Have you ever tried to practice something smaller? Like working on creating thumbnails in black and white first?

I am aware that this is a monochrome palette, so that probably qualifies as such. However, what I usually find much more useful is practicing on a small scale.
For instance making 5 or 6 thumbnails and then pick the best one.
You will notice that it usually takes you a while to get yourself warmed up and therefore, your thumbnails will generally improve after within one sketch session already. :slight_smile:

If you want to talk about this more or share some more sketches or progress shots, just let me know. :slight_smile:


What are your current goals regarding what you want to learn from the new workflow? What was it about your previous workflow that you weren’t happy with and wants to improve on?

The tutorial you’re following doesn’t show any perspective lines, but I assume in your workflow you made them while planning the scene? The clouds looks a bit odd because they seem to be following the way a wide-angle lens distorts, but that distortion doesn’t seem to match the scene very well. If you are basing this on a photo reference, can you please post it?

Do you plan on turning this into full color? If it’s meant to be monochromatic, then the sky wouldn’t be a different hue, so I’m not sure what you have planned for this.


The tutorial is much like “quick draft, more shades, more rooftops, more details, and keep it up use your patience”

Thank you for an advice. I’ve read much about how to paint, but I hardly know how to practice.

My old workflow was made entirely by a computer mouse and it is more like I patch things together (>1,000 layers). I think the way I can learn is limited when drawing this way. I just got my first tablet not long ago and I think I should work from overall image first and go into details, following the tutorial.

The work is painted over Google Earth references patched into a panorama image. It will be a background scene for characters, like that picture in the old workflow. I planned on turning it into full color, but you reminded me I don’t even have any good idea how to do it :frowning: I am just blindly diving in.


Jesus you painted all those thins on your DA with a mouse??
I bow to your patience!! (But I must admit it also took me a while to finally use a tablet when I started out. XD)

Is it a WACOM? :slight_smile:

I see that doing some exercises to get yourself used to working with the tablet is a good idea. Though for the next one I’d recommend you to look up some stock photographs or some old masters work. With the photographs it’ll be a little more like learning from life, and studies of old masters can be used to learn essential things about compoisition, colours and brushwork.

Have you also been working traditionally in the past? Or is it mainly digital?

Also your saying of patching things together with so many layers makes me curios. Is it like you finish one detail and then move and scale them around within the composition? Doesn’t this take a lot of time? :slight_smile:


Thank you. My tablet is Intuos 5, and I have little to no experience with traditional painting.

For that image, the background was separated, made one part at a time, and joined again in one final picture (the file size will kill my RAM if I don’t do that.) I use 5 PSD files for the characters, and another 5 for different levels of background. Because I could not use brush efficiently, most of the background objects are made with lasso, gradients, vector shapes, and smart layers.
I don’t know it’s just me or each part are not blending well together, but it is an obvious issue for my older works.

I am still not sure what I can learn from the style I am trying to achieve, but I will see what you suggest.


If your image is meant to be a full-blown scene with characters and background, then it’s very important to plan it that way, doing thumbnail sketches/color studies to work out all of your visual design problems first, as simplified versions of your finished image, with just basic shapes, values, and colors. The mistake that too many inexperienced artists make is to jump into a full sized image with lots of details without having done any planning at all, and just start to arbitrarily add stuff to it. That’s like shooting a movie without a screenplay, or performing a piece of music without having composed it first. Serendipity might give you a few pleasant surprises here and there, but overwhelmingly, lack of a plan will have dire negative consequences in your work.

1,000 layers is absolutely overkill and unnecessary, taxing your computer’s resources, slowing down the workflow, and a nightmare to manage. There are far more efficient workflows and management methods–ones used by experienced professional artists. The most important thing you want to keep in mind when managing your layers, is that anything that could be repainted fairly fast and don’t need to be isolated for random, sudden changes in value/color/size, do not need to be meticulously kept on separate layers. What you need to keep on separate layers, are the shapes and objects and characters that might require random and sudden changes as you move them around, resize them, change their values/colors, etc, as well as elements that would be too much work to separate from other elements via lasso/magnet selection (for example, hair, with all those detailed strands). At the basic level, you just need to keep your characters separated on layers from the background, as well as any background shapes overlapping other background shapes (such as buildings that are closer or further back at different Z-depth planes).

