This tutorial is all about learning the basic toolset of Blender 2.5x series, and the techniques we use to make renders more alive and believable. I won't provide you with step-by-step tutorials, since you won't focus on the real issues. I will show you the techniques and the correct mindset that you need to achieve before getting into serious work. There are lots of people with their own work style, and documentations on Blender 2.5x are piling up, as well as good tutorials, but that does not mean that we cant see things from another perspective, right? I am not a professional by any means, and that might be the key, since I am very familiar with the struggling issues that a newbie has to face. That's why I will try to give you as much basic information as I can. The Toxic Storm is a small piece of toy, or maquette I made for the purpose of this tutorial. I have used different shading, texturing, mapping and modelling methods to accomplish the final scene, and you will read about those in the following chapter. I will try to cover everything from alpha to omega, starting with the props and ending with the composition node set-up. You will find illustrations in the tutorial, but I have also made some video tutorials, so the more visual ones can learn as well. I hope that you will enjoy reading this untraditional collection of lectures, and learn a lot by doing so!

This chapter will talk about different lighting, shading and coloring techniques. I will try to explain the basics of a single lighting set-up, we will learn about using Normal maps and Specular maps, fake rim lights and fake ambient occlusion as well as real ambient occlusion and environment lighting.

Lighting can generate loads of different feels to a certain scene. Night and day lights have different moods, but if we have a well lit room with a warm fireplace in it, or the same room with demonic signs glowing in green, we get different stories out of them.
Much can be told by only using different techniques on a character as well. If you light the face evenly, using the standard three point lighting, you will get a warm, but general picture of the character. If you only use rim lights, the character becomes mysterious, and if you light the face from below, it immediately becomes intimidating.
The same goes for environments, as I wrote before. The amount of light and the hue of the color have their own meaning in a certain set. There is a good joke about this online. If you didn't play Limbo, then you don't know, but there is a certain spider that is chasing the main character in the game. Since the game is set in a very dark environment, you can only see the silhouettes, so you can only guess what lurks in the shadows. And someone did a joke about this, making the character aim a light source at the spider, revealing it to be having a big Hello Kitty head and pink skin, rendering the whole scene funny instead of scary.
I am going to use the industry standard three point lighting, with two Sun lamps and a Hemi lamp. This set-up was used in the final render as well. The tree point lighting set-up stands from a Key light, giving the most amount on the scene, a Fill light for making the objects evenly lit, and a Rim light to cut the objects from the background. The Key light has a mildly orange tone in this case, the Fill light a light blue and the the Rim light or Backlit has a very vivid and bright blue color to it.
This hot to cold gradient usually gives a good look on our rendered characters, and the color palette is this because usually the Sun gives a warm Key light, and the Fill is usually the colder color of the sky's reflection. The Key and the Rim light usually have a greater energy value to them.
This technique works well with Spot Lamps and Area Lamps as well, but I found the Sun Lamps and one Hemi as rim light just fine for this project. Feel free to experiment with the lamps, you will become more confident with time. Just remember, if you turn the samples way up, you will get longer render times. Spot Lamp's Buffer shadow has lots of cool uses, and it wont generate too much render time neither, so make sure to look into them as well.
My usual settings for the Spot Lamps: Buffer, Soft:8-12, Size:2000-3000, Samples:6-8, Autoclip Start&Stop enabled. Blend and size depends on the scene, everything else is default.

Slight warm to cool gradient was applied with a more visible Backlight, making Suzanne pop out.


Lighting set-up used fort the effect. The Key light is pointing down from above, the Fill sits in roughly eye-height, and the Back light faces the camera from behind Suzanne. Note that the position of the sunlamps does not matter, just the direction of them.

I have also used Ambient Occlusion for my final render. Ambient Occlusion gives the scene even lighting, generating shadows based on the distance between objects. Ambient Occlusion has two types: Raytrace and Approximate. Raytrace is more accurate physically, but it needs more samples and more time to generate a good looking and noiseless render. If you are aiming for quick renders that don't critically need accurate shadows and lights, feel free to use Approximate. Approximate will even allow you to use Indirect Lighting, which will let you use object as light sources. Even though this is not perfect yet, bouncing balls of light are fun to play with.

