Live Perspective: Draw Animals in 3D Space

What You'll Be Creating

If you treat your artistic development seriously, I'm sure you've tried to learn perspective and to apply it to everything you draw. However, even if your animals in perspective looked 3D, they also became as stiff as stone statues.

Why? Because if you wanted to use only one perspective grid, all the lines of the picture were bound to it, which means they turned out having all the same rhythm. Yes, you could add another perspective grid, for example for a turned head, but can you imagine all those converging lines? And what about the neck: wouldn't it have some kind of "transition perspective" between those two?

All these problems come from one source. Perspective as you know it, called linear perspective, was created to simplify the visual phenomena occurring when observing a big area, e.g. a city or a landscape, or maybe a herd of bison. Never a single bison! Why? It's simply too small to be affected by converging lines. It doesn't mean it's not affected by perspective at all—but unfortunately, the tutorials you might have read focus completely on large-scale scenes, as if small scale had nothing to do with perspective.

Of course, perspective is inevitable in realistic drawing. At the same time very few tutorials teach how to use it without sketching a grid at the very beginning of the drawing process. That's why usually artists need to work it out themselves, as I did, but this time I'll try to show you the way.

This tutorial is a practical extension of this article, and I strongly suggest reading it thoroughly before trying this one. Also, keep in mind this tutorial isn't "how to draw animals in perspective" or "how to draw animals in perspective, but rather "how to draw animals in perspective. It's directed to people who already know how to draw animals, but struggle with creating more interesting poses for them.

Still Perspective vs. Organic Perspective

Converging lines appear when objects follow the same rhythm. They don't occur in nature on a large scale too often, but humans love order and arrange their creations this way. This is the first reason why perspective seems to be necessary only to draw man-made space.

Some pictures simply shout "perspective!"...

...while others don't

Linear perspective is all about right angles. They're predictable and neat, and can easily be organized with straight lines. Unfortunately, nature doesn't know anything about right angles. It creates them spontaneously (usually on a very small scale), but it doesn't favor them in any way, as architects would. We could say that chaos is far more natural, but how can chaos be constrained with perspective rules?

There's no such thing as organized chaos

Even if animals had a silhouette based on right angles, in most cases they wouldn't follow linear perspective rules. They're just too small! Let's take a closer look at this "horse". Its lines converge so far from us that we can safely say they don't. Where's the vanishing point then? Where's the horizon?

That's how it's going to look in most cases. Even if you manage to find straight lines in a silhouette, they won't converge and won't show you the horizon. And since the horizon is the first thing you should draw in classic linear perspective, how can you even start without it?

Perspective Without Vanishing Points

On a small scale, the scale on which humans observe animals, we don't need vanishing points, but the horizon is still crucial. This is what a human eye sees when it comes to a small area (no converging lines visible). Notice two horizons—horizontal and vertical. They're actually the same, but we favor the horizontal one because we rarely move up and down, and our eyes are placed horizontally, too.

The point in the center is the one you're looking at. When you look up, your horizontal horizon slides up along the vertical one. You can't look anywhere but in the center of your field of view (FOV). However, for simplicity's sake we can ignore the motion of the eyeballs and treat only the motion of the head as a change of perspective. Both horizons are simply lines crossing the center, and they have nothing to do with Earth's horizon.

A cube looks 2D only when it's in the center. Every motion away from the horizon (no matter which) results in showing us another side of the cube, at the expense of accuracy. A cube moving away from the center seems to be rotating:

• If it's moving up, it gradually shows you more and more of its bottom, until it becomes a square and the front becomes a line.
• If it's moving down, it gradually shows you more and more of its top, until it becomes a square and the front becomes a line.
• If it's moving left, it gradually shows you more and more of its right side, until it becomes a square and the left side becomes a line.
• If it's moving right, it gradually shows you more and more of its left side, until it becomes a square and the right side becomes a line.

These movements can be combined, e.g. a cube can move up and left, giving you three visible sides (bottom, right, and front). What's interesting, a cube rotating in the center will look exactly the same!

This example shows how a cube changes when going up, right, or combined

How to Draw a CubeAlthough the picture above makes a good reference, you don't really need it. There are simple rules for drawing a cube in every perspective, and they can be used for drawing other forms, too. The most important lesson from it is: how to draw something in perspective without setting it at all. Because we don't want perspective—we want what perspective gives us!

Basic Rules

You can show one, two, or three sides at a time. It isn't really like "the more sides, the better", but rather "the more sides, the more distortion".

