This is the post I will be updating as I’m working through Scott Robertson & Thomas Bertling’s How to Render book and doing the 13 day challenge. I’ll be letting you know what I’m doing, why I’m doing it and to what outcome I expect to have after the 13 days. I’ll also be providing all the custom exercises I’ve made – assignments and solutions to check your accuracy. Feel free to change light & shadow direction for the shadow constructions, the solution images are just one of millions of possibilities. They’re also provided so you can do some detective work and figure out how the forms actually relate to each other, it’s not always clear if forms are parallel or the same dimensions. I thought that making some exercises deliberately ambiguous would test your perspective and logic skills… it certainly has mine.
Day 1 was spent primarily going over the whole book, figuring out the structure of the content, figuring what I know and what I don’t know and prioritizing how much time to spend on each subject. I’ll most likely have at least 2 days of shadow construction, 5 days of reflectivity (construction, Fresnel, chrome, etc.), 2 days of rendering and the other 4 days would be just padding, if I need an extra day here or there for something I don’t quite get or just going over material already covered, making sure I understand it.
Rendering, for our purposes, is simulating the effects of light interacting with a surface. I emphasise simulating since we’re learning how to calculate and figure out what would happen in a scene that does not exist. We do use reference, we constantly observe the world around us for more information, but when looking at an object, we understand on deeper level the interactions of light and surface that present us with the final, visible result.
The book, to simplify the organization of what I’ll mainly be looking into, consists of Light Sources + Cast Shadow construction, Matte Objects + Value Assignment, Reflective Objects. The other chapters give specific examples of materials, more complex objects, etc, but the main Key Points would be the ones listed previously. Those would be the fundamentals that everything else is built on.
There are a couple of variables to rendering that you need to be clear about, these are always present in absolutely any scene, we can recognize them easily enough, but when we invent our own, I think we very often deviate from the check-list because these things seem so simple and are taken for granted:
1. The quality of light – Hard or Soft, Local or Sunlight (This has profound implications for the shadows in your scenes and how your forms will read)
2. A light either reaches a form and illuminates it, or is blocked from reaching it, resulting in a shadow (don’t forget your cast shadows)
3. Forms can reflect, absorb or transmit light.
4. A Matte form, when illuminated, will show itself
5. A Reflective form, will show the environment
In this post we won’t be looking at refraction, caustics and sub-surface scattering so no light transmission.
Numbers 4 & 5 might need a bit of clarification, I try to think of it this way – to simplify the constant dilemma of “What the @#$! am I looking at?” A matte form will always show itself when illuminated. If it’s a blue ball, when illuminated with a white light, it will show it’s local value of blue, it will have a core shadow, you will not be able to tell what environment it’s in, but you will be able to get a lot of information about the ball. In other words matte surfaces only tell you about themselves, they don’t care about other things. Just remember it as matte objects being super selfish and self-absorbed.
Reflective objects on the other hand will tell you very little about themselves and a lot about the environment. You will be able to infer about their character based on how they display the environment – a chrome ball will warp and skew everything around it, but the environment will be bent to the surface of the ball, so you’d be able to recognize the underlying geometry. If it’s chrome it will have no color of its own, it will only darken the reflection slightly, if it’s colored, then everything will be tinted. Think of reflective surfaces as gossipers. They’ll talk about eeeeeeeveryone else, show you eeeeeeeverything around them, but tint and skew everything with their own little geometry and agenda. They’ll also be showing you all kinds of super weird and crazy reflections of others that seem mind bending and impossible – beware the lies of the little, curvy, reflective surfaces… Just look at your head in a spoon and you’ll know that spoon’s a frickin liar!
So – to summarize – Matte forms will only tell you about themselves, they’ll show you their color, you might also be able to infer some information about the environment through detective work, but that will not be explicitly communicated to you. Matte objects only show themselves. Reflective forms will show you everybody and their mama, every single detail about the environment will be explicitly given to you… even if you don’t want it. Light sources will be reflected perfectly, all surrounding objects as well. If it’s a curved surface expect some crazy distortions. Straight planes will reflect perfectly. Straight planes tell it straight, curved objects are liars and will bend and lie about everything around them.