Gobo (Goes before optics) is a piece of plane with a pattern of holes in it, set between the light source and the scene, creating more interesting shadow and light. It is also called Cucoloris. I used this method to generate shadows that look like cloud shadows, and you cant really see it unless you see the difference between the two renders. It gave the scene a good feel, even though its not too visible.

Gobo used in top image, and not used in the one below.

Neither Gobo nor light texture were used on this scene. Note that the scene became way brighter and lacks shadow variations.

For those who are new to Gimp, this free, open source program can be downloaded from here:



Open the Color Texture in your preferred image manipulation application.


Desaturate it, using Luminosity in Gimp. That option gives the most satisfying results most of the time. This is the part where we are getting rid of the unnecessary color information.


Adjust Color Levels to the point where you have the Specular and dark parts separated by hard black and white colors. The white parts will be shiny on the finished result and the black parts won't affect the outcome at all.

Now, since we have a Specular map, lets use it! Add a new texture to your material, with the following settings: Image, Mapping as preferred, Diffuse: nothing enabled, Specular: Intensity: 1, Blend: Multiply, RGB to intensity, and use a Wardiso with a 0.160 slope. This should work in most cases, but if Wardiso wont bring the desired result, try to use different settings for the shader.
Second method is when we use the same color texture and modify it in Blender. For this, simply add a new texture, select the color texture, and modify the Brightness and Contrast values: Brightness: 0, Contrast: 5. All the other settings are the same for this one as before (turning saturation to 0 is optional).

There are also two methods for this. The simple one is when you just enable Normal under Geometry for the Color texture. The other one is similar to the technique used for creating Spec maps. The sad part is that we cant do this with simply using Gimp. We need something more for generating a Normal map.
Good news is that there is a solution, in a form of a totally free web application! This little program is called SmartNormal, and it can be used as a desktop software as well as online:

Using the color information from the texture for creating displacements.

If you are interested in a more throughout article about Normal maps and Specular maps, you can check my free online tutorial about the topic here.

For creating the Toxic Storm scene, I linked the objects from different files. I used different color maps, shading and texturing techniques on all of them, and that creates a little bit of difference of color between them. Correcting this difference is possible by lots of ways: you can either use a set of textures that match each other's tone, color corrected before using them, or you can use Blenders Color Balance tool to shrink the difference of the tones. However, using the Color Balance node wont solve the issues, it will just render the Shadow, Mid Tone and Highlight values to a desired color, fitting the overall light and color values a bit better, and letting you create different moods for a certain scene. I usually use this node to finalize the mood, not for correcting the tones.

Lift stands for the black levels, Gamma for the mid tones, and Gain for the amount of white.

Since I used mostly desaturated, rusty and old objects on my scene, I could go with just desaturating the colors in Blender. For the barrels, I created two versions: the red one uses a mildly desaturated color texture, and the basic rusty ones are using a strongly desaturated color texture. With this, the barrels don't pop out from the environment, only the one I am focusing the camera on.
For adjusting the saturation, brightness and contrast levels of any textures, go to the textures panel and look for these sliders under Color:

R,G,B values stand for red, green and blue colors. You can recolor the image by using just these sliders, to a desired tone, it lacks accuracy though. I prefer to use Gimp for coloring instead. Brightness, Contrast and Saturation can be manipulated here as well.

Stencils are good for blending two different textures in real time, using 3d tools instead of 2d applications. Blender is able to use stencils, and with the help of them, we can easily create muddy paths in a grassy field, or paint the soil inside and around a lake without worrying about the heights of the ground.
To use this technique, you have to unwrap an object. For practice, use a simple plane, as in stencil example.blend. Create three images, the same size, two colored and a black one and save them. The first texture layer is the first texture with color information, the second one is the stencil layer, and the last one is the second color texture.
For the stencil map, under Influence, enable RGB to intensity and Stencil. Now, you just have to enable GLSL as the drawing method, and in Texture Paint mode, you can blend the two textures. When you are done, remember to save the stencil image, because it wont save automatically, leaving you a headache. Also, it is important to have the stencil image as the active image. If you cant paint anything, or Blender paints on one of the color images, its probably because the active image on the UV layer is one of the color textures, instead of the stencil one.
To check this, go into edit mode, and in the image editor see which textures is under the UV islands. It should be the stencil image. If this is a bit confusing, I don't blame you, I always mix these setting up myself, and mostly because I don't use this technique often. If you practice it, it will become clear quickly, so have fun with the example file.