1. Only one side will give you perfect 2D view. It's the easiest to draw; no space for mistakes here. However, it's also flat and boring.
2. If you want to show another side, you'll need to compromise. The front can't be a square any more, if you want to squish something next to it. The more complete one side is, the more squished the other one.
3. If you want to include the third side, it'll push the angles, too. There is no way to preserve right angles with three sides visible (unless you get rid of the depth by dividing them).

How to control this chaos? You only need to remember that the rotation leads to a point where the current front becomes a side, and a side becomes new front. In order to reach this state, the current front (A) needs to get smaller as the side (B) gets bigger.

Basically, every cube is made out of two squares with a distance between them. When you connect their corners, all the sides appear by themselves. However, our vision doesn't work that simply:

 

Let's take a look at this scheme once again, trying to find the connected sides. There they are!

But hey, there seem to be four of them in combined views! How does it work? How can we foresee where they should be to make an accurate cube?

Step 1

Start with a cube in the center view ("no perspective"). This is going to be a base for the front, no matter how visible it'll be.

Step 2

The front needs to be changed if other sides are to be introduced. It has two lengths: width (horizontal length) and height (vertical length). If you want to add a side that's placed vertically, you need to shorten the vertical length of the front:

• A—a bit of the bottom will be visible, so I shortened the front slightly.
• B—I want the bottom and the front to look the same big, so I made the front half smaller.
• C—I want a bit of the top to be visible, so I shortened the front a bit.

Step 3

We need to do the same with the horizontal sides and horizontal lengths:

• A—I shortened the front a bit to reveal part of the left side.
• B—I shortened the front a bit to reveal part of the right side.
• C—I shortened the front a bit to reveal part of the left side.

Step 4

The front is done, and since we need two sides and a distance to make a cube, let's add another side. The one we've drawn before will now be the back. Draw it once again higher (to reveal the bottom) or lower (to reveal the top). The rule to evaluate the distance is:

The shorter the vertical length, the bigger the shift and the more of the upper/lower side will be visible.

Step 5

That would be enough if we wanted to rotate the cube only vertically, but we want more. Let's move the front again, to the side we don't want to see:

The shorter the horizontal length, the bigger the shift and the more of the left/right side will be visible.

Step 6

To make a base for the angles, we need to copy both front and back, and move it to the same side as before. The distance between the original and the copy is based on this rule:

The bigger the horizontal and vertical lengths, the bigger the distance.

Step 7

A cube has only eight corners, but our sides all together have sixteen! That's way too many. To pick the correct ones, first focus on the top/bottom side (depending on what's not visible in our perspective). Select the corners of the "original" sides.

Step 8

Now move to the border of the side we will see (top or bottom) and select the corners of the copied sides.

Step 9

Now simply connect the dots:

• Copied corners will make the visible top/bottom. 
• Original corners will make the hidden top/bottom.

The other sides should be easy to create out of them.

Why no vanishing points? Here's the answer. Converging lines look converging only when you look at them from a distance. If you crop a smaller area, the lines will look almost parallel. That's why linear perspective works so great with cities, big buildings, and man-planted forests. You'd need a whole herd of animals standing in perfect rows to see the converging lines. You don't need them for one animal seen by an observer of similar size. No vanishing points are necessary, until you decide to draw a huge monster (or a normal animal seen by a tiny observer).

In most cases the bending is so subtle that we can ignore it

This basic rule about visible sides may seem confusing, but I'm sure you've used it even unintentionally. Just put a simple animal in place of the cube:

This is perspective in action, though without right angles and vanishing points!

Cubes in Nature

However, cubes are very hard to find in animal bodies. The only useful application for them is a bounding box. It's usually a cuboid, with its longer side symbolizing the length of the body, and the shorter one the width of the body. It defines the perspective by showing the ground and the "ceiling". However, the perspective of the animal inscribed into the box doesn't need to define it—the box only shows us the ground as the observes sees it, and a default position of the animal.

If the animal shares its perspective fully with its bounding box, it will be literally bound by it

For elements of the body inside the box, it's better to use balls and their derivatives. Look what happens to the center lines of a ball when the perspective changes:

These lines are crucial to understand perspective and 3D overlapping. For animal (and human, too) drawing, they're as important as vanishing points in architectural drawing!

These bending lines add a lot of volume to forms and show the perspective in a clear way

Notice how these lines define the forms in 3D, showing us when and how they cover each other. Compare them to the balls to remember this rule.

