5 minute read
Mattes are the foundation of compositing, allowing you to combine multiple images together. In essence, a mat is like a cut out – it tells the computer which areas of the image should be visible and which should be transparent, allowing other images to show through. In a green screen shot, a matte is created by removing all of the green from the image.
Once a matte is generated, you can enhance, repair, or create entirely new shots by combining it with other mattes in the foreground and background.
You can sift mattes into two categories: static and traveling.
If you want to isolate a shape in your shot that doesn’t change over time, you create a static matte. Why?
Think of a wide establishing shot of a house executed from a camera on a tripod. The lawn has a garden gnome. The gnome wasn’t supposed to be in the shot. To remove it, you could create a matte to separate it from the rest of the shot, then use compositing techniques coupled with your software of choice to do so. Since the shape of the gnome doesn’t change over time, this matte would be static.
If you’re isolating a shape that change shape over time, you create a traveling matte. Why?
Think of a hero leaping from one skyscraper to another. The actor is running, jumping, flailing in the air. If the actor’s performance is shot in a stage lit and designed for visual effects work, you can generate a matte by removing the existing background to isolate the actor’s performance, which can be combined with shots of the desired environment. Since the shape of the actor (definitely) changes over time, this matte would be traveling.
Matte creation has evolved over the years. When computer-generated imaging became a visual effects fixture, mattes took the form of transparency information in digital images.
Digital images typically contain three channels of information to represent color: red, green, and blue. Those three colors are combined to make the final image. But some images have a fourth channel, an alpha channel. Alpha channels contain transparency information that tell your image compositing software which parts of the image are visible, and which are transparent. In other words, mattes can be generated in one piece of software and then bundled with your image to be sent off to another piece of software.
Here are some common methods for generating mattes in computer-generated images, still and motion, with alpha channels.
Luma keys operate on the brightness levels, or luminance values, of an image. A matte can be generated by applying a luma key and designating which brightness levels should be transparent and which should be opaque. This is known as pulling a key.
Chroma keys let you select a specific color in an image (i.e. the chrominance value), designating anything with that color transparent. This is also known as pulling a key. Performances shot in green or blue physical environments were executed with chroma keying in mind.
Transparency can also be defined using masks, which is a shape is drawn over a portion of a shot. You can then define which area in relation to the shape will be transparent, inside or outside. You might think of masks for generating static mattes only, but traveling mattes can also be generated by moving mask points over time as the shape of your masked element changes. Rotoscoping is the process of using masks to create mattes.
A matte defines which areas of an image should be transparent, but a matte can also have many levels of semi-transparency. It’s very common for an artist to fade the edge of a matte so that it blends in smoothly with the background image, for instance.
Knowing what a matte is and how it can be defined in software, let’s look at some common applications of mattes in filmmaking.
Matte paintings can add visual elements in the frame, extend existing sets, or serve as the set itself. For a long time, matte paintings were actual paintings on glass.
Although the medium changed, the principles haven’t. Composition, perspective, and lighting are all carefully considered to execute a convincing matte painting.
For years, matte paintings were mostly static. But with software and increasingly sophisticated techniques, many are now traveling mattes, allowing filmmakers to replace just about anything, foreground or background, in a shot.
Most visual effects shots start out with ‘something in the shot’. Could be lighting gear. Could be some other rigging. Could be a crew member! Usually, all of these are in the shot out of necessity.
To hide these, a garbage matte can be generated. The intent is to keep the wanted elements visible in the shot while hiding the unwanted bits (i.e. the garbage). Depending on how complex the desirable elements’ shape changes over time, it could be a static or traveling matte. In software, these usually take the form of shape masks.
Holdout mattes are connected with keying. When you’re pulling a key, you might discover your subject was reflecting a bit more of your chroma key background than imagined. Portions of the body might disappear despite your best keying efforts. Using a holdout matte, you can duplicate a piece of footage in your software, draw over the portion you want to re-include, and remove the keyer applied. Again, depending on how the desirable elements’ shape changes over time, it could be a static or traveling matte. These are also usually in the form of shape masks.
No matter how sophisticated or layered a shot might be, it all comes back to the matte. Knowing the difference between a static and traveling matte, how they can be applied, and how you might wield them in service to the story will only make your story better. It will also make you a valued team member and provide even more value to the client.
Unlock all 100,000 words of the Frame.io Workflow Guide and learn how the pros do workflow.
Video collaboration solved.