Let's discuss something that is very crucial in creating art. When I started using Blender, I usually went after my first impressions or thoughts, and started projects without digging deeper into the topic. I didn't bother with lights, with correct texture size or object size proportions, or a certain goal for the project, just did it. The brightest example is my first "animation clip" called Paradise. If you watch that short, you can pretty much tell that it had no concepts for the scene at all, and that it has lots of issues regarding the textures, lights, etc. I could have done better, if I would sketch more ideas and the scene better. Now, the entire short video can be called as a sketch instead.
If you are interested, you can find it here.

• Google - your friend
• deviantArt - biggest art related community
• Flickr - stock photos
• CGTextures - free texture database
• SXC - lots of free stock photos

The way I did my scene let me modify the props easily, since basically all the props have their own .blend files, with their own textures, materials, groups and whatever, and I have simply linked them onto the scene. With this, whenever I modify a prop, it will be automatically updated in the scene.blend as well. This method let me experiment with different textures and looks for each prop I have made.
The first things I wanted to do, were the barrels. You can see the whole modelling and texturing process in the videos directory. Once you have a prop modelled and textured, make a group for it and use the linking feature in the scene file. Textures and colors are not much important at this point, since I was going to place the props on the scene first, to make it perfect, and I'd move to the correction of textures and matching of colors after that.
After finishing the barrels, I thought about the scene more, and found a purpose for it. I wanted to create a once cheerful and bright playground in its worst nightmare, set after a nuclear explosion, right after a toxic storm. For that, I needed to create some swings, benches and even some rubber wheels to put into the ground and hang up as a swing. I was thinking about having some pallets around with wooden debris as well, but after having too much props on the scene, I decided to remove those, and even fully textured bench, since I didn't want undamaged things around. You can find all unused props in the creatively named "not used" directory.

I have ended up with lots of different barrels, benches, swings and even pallets. I had to decide what I wanted exactly on the scene. I knew the plan and I had a sketch already, but I wanted to only have the things which are really necessary for the scene. I wanted the barrels, since they would represent the lost innocence of the playground, creating an ugly presence on themselves. Even the camera focuses on the middle red barrel in the finished scene, and I made that one look a bit brighter, further catching the eye of the watcher. The second thing I wanted for sure are the different swings, bringing back childhood memories: the basic wooden one, and the car wheel one. Benches were needed as well, giving the scene a hint that its not just a small thing, but a part of a bigger park, and further strengthening that idea, I have even developed a small pool of lake on the border of the turntable.
Everything was settled, and I moved on to selecting the best looking props I have made. This part is very subjective, since I love all of my creations, but I had to focus on what would do a proper tutorial subject. For that purpose only, I have selected a half opened barrel, the tire of course, the broken bench and two versions of the swing. I used different texturing and shading techniques for them, but lets take a closer look.

If you have watched Creating the bench and the swing.mp4 already, you have probably noticed that I used multiple objects for the metallic parts and for the wooden parts for each prop. I did this, because I wanted to mess up the planks, and I wanted to keep the frames symmetrical. The whole process stands from basic steps:

• Extruding faces to get the shapes: E
• Creating edge loops for more details on desired parts: Ctrl+R
• Setting the whole object to smooth and applying an Edge Split modifier
• Messing it all up by selecting everything, and drawing a line while holding K for cutting edges where I want the planks damaged and deformed.
• Refining shapes.

Most of this process is documented in the video section, as before. Basically this was the part where I could give the objects their finishing tweaks and this is the part where I corrected the colors. For all of this, I used Linking (Ctrl+Alt+O). Remember, that you can modify only the linked groups, if you link meshes instead, or objects, they wont rescale or move, not even rotate.

When I started using Blender, this was the part I feared the most. Fear comes from not knowing of course, and after I have learned about it, I felt like I've just arrived home. The system begun making sense, and lots of things were finally available for me. I could use Depth of Field or DoF, I could create a simple vignette, or correct the tone of my scene. Lets create some different effects using the scene, just to learn a bit of everything overall. You can check my final composition of course, but it might not be as helpful as just starting from scratch and finding the effects you would like use.
The first thing you need to do for compositing, is a render. Open the composition scene.blend file, switch to composition view (Ctrl+Left arrow), and hit F12 (you might want to simplify the render options first, maybe turn the resolution from the default one down to 1024*768). If you want to use backdrop, use a viewer node. Backdrop image can be resized if you open the N menu (or use V for scaling down).