They're also indicators of the bending of the surface. Notice that in the case of a cube, when the sides aren't bent, the lines only change length, but not shape. When we want to draw rounded forms, these lines come in handy. Even when we don't sketch them, they appear in the form of shading and texture (yes, a texture can't be flat on a rounded surface!).

Enough theory, let's see how to use it in practice!

Simplified Animal Skeleton

The pose of an animal is defined by the pose of its skeleton. This is why it's so important to learn about the bones of an animal, their construction and proportions. Fortunately, most of the animals you'll draw will share the same skeletal structure:

1. Skull
2. Neck
3. Torso
4. Hips
5. Arm with shoulder (A)
6. Forearm with elbow (B)
   
7. Hand with wrist (C) and fingers (D)
   
8. Thigh with knee (E)
9. Calf
   
10. Foot with heel (F) and toes (G)

The shape of individual bones, though helpful, doesn't need to be memorized. Look at the scheme below—it's simple, easy to remember, and you can still find all the structures described above!

However, this simplified skeleton is still too complicated and stiff to start a picture with it. If you want your lines to be truly free, simplify them even more!

This way you can draw every pose you imagine. However, only as long as you stay in two dimensions. It's a bit boring, isn't it? We perceive three dimensions every day, and we don't want to be limited to only two of them!

If you're having problems at this point, check these exercises before going any further. It'll spare you a lot of frustration! To understand movement, check my article about poses and the introduction to this tutorial about run cycle.

The bad news is, we can't draw in 3D yet. The good news is, we can pretend we do—and even so well that others will believe us. To do this we need to convert the simplified skeleton to a simulated 3D form.

From 2D to 3D

We're going to work on every part of our simplified skeleton to understand how to add depth to it.

Torso

This is the simplest, and also the most important part. The torso is the part that gives direction to the rest of the body. It's the front of everything. Even though the head seems to lead the body, it can easily be turned to a different direction while running—you can't do the same with the torso without changing the direction of the main movement.

The 2D base for the torso is an ellipse. An ellipse is symmetrical in both axes—no matter how and when you draw it, it can't be changed. It has two axes—a major axis (longer, A) and a minor axis (shorter, B)—that define its shape. The axes are always perpendicular to each other.

A rotated ellipse stays an ellipse

The easiest way to draw an ellipse is to sketch a cross of perpendicular axes, and finish them partially—first with straight lines and then with round ones between them.

A simplified torso we're going to use is made out of elliptical cross-sections. When shown in simple views like top (1), side (2), and front (3), they're either full ellipses or straight lines. To create an illusion of depth we'll need to break this rule.

The shape for the torso can be a barrel or a capsule. No matter what you choose, an ellipsoid will be the best base for it. Most of the time you'll want your torso to be oblong and rounded, just like it:

It doesn't look very easy to draw, does it? Fortunately, an ellipsoid will follow the rhythm of its bounding box—and we already know how to draw a bounding box in every perspective! There's only one problem with the perspective of the ellipses, so let's investigate it step by step.

Step 1

Let's start exactly as we'd do with a cube, but this time use a rectangle as a base from the center view. The procedure is exactly the same, except that the front of the torso is the left side of a cube. So, let's say we want the front (left side) and bottom visible. What can you do with a rectangle to achieve this?

Step 2

Draw diagonals on every pair of corresponding sides and connect them with a line.

Step 3

The axes show us the points where the torso touches the sides of the bounding box. That's a good start!

Step 4

These points are also the centers of every side. You can use them to draw rectangular cross-sections right in the middle of the cuboid. These will be bounding rectangles (looking like trapezes now) for our ellipses.

Step 5

This is where all the problems start. Take a closer look at one of the bounding rectangles.

If you try to draw a simple ellipse inside, it won't touch the the sides of the bounding box, which makes it useless. We can always draw a rotated ellipse, but where to put the axes? What angle do they need?

Step 6

To find out, draw the diagonals of the rectangle.

Step 7

Let's draw an ellipse, starting as we would normally. Sketch two short, straight lines at the ends of the "fake" minor axis.

Step 8

Now things go a bit differently. For an acute angle, draw a normal straight line, but when it comes closer to an obtuse angle, turn a bit to the inside.

Step 9

The other guide lines will look different, too. Draw them long for the area of an obtuse angle and short for an acute angle.

Step 10

You can now finish the ellipse. Draw the opposite arcs in pairs—they should be identical.

Step 11

Do the same with the other rectangles.

Make sure your ellipses are symmetrical, and if not, correct the arcs.

Step 12

You can now draw the outline of the main, outer ellipse. It doesn't need to be perfect, but make sure you enclose all the other ellipses inside.