Doing monochromatic versions first and then colorize later has its advantages, but it also has its drawbacks. For example, any kind of minute detail you do, you’ll have to carefully redo them later on the color layer to match, and that’s unnecessarily doing twice the work. If you do effective planning in your thumbnail sketch/color study stage, you’d have figure out all those problems early on and don’t need to do a monochromatic version first. Also, even if you work full color right away, you can still add colorize layers later and completely change the colors if you want to, so you’re not missing out on anything by going full color right away. Some people work with monochromatic versions first because they don’t want to be distracted by colors, but I think that’s just avoiding learning to work in color properly, and one of the things I really train my students to learn, is to develop the ability to not be fooled by the deceptive values of colors, so they can make color judgments effectively and not be afraid of it. I have very challenging exercises I have them do to force their brains to split all color information from value information and deal with both simultaneously, and still nail the correct chroma they are aiming for. This is extremely difficult for some students, but with my guidance, they almost always end up getting it right.

My suggestion for this piece, is that do hold off on going any further and do some thumbnail sketches and color studies first, of the full scene as you see in your head, with the characters and the background, and work out all the visual design issues like composition, value management, colors, lighting, etc. They can be quite simple–something like the color keys we see in the “art of XXX” books from Pixar. If you don’t know what they look like, just google for “pixar color keys” That’s pretty much the golden standard for clearly readable ye simple thumbnail planning that works out all the visual design problems, without going into any details.


Is it literally 1,000 of layers or is this just a figure to illustrate that you are working with crazily many layers? I agree with Lunatique here that whatever number it is, it is probably too high. You also lose track of so many layers easily and it slows you down to search for something within the process.

When teaching students at our academy, I noticed that especially students who were very insecure created too many layers and were much too slow and too distracted with their insecurity from actually proceeding with their work.

To prevent this I gave them the “4 layers rule”. You can paint whatever you want as long as you keep it on 4 layers. For 1 figure with 1 background that would mean you have 1 layer reserved for the background, 1 for the figure and 2 for experimenting.

Once a student is happy with what he has tried out on his extra layer (for instance for the figure) we use apple+e or ctrl+e to merge this layer with the layer underneath. This way you can work with new layers without having hundrets of them in the very end. You create 1 new one once you experiment and merge it down once you’re done. Takes a little bit of training but will help you lots with decision making and becoming faster.

What can also help here is to think of it in a less technical way.
You can always paint small things again. It is in many cases actually much faster to see it like that (a more “traditional” approach like on a canvas) than having every tiny thing on an extra layer and never being sure if it actually “belongs” there.

Here again thumbnailing and brainstorming is a perfect way to get yourself making a step forward. Too many students are afraid that their work could look ugly so I had them do blind drawings to relax. (Blind drawing means you look at a friend or a model and draw them without looking on the paper. It helps you to train your observation skills, your feel for paper and pen, and it reduces your fear of “uglyness”) Here is an example of what blind drawing looks like, they had the most fun ever ;D

I would also say: Things need to look ugly in the beginning. And they are allowed to since we are learning. They actually will turn out pretty and advanced much faster when you dare to be ugly in the beginning than the other way around. ^^

And finally, to round things up, I would also suggest you to go back to small things in the beginning and practice thumgnailing etc. It will help you learn much better because you won’t be ovewhelmed by wanting to do everything at once, plus you can easily train to paint on a small number of layers this way. :slight_smile:


I just had a habit of keeping layers for some reasons. I also relied a lot on vectors and smart layers. I just thought I cannot follow a traditional way if I still use a mouse.

Here is what I can do with the thumbnail:


I choose from 6 variations of the poses. I still think I have a bad sense of composition and movements. I read so many about these but I can't just do it myself.


You have some composition issues in the thumbnail. The visual weight is too much at the top, with the larger character as well as that building’s dome–both are the larges shapes in the image. You might want to try swapping the position of the two characters so the visual weight is a bit more balanced throughout the composition.

Also, the width of the two character’s overall shape is too similar (considering the wing-like thing and whatever the top character is holding). You never want to have unintentional mirroring of sizes or shapes or orientation like that, because it locks the visual flow of your composition. Variation is the spice of life, as they say. :slight_smile:


Oh thank you very much. I should have done this thumbnail thing a long time ago. I can see the balance is obviously off when I left it long enough.

Any suggestions on the coloring? Japanese landscape illustrations tend to have this vivid sky-blue lighting and shadow and I think I still have something to learn.


Values are actually much more important than colors. The general rule is this–if your values are messed up, no amount of great color palette will save it from looking like a chaotic mess, but if your values are coherent and well-managed, even weak color choices won’t diminish the clarity of the image.