This is the simplest technique to catch someone's eye. A vignette darkens the four corners of the scene, forcing you to focus in the middle immediately. There are two methods for this: using a transparent image file with alpha transparency, or creating a nice blend directly in blender. Lets check the last version, since that is more flexible, and it can be used with every aspect ratio. The drawback is that it wont be much visible on a dark scene to begin with.
For this, you need to use a Lens Distortion, a Math a Blur and a Mix node.
Add->Distortion->Lens Distortion, connect your Render Layers node with it, change the Distort value to 1. Then Add->Convertor->Math, set it to Greater than, connect the Distortion with Math's top value and set the bottom value to 0. Now, we are going to blur this out: Add->Filter-Blur, set it to Fast Gaussian and Relative, and set the X and Y values fit your aspect ratio. Next, we are going to mix the result with the rendered scene: Add->Color->Mix, set it to Overlay, connect the Render Layers node to the top image socket and the Blur node to the bottom one, and crank the value up to 1. If you connect this Mix node to a viewer, and you do the same with the render layer, you can switch between the two results without the need to render the scene twice. As you can see, the Vignette we created softly darkens the corners and gives more light to the middle of the screen.

Depth of field is a great effect, rendering everything out of focus blurred and leaving the focus subject blade sharp. This effect is often used by next-gen games, and in cinematography, catching our eyes right at the horns and bringing them to the real deal. As before, this one has more approaches as well, but I am going to introduce you the easiest one. For this, you need a focus subject for your camera, for me, this is usually an Empty called DoF. Fear not! We have this already set up in the composition file. But if you are working on an own scene, just add use an Empty or an object. Select the camera, go to Object Data, and under Depth of Field, choose your focus object.
Next thing to do is compositing of course. This is an easy one, we are going to use only one node this time: Add->Filter->Defocus. Connect the Render Layers node to it, even the Z value, and connect the Defocus node to the Composite node. Defocus settings depends on your needs, but basically you need to enable Z-Buffer, and set the fStop value way down, around 20 or so. For the time being, I use 1 for that, exaggerating the effect to the top. Setting the samples higher will make the blurred parts smoother, and setting the Bokeh type and angle will affect the small parts out blurred.

Sharpening will bring the colors and details back to a certain degree, but they can ruin the image totally. I am usually trying to use a combination of two filters when I sharpen something, and those are the basic Sharpen and Kirsch filters. For this effect, you only need to add the two filters and combine them, so let's do it quickly.
Add->Filter->Filter, set it to Sharpen, and set the value way down to 0.15. Add a new filter and change it to Kirsch, with 0.1 factor value. Connect your Render Layers node to each of them and add a new Mix node. Combine the two with a single Mix blend type, but keep an eye on the factor of the combination, 0.35 was just fine for me. Using too much of these guys can blow your scene away with ugly blurry dotting. Also note that in some cases, only one layer of sharpening is enough, or even the other filters can be of help. For example, the Soften filter can help you in getting rid of unwanted dizziness caused by too much sharpening, and this can be aimed to certain spots on the scene as well. Changing the Mix type can help in some case as well, cutting away some edges or light from parts.

This might not be appropriate for my scene, since Bloom is generated by highlighted objects in a very sunny environment, but I hid it in my last composition anyway. For this, you will need an RGB node and a Blur node Mix. Add->Color->RGB Curves, and on the C channel, create a curve similar to the one shown below. I had lots of red highlights, but the outcome of the operation before was too dark to give a good Bloom effect, so I changed the R channel as well, rendering it more visible. After that, you have to add the Blur node, Add->Filter->Blur, and connect the RGB Curves note to it. Use a Fast Gaussian blur with 30-40 values for X and Y. Next up, you have to add a new Mix node, and connect the Render node to it, alongside with the Blur node. Set the blend tipe to Add, and connect your Mix node to the Composite node.

Initially, I wanted to go trough only with my node set-up, but now I think that the basics should be enough, since I didn't wrote the tutorial in a manner that my scene could be reproduced step-by-step. If you want to learn about it, you can check the blend file anyway.

Final node set-up

Finished result