Step 13

To make the depth more apparent, fade or erase the lines lying on the hidden sides.

It looks very complicated, I agree. The good news is you won't have to repeat all this process every time you want to draw an animal. If you practice long enough, you'll understand what a proper torso looks like, how it changes in perspective, and what your favorite views are. The more experience you get, the less you'll have to resort to drawing a cuboid for the base. But until then—practice!

Hips

Hips are very complicated structures, hard to picture in the classic front/top/back view. Fortunately, we don't need them as a whole in a drawing. The most important elements for us are:

1. Iliac crest—you can feel it in front of your hips, and even see it, if you're skinny.
2. Acetabulum—the "hole" for the thigh bone.
3. Sitting bone—the part that you basically sit on.

As you can see, our simplified animal hips are built of two flat cuboids and circular spots for leg bones. The longer, rotated cuboid may look a bit hard to draw in perspective, but fear not—there's a simple trick for it!

Step 1

Inscribe the side view into a bounding rectangle. Then treat it as a regular side of a cuboid. We want the left and top visible, so we shorten the side properly.

Step 2

Let's use only the bounding rectangle for a while. Build a bounding box out of it.

Step 3

Draw one part of the hips on both sides of the box, then connect them. Remember to follow the rhythm defined by the bounding box!

Step 4

"Cut" the opposite corners of the box with short lines, making an angle similar to the original.

Step 5

Connect the upper points with the upper points on the small box.

Step 6

Copy the lines you've just drawn to find the lower part of the box.

Step 7

Connect the lines to finish the box.

Step 8

You can now add the circles. Don't worry about their perspective—they're spots, so their shape isn't that important. Clean up the picture to define the visible and hidden lines.

Skull

The skull is the most complicated part of a simplified skeleton. Even though you don't need to draw teeth, the nose, or the complicated curvatures of its surface, there are a few elements that need to be included for the head to have a proper shape.

The method I'm going to show you works for almost every skull you can imagine. However, you'll need to understand what you're drawing in order to modify the elements. The most important thing to determine at the very beginning is whether the animal is a herbivore or a carnivore. Though their skulls can be very similar, there are some general differences:

• Herbivores usually have longer snouts, with big incisors.
• Carnivores usually have wider, stronger jaws, with big canines and small incisors.
• Herbivores generally have eyes on the sides, carnivores—on the front.
• Herbivores may have very big jawbones, designed for chewing, while carnivores can only move their jaws up and down.

No matter how strongly you simplify the skull, it will always be complicated if you want it to be accurate. What you need to define are:

1. Upper jaw—fused with the rest of the skull
2. Lower jaw—mobile, with big bony "hooks" attached to the upper part
3. Eye sockets—no matter how big the eyes are, keep the sockets wide and round
4. Brain case—usually quite small in comparison to the jaws
5. Zygomatic arch—the point where the "hooks" of lower jaw hangs on

Remember that when the lower jaw is open, you need to use a slightly different perspective for it, treating it as a separate part.

Step 1

Because our skull is made of cuboids, it shouldn't be too hard to draw them in perspective. However, it can be time consuming to draw them all separately. You can use this trick instead.
Find a bounding rectangle for all the views to see what kind of cuboid you'll need to use. Then modify the front as a whole, using the method for drawing a cube.

Step 2

Add guide lines to see where the elements should be placed inside the bounding box.

Step 3

Prepare a bounding box in the same size as we've just defined for the side.

Step 4

Draw the elements on the side, following the rhythm of the bounding box.

Step 5

Replicate the elements on the next side.

Step 6

Add other sides, if you need to define other distances too.

Step 7

Connect the sides, still following the rhythm.

Step 8

Add any necessary details. Now you should be ready to draw the skull or head based on the perspective.

Again, using the bounding box isn't obligatory, but it's very useful when you don't have too much experience with cuboids in perspective. Also, once you understand the elements of the skull, you'll be able to create your own method of simplifying them.

Spine

The skull, torso and hips are linked together by the spine. It's this element that makes animals so different from buildings or cars—thanks to the spine, the elements of the body can rotate very independently. That's why the whole body shouldn't be enclosed within one bounding box and its perspective. Accordingly, by using the power of the spine you can add a lot of depth and realism to your picture.
To understand the flexibility of the body, we need to learn the parts of the spine:

1. Cervical spine (neck)—for quadrupeds it starts in the back of the skull and ends in the upper part of the torso. It's S-shaped, which means it can be pretty flexible on its own.
2. Thoracic spine—not a very flexible part. It's good to treat the whole chest as one big block.
3. Lumbar spine—the most flexible part. The longer it is, the more elastic the body (compare a cat and a horse).
4. Sacrum—fused vertebrae make this part stiff. You can treat it as a part of the hips.
5. Caudal spine (tail)—made of very small vertebrae and therefore extremely flexible. However, notice where it starts—the sacrum isn't part of the tail!