Having said all that – bear in mind that reality is never as straight forward as a 3D renderer, objects will most often blend multiple characteristics in a single object… just to make art that much more difficult. Below is an example of a matte & reflective spheres. Going from left to right the spheres are 100% matte, then incrementally increase in reflectivity by 30%. The sphere far right is 100% reflective, the one in the centre is perfect chrome.
This is the first major element of construction in the book. This is probably because regardless of surface quality all objects will block a certain amount of light and thus cast shadows. So it makes sense to start with something that must be considered throughout the whole rest of the book and will be unchangeable regardless of all other variables introduced. Casting shadows is also relatively more simple to the constructions following… especially the curved surface reflections.
The principle is simple enough, cast shadows are basically projections of the form onto the ground plane. I often think of just stepping onto something very rigid and flattening it onto the ground – that’s pretty much what happens. At first the constructions will probably mess with your head a bit, but you will soon begin to see patterns. A curved surface will almost always have a curved shadow, sharp edges and straight lines will be straight on the ground plane as well, the actual flattening is what causes the confusion – which corner flattens over what and which parts of the form will be present in the shadow construction. As with all of these – don’t get too obsessed, although try your best to figure out what’s going on and do it as perfectly as you can. Ultimately though you will probably never be doing this for any other purpose rather than just being able to estimate things correctly. No one would ever think to check your shadows… not even your mom. Below are some examples of cast shadow constructions and also some files for you guys to practice with. One file for the exerciese, one for the solution to check yourself against. You can grab the light direction and shadow direction from the finished image or just use forms and light them as you wish. Often times you might not be able to tell how forms relate to each other, the solution images will give you more information about that. The detective work is up to you to do 😉
The rule seems to be that curved forms are always more complex than planar surfaces and cast shadows are no exception. The reason for the complexity is that normally, to properly perform a construction, you need to put the curved form inside of a planar one and then perform all calculations. What does that mean? It means that if you have a horizontal cylinder positioned on the ground, to properly predict or transfer a point, you’ll need to put that cylinder inside of a prism or at least construct a few straight planes around it and then transfer points from one to the other. I’ll post a few examples below.
Shading curved forms is also more difficult in terms of technique. With cubes you can just assign a value to each side and apart from a little gradation from the reflected light or light falloff from a local light source – there are no other gradations to worry about. Curved forms on the other hand are made up entirely of gradations. Shading a sphere can be particularly challenging as it is nothing but a continual, smooth gradation. To properly shade cylinders, cones and spheres will take some practice, not only understanding why and where all shading elements go – such as the passive highlight, terminator & core shadow, but also executing the value transitions smoothly enough that you don’t create unintentional form. The first few times I shaded a cone it looked like a party hat that someone had stepped on, bumps, lumps and dimples everywhere. Don’t be discouraged. As with everything else you’re learning – complexity is only being layered and added, things don’t get simpler – you get better, so just do a few of these each day and within a few weeks you’ll be a master.
Predicting the core shadow and highlights properly, involves cutting a section through the object. Ouch… To figure out how an object casts a shadow over a curved form you need to cut through it again. There is no going around that. Plotting sections of a curved form is often times like a chess game and can take as long… It is a very logical process though and with some practice can become routine. Develop a system as to how you approach these and don’t change it very often. Your brain will automate your actions and make it easier for you to perform in the future. Ellipses are a must. If you’re unsure about plotting circles in perspective you will have to practice some of that as well. Knowing how the minor axis relates to the ellipse is crucial, you’ll be using that knowledge very often, especially with spheres. Also understanding the degree of an ellipse and how that relates to your point of view is very important. It might sound like a lot, but if you spend the time doing it, instead of thinking about it – you’ll be there in no time. Below is a free video from Scott Robertson on drawing ellipses.
Some images below to show some construction and shading, you can download the first images and practice doing these yourself. Try and change the light source and its direction, if you feel like there’s something you don’t quite get – keep focusing on that problem, you can only iron it out by practice and reflection. There is no quick route.