What you want to do, is make sure that the main shapes all read clearly, and the values are managed well, so contour shapes aren’t blending together because the values are too similar. If the local values/colors can’t be changed (established design for an IP, for example), then you can use lighting to help separate the foreground shapes from the background one (for example, rim lighting is a nice trick to use).

Ideally, you want to contrast color temperatures so that warm is contrasted against cool and vice versa, but there are situations where you want a stronger sense of harmony–it really depends on where you want to draw the eyes of your viewers. How you control the values and colors will create selective focal points where you want your viewers to pay the most attention (using color contrast).

There’s a lot of consideration that goes into the basic visual design of an image, and the better you plan it from the early stages, the more you can avoid huge mistakes that you’ll have to completely repaint later.


I tried these variations following your suggestion:

I choose the last one. I don’t know why I am not satisfied though. Maybe it is hard to work this out without changing the background but I just want to save the time.


The first one looks the best to me. The last one is actually one of the worst. I’ll explain why.

When you have characters and objects in a scene, you want to have their contour silhouette read as clearly as possible, without having them blend into the background due to similarities in values, colors, shapes, sizes, density/clarity of details, etc. If you look at the last one, see how the two figures just kind of blend into the background, but in the first one, they do not blend into the background because they are overlapping the sky more than the cityscape, and this makes their silhouette contour read a lot more clearly.


Thank you for the reply. Now I just have to remind myself of those principles every time I start a painting.

I suppose I can continue with the drawing. But what is the best approach with character and background? I heard we should start from background first so we can match the characters with it, since my characters are cartoons made in digital way.


Ideally, you should work on background and foreground at the same time, in passes. So for example, you first block in the basic shapes with local values/colors for the entire image (foreground and background–everything), and once you have the main shapes established, you can do the next pass, which will be establishing the general forms according to the lighting setup (keeping in mind the light source’s color cast, as well as ambient light sources, radiosity/color bleed, etc). Then you do the next pass and add the more intricate details, and so on. This is the best way because every step of the way, you are considering the entire image’s composition instead of working on only an isolated spot without considering how it interacts with the rest of the scene.


More detailed one with characters drafts

The colors are different from the thumbnail. I think I will have to rely on this forum for a long while until I can spot the problems myself.
Right now my guess is there is something wrong with the bottom-right character’s shape or pose, and the dark shadows of the background buildings.


You have some problems with tangents (common problem in composition for beginner/intermediate artists). The left character’s sword is “resting” on the top edge of the dome building’s first circular level. Her left foot is also “stepping” on a distant building’s top corner. These accidental tangents show up very easily if you don’t pay attention and try to eradicate them as much as possible. If you don’t, the readability of your contour silhouettes and scene’s Z-depth hierarchy gets weakened.

The right character’s legs look like they are retracted because she’s trying to avoid the bottom border of the image, or that she’s about to land on it. Try angling her more so he legs are foreshortened behind her and not about to rest on the bottom border of the image.

The right wing’s bottom corner and the two little wings on her hat have tangent problems too. Try moving them away from “resting” on the distant buildings and roads’s contour edges. Generally, you want your foreground shapes to either clear the background details entirely, or you try and make sure any overlaps are only at spots where the contour edges of both foreground and background elements aren’t touching, but clearly overlapping–this helps create a stronger sense of dimensional depth and better readability.

If you are serious about learning/improving in an effective manner so that you can excel in your artistic development far faster than you are now, you might want to consider taking my workshop (linked below in my signature. It just started today, but you can still get in if you want to join us). It is designed as an intense boot camp for artists who are tired of not seeing much improvement in their work even though they seem to be working hard at it, or feeling like they’ve been stuck in their artistic development and improving too slowly. Many students have said that 8 weeks with me taught them far more than they ever learned in all the years they went to art school (you can read the numerous student testimonials and see how the workshop had a profound effect on their artistic development).


I have read further about tangents, but this is some changes I did with what I understood from your comments.

I am really going to join the workshop, but my financial situations are not good right now, thanks to a career mistake. Maybe next time or so, at least not before I am clear of some problems. Do not write me off yet :slight_smile:


The update looks much better. You’re learning fast–I’m looking forward to having you as a student.

The bottom character’s feet now has a different problem. See how they are right at the separation line between the two different values in the background? Try having one of her legs drop lower (maybe her left leg), and then have her right leg drop just a little bit so her foot is overlapping that line more obviously.