No matter how long the neck, it takes a similar S-shape—stiff just by the skull and chest, and more flexible in between. The neck doesn't really move on its own—it's the head that directs it. When drawing it, decide how long the neck can maximally be, and then place the head somewhere in the area within its reach. Afterwards, add the neck.

The lumbar spine adds shape to the back. It's usually straight in the default position, but it can't be bent to make the back more concave or convex. The longer this part of the spine, the more extreme an angle the animal can reach.

Why aren't we saying anything about the perspective of the spine? Because for us it's just a line of no width. It symbolizes the connection between the elements in perspective, but it doesn't need to be changed itself. What changes is the perceived length of this connection, and that's where foreshortening comes to play.

When the elements in a line rotate at the same angle, they all get proportionally shorter (we've already learned how), but not only them—the distance between them gets shorter too! We can say the distance is longest in the side view, and then it gets gradually shorter, until the point where it's equal to zero. Keep in mind it's proportional change—e.g. when the hips and chest are half shorter, so is the spine between them. Don't make it shorter by 2 cm or some other value!

Legs

Just like the spine, the leg bones don't need to be managed in perspective other than by foreshortening. They are indicators of position, with no special width.

It's important to understand the relation between the legs and the other elements of the body. Both the forelegs (arms) and hind legs (legs) can move on their own, but they have their limits. When you want to move them farther, you also need to move the element they're attached to.

Paws

Although the legs themselves can be drawn as simple lines, paws require a different treatment. If you want to be fast, you can draw the fingers as simple cuboids and add details to them later. For slightly better accuracy, you can use the method below.

Step 1

As usual, use a bounding rectangle to create the bounding box.

Step 2

Divide the bounding box into sections just as on the template.

Step 3

Using the template as a reference, connect the points that will make a finger.

Do the same with all the fingers:

Step 4

Later you can add the details, like paw pads, claws, and fur. Keep in mind that these are only a finger—you need the rest of the hand (or foot) to make it complete.

Practical Exercise

Now I'm going to show you how all this information can be used in practice.

Step 1

Start with a basic sketch of your idea. It can be hard to plan depth when you're a total beginner to it, but starting with a perspective grid kills the composition and liveness of your animal, so it's worth trying.

To keep it free from formal mistakes, I decided to draw a made-up species, some kind of feline with its baby. This way we'll focus on creation, not on accurate re‑creation.

Step 2

Try to guess what kind of bounding box your sketch brings to mind and draw it.

Step 3

If we use the bounding box's sides, we don't need to draw them for the chest's bounding box. Just follow the rhythm, to create bounding rectangles for cross‑section ellipses. Then use the usual method to draw the chest.

Step 4

Draw the box for the hips. See how I simplified it?

Step 5

I repeated the steps for the body of the baby:

Step 6

Now, skulls. For the baby I borrowed the rhythm of the big bounding box, but the mother's has a totally new perspective that breaks the monotony of the composition.

Step 7

Add the connecting lines: spine and legs. You can be free here!

Step 8

You can now build the body on the guide lines.

Step 9

When it's all done, you can finally draw the details and refine the picture.

It's the End... Finally!

That was a long tutorial, wasn't it? I know for now it may look very complicated, but with practice it'll become second nature and you'll just feel how to use a certain perspective without all these guide lines and bounding boxes.

Practice a lot! Don't refine every interesting pose you've managed to draw just to show others how good you are. Draw for as long as is necessary to gain confidence—it must come naturally to you, without "unexpectedly good results".

Perspective is the hardest topic in drawing, but the methods I've shown you should help you at least get close to it. Finally, you can move from non-intuitive "where should I place the vanishing points?" to "what sides do I want to present?"

Of course, it doesn't mean vanishing points are totally useless. They're useful when you want to picture motion or scale, but in most cases you can forget what architects teach you—for your own good!
The best thing about drawing animals in perspective is that you can make a whole lot of mistakes without any risk that someone actually notices it. It doesn't work this way for architecture, and that's why perspective tutorials usually require perfection. Don't try to be perfect—learn how to delude people that they're looking at something 3D, and that is all you need!