Before moving on to reflective surfaces, I recommend doing some self-testing. Complex curved forms are the best measure of your understanding of the principles studied so far. Reading is not enough, shading a few spheres and cones is better, but still not enough. When you’re trying to judge whether you’ve mastered a particular item of study, keep this quote from Confucius in mind “I hear and I forget. I see and I remember. I do and I understand.” Your brain will not create deep level encoding for something that you did not spend enough time and resources on, which would mean that you might retain the knowledge for a few days, but don’t practice it enough and it will be gone within a week or two. When studying new information you need to put it into practice and you also need to explore the relevant contexts that you’ll be using that information in. You need to see the implications and problems posed by every single element in what you’re exploring. Rendering and shading are useless in theory only. Shading cones and spheres is theory only… Life is more complex than that. Cars, products, figures, nature – chances are that anything you look at in your surroundings will be a complex compound of all basic shapes, so if you’ve done your shadow constructions and your basic 3D figure renderings – it’s time to move on to complex objects and shade some of those.
You can create your own objects by simply drawing them or using basic 3D. If you want some forms to practice on – do a google image search for Scott Robertson, you’ll get many, many awesome line sketches that you can try and render. You could also google line drawing or line drawing + specific object. Another source would be Scott Robertson’s YouTube channel – https://www.youtube.com/user/scottrobertsondesign Just do a screen grab of something he’s about to render, give it a shot and then see his process. The examples I’m posting below are all from the book, so I can’t share the drawings since they’re copyrighted, I can only show my renderings.
You can see that for some objects I’ve done the same setup multiple times. Try and render out the same form, but lit in different ways or being a different value. I used to have a lot of problems rendering objects in different ways. I couldn’t understand the geometry well enough to manipulate variables and not being able to change the lighting might mean that you have practiced one particular setup so much that you’ve created a habit for it and by default that’s your go-to strategy. I normally get very rigid in my practice and end up doing one thing many, many times. When I was doing cast shadows they would almost always go in the same direction. Any time I’d notice that and change it – it would take me much longer to do the study since I’d have to fight that original habit of default direction. Make your practice as varying and as rich as possible, explore multiple options and exhaust possibilities. Do quick sketches to try lighting angles, use multiple light sources, change values, change materials.
Final chapter of this mammoth post and by far the most mind-bending aspect of the whole book – reflections. Reflective surfaces are pretty complex, also very interesting. The two almost always go hand in hand. Don’t be discouraged about the amount of time it might take to master – remember you’re in this for the long haul, even if you’re desperate to improve as fast as possible, get the principles, then just spend time observing the world around you. Your ability will increase with time and you don’t have to spend the next year obsessing over constructions. You don’t have to know everything perfectly when you study it, there will always be more layers of complexity.
So having said that – let’s see how we can simplify the physics of reflective objects. The surface of an object determines its rate of reflectivity. The smoother the object – the more it reflects its environment. Light rays bouncing off of its surface get reflected in more precise angles, giving you a good idea of the environment surrounding it. A matte object has millions of very tiny irregularities across its surface, so when light hits the object it gets scattered in many different angles. A reflective object is just like a mirror, light rays bounce together to give you an exact replica of the environment. Think of light as a source of information. Each ray is carrying a fragment of the environment it bounces off with it. If it strikes an object with a lot of small irregularities on it – all the information gets scattered in millions of direction – consequently you can’t see the source of the information. The mirror would reflect all light coming to it – back to you, preserving the configuration of the information. This might sound a bit confusing, but spend some time playing with a mirror and looking at objects with varying degrees of reflectivity and the point should become clear. The illustration below should also hopefully help a bit with understanding what’s happening. It’s a top view of an environment with you standing directly in front of a mirror, 2 cubes are also there, because cubes are cool. The sight line of your reflection bounces right back into your eye. The 2 sight lines going to the cube to find the extent of its reflection are equal angles going out of your eye and travelling to the cube. In top view you can just make an equilateral triangle and scale its width to find points in the environment.
Predicting reflections is all about understanding light bounces. These are basically the physics of a mirror, if you’re curious – check this link out you’ll understand concave and convex surfaces a lot better. Or if you’re even more curious you should check out the physics of optics, which is the main branch of physics all this comes from. Sight lines are imaginary straight lines going out of your eye, bouncing off of a reflective surface and hitting an object in the environment. This is how you get to figure out where, in a reflection an object will be. The golden rule here is that the angle of incidence is equal to the angle of reflection. Said more simply – the angle at which the line travels from your eye is the same as the angle it reflects at. More examples and images for you to practice on below.
