We’ve put together a series of dedicated Insights into our work at LaserLines Creative.
This issue gives details of the processes behind our 3D Illustration and Animation work.
If you would like to know more, or have a project in mind, please Get in Touch
3D Illustration and Animation, what is it?
Firstly we use a process called 3D modelling. We create a virtual wireframe of a three dimensional object using a 3D modelling software package, very similar to creating a basic shape out of chicken wire.
Setting the Scene – Textures
Creating the 3D wireframe is only the first stage in generating a realistic 3D illustration. The next stage is to texture the 3D object. A texture is an image or pattern which then gets wrapped around the wireframe model like a plastic shrink wrap. We can set numerous properties for each texture such as colour and surface appearance, whether it’s rough or smooth, reflective or matt, dull or glowing, the options are endless.
Lighting the scene is one of the most critical stages in achieving a photo realistic result. The 3D software is like a virtual studio where we can position our 3D object in the scene, set the surroundings and light it to get the effect we want to achieve. For example, to mimic distant sun light we’d use distant blue and yellow lights to simulate the sun and reflection of the blue sky. If we want a more focused setup, in a close-up product shot for instance, we can create a more vivid effect with a mix of spot lights and subtle ambient lighting.
Finally, we set the camera angle by moving the virtual camera around just a we would in a photo studio. We’ve also got control over settings such as focal length and depth of field, so we can choose how far away we want to focus, how much we want to blur the background and whether we want a wide angled shot with lots of perspective or a flatter less dramatic angle.
3D Rendered Still Image – The Final Output
Rendering is the process in which the software paints the picture. It runs thousands if not millions of calculations, working out how the light will bounce off of different objects and the light beams then scatter around the scene. The more objects that are in the scene, the longer the render will take. The more complex an objects surface, the longer a render takes, the higher resolution the render, the longer the render takes. This is the stage that all 3D studios hate. If everything was setup up correctly however, the results are worth the wait.
3D modelling is very flexible solution and allows us to create images at any size. We can produce low resolution (1200×900 pixels or lower) images for use on the web, high resolution images suitable for printing full page (5000×3000 pixel) right up to super high resolution images (20,000 x 15,000 pixel or higher) for bill-board advertising.
Once the scene is setup, we can take the project one step further by animating it. We can animate and move each individual object in the scene. We can control how fast it moves, its direction, speed, acceleration / deceleration rates and much more.
3D animation can show things that are simply not possible with video alone.
Terminator 2, a liquid metal human figure, melts into the floor. This was probably the first time we witnessed realistic human movement that was 3D computer generated.
Remember how impressive the fully animated dinosaurs were the first time you saw Jurassic Park? And finally Toy Story, the first film to be completely 3D computer generated, without a conventional film camera in sight!
3D animation and special effects are now heavily used in most of today’s feature films. You probably don’t need 3D generated dinosaurs running around terrifying people in your corporate video (although we’re happy to provide them if you do). That doesn’t however mean 3D animation isn’t for you.
3D animation is a massively powerful tool and can be used to demonstrate all sorts of products and processes
Simulating Real-World Forces
One of the most exciting recent advances in 3D animation software has been the introduction of powerful physics engines that help simulate real-world forces such as gravity and collision detection. This lets us create very convincing movement when one object hits another. We can alter the weight, softness, bounce, friction levels and gravitational pull of an object to influence how it interacts with other objects.
These forces also come into play when we animate a character. We also need to set other parameters such as limiting the angles that legs and arms can rotate and how they react with other parts of the character.
What if we want to make waterfalls fall, fire burn and smoke rise? For this we use particle systems. These can generate a mind-boggling number of individual particles that can take any form we choose. To generate a realistic looking firework for instance, we would only use between 200-500 large particles that all emit and blast out from the same point and have a set life span of say, two seconds. We can also make the colour and brightness of the particle vary over its lifespan.
When we move on to large realistic clouds, the computation gets much more complex. Now, instead of 100-200 particles, we may need in excess of 20,000 pixels depending on the size of the cloud. We can then set properties that effect how the clouds react when an object passes through them. The examples given barely scratch the surface of what can be achieved, with software getting more and more advanced and computer processors and graphics cards getting faster and faster, what can be realistically created is changing all the time.
As with 3D illustration, the final stage is the rendering where the software creates the final images. It can take a few hours to render a single image in a complex 3D illustration. Now take into consideration that to generate one second of animation we need to generate 25 individual frames. So even a relatively short 30 second clip would contain 750 frames. Now obviously we can’t set about rendering a 750+ frame animation if it’s going to take over an hour a frame. This is where we need to optimise things a little more than we would with a single illustration.
For large jobs we render over our network using every multi-processor workstation we have so we can achieve what we need within the deadline. Our average animation job can be rendered over night and is ready waiting when we arrive to work the next day.
If it’s a very large job with a tight deadline, we can use dedicated render farms which have 200+ processors at their disposal to get the job done in record time. 3D feature film Avatar set a high benchmark for realism. But realism comes at a cost. They had to use a massive render farm comprising of over 40,000 processors (yes you read that right) to get the rendering done. To render a single frame of Avatar on a conventional workstation would take over 48 hours!
We can output our animation in number of different formats optimised for online streaming or output at higher quality in either SD, HD or 4K for DVD, Blu-Ray etc.