From the illustration above we get a pretty good idea about the fact that reflections are predictable, though with more complex objects it’s often too laborious to plot everything. It is possible, but often unnecessary. Knowing some general facts is good enough, at least for a start. Convex objects, like a cylinder, will normally compress reflections, convex objects will sometimes turn things upside down and could also stretch them.
The examples above have been mostly related to chrome or perfect mirrors. But wait… there’s more! In fact the majority of reflective objects live here. There are surfaces with reflective properties, that are a blend of matte surface shading and reflectivity. They exhibit characteristics of both, such as core shadows and passive highlights, along with environment reflections. Cool…
The main key factor with these surfaces is the Fresnel effect. It basically states that the more perpendicular a surface is to your line of sight – the less reflective it is. Meaning that the more the sight lines from your eye bounce right back at you – the less reflective the object, the more they bounce into the environment – the more reflective the object becomes. Said in a different way – the more oblique you become to the object ( the more you travel away from being 90 degrees to the object), the more reflective it is. Get a piece of reflective plastic and look at it straight on – little reflectivity. Tilt it away from your eye – it becomes very reflective. Below is a digram and a real-world example of the same setup.
The Fresnel Effect can be seen anywhere in nature or in man made objects. The logic behind it is that every form has slight imperfections to its form, looking directly at the surface, while standing right in front of it, some light rays get dispersed, hence you don’t get the mirror reflection expected from something like chrome. Stand oblique to the surface though and more points on the surface align together to bounce light rays in a similar direction and give you a more reflective form overall. The best way to understand it is to just look at more examples.
The easiest way to think about the Fresnel Effect is – the more the surface faces away from you – the more reflective it gets. In situations where there’s nothing to reflect, but a cloudy sky the surfaces get brighter, but have a look at the water photo or the red car above and you’ll see that it could also mean a hue shift. The surface will literally become more like a mirror, so it’s not just brightness that changes, the surfaces will tend to be more like chrome, reflecting whatever the environment happens to be.
There’s a lot of material to take in here, so what’s the conclusion to all this? How can we think about all we’ve learned and what should we expect now that we have this information. First thing is that having the information about something does not mean you know it. I make this mistake constantly. Just because you understand the rules of perspective does not mean you’ve mastered perspective. Constant observation & testing is what makes you proficient at something. The rules are all very simple, but their implications are millions. It’s like they say about certain games – 5 minutes to learn, a lifetime to master.
Don’t despair if you can’t get how sight lines bounce, can’t think of objects in different views or have no idea how the environment could influence an object. This is all learned knowledge, no one is born knowing it. Arm yourself with the information above, remind yourself of it every few weeks until it’s part of your long term memory and then just go outside. Observe the world. Stage some objects in your house, get yourself a flashlight. Everywhere you go – look for surfaces, make mental notes of everything that seems odd or you were not expecting, that’s a good indication that you’re learning about how the world actually works as opposed to how you thought it does. There will be a lot of “What the F$#!@#!#@!#!” moments as you’re adding more and more complexity layers to your understanding. Don’t fight those, as frustrating as they are, they are your best friends, they will always point to what you need to explore next and are always landmarks of your learning simply because you can now notice them, but couldn’t before, it’s like more of the world has been revealed to you.
I can only cover so much in this post, the book has a huge amount of content, videos, examples and a lot more theory. This post can hopefully give you an overview of the contents or be a good place to explore a different way of looking at things if you’re having difficulty understanding something. Keep learning and keep growing.
Some more examples below of reflections, always an interesting thing to look at. And also a ball having the same texture as the ground, vs an actual chrome ball reflecting the environment. Notice that while the colors and texture are the same, indicating it could be a mirror surface, the distortion of the environment is what’s necessary for the brain to interpret a material as reflective. We normally avoid anything to do with distortion, whether in perspective or photography, etc, it’s a word that normally denotes something to be avoided. When you’re trying to learn how to replicate the actual world – there’s plenty of distortion all around.