Simply put, an aspect ratio of an image is its displayed width divided by its displayed height. Aspect ratios are commonly represented as x:y (for example, 4:3) or x (for example, 1.33), in which x is the width and y is the height. A list of popular aspect ratios, their origins, and representative image resolutions follow (see Figure 10.1).

Figure 10.1. Common aspect ratios

The NTSC D1 and DV video standards require nonsquare pixels. NTSC D1 and DV nonsquare pixels carry a pixel aspect ratio of 10:11, wherein the height is greater than the width. (10:11 is often expressed as a rounded-off 0.9.) Nonsquare pixels are a legacy requirement of 4:3 SDTV television technology. (The electronic design of a 4:3 SDTV television allows the image to appear normal.) By comparison, computer monitors and HDTV televisions use square pixels. Thus, when a D1 or DV image is displayed on a computer monitor, it appears vertically squished. Editing programs such as Final Cut Pro and Adobe Premiere avoid this problem by offering Show As Square Pixels options.

You can set a nonsquare pixel aspect ratio in Maya by changing the Pixel Aspect Ratio attribute in the Render Settings window. For instance, you can set the value to 0.9 for NTSC. However, in order to maximize render quality, it is generally better to render square pixels and "squish" the image in the composite (see Figure 10.2). Thus, when preparing for NTSC DV in Adobe After Effects, a square-pixel 720 × 540 render can be converted to a nonsquare 720 × 480 composite by scaling the image by 90 percent in the X and 88.9 percent in the Y. The size reduction will also serve as an additional form of anti-aliasing. When preparing for NTSC D1, a square-pixel 720 × 540 render can be converted to a 720 × 486 nonsquare composite resolution by scaling by 90 percent in the X and Y.

Figure 10.2. (Left) Square-pixel 720 × 540 render in After Effects. (Right) Same render fit into nonsquare 720 × 480 composite.

In Maya, a camera's virtual optical properties are defined in the Film Back section of the camera's Attribute Editor tab (see Figure 10.3). The most important attribute in this section is Film Gate. With a real-world motion picture camera, a film gate is a plate with a rectangular opening that sits behind the lens and in front of the film stock. The gate controls the exposure of the film by allowing only one frame at a time to be struck by light. Commonly used cameras have distinct lens and film gate setups with specific physical properties. These properties cause light to strike the film in a distinct way. The Film Gate attribute's drop-down menu (see Figure 10.3) provides a selection of common motion picture cameras. When one of these camera presets is selected, the Camera Aperture, Film Aspect Ratio, and Lens Squeeze Ratio attributes automatically update. A description of these attributes follows.

Figure 10.3. (Left) The Film Back section of a camera's Attribute Editor tab; (right) Film Gate presets

If an animation represents an abstract, fanciful, or imaginary scene, you can leave the Film Gate attribute set to the default User. For exaggerated camera effects, you can also enter custom numbers into any of the attributes in the Film Back section. However, if an animation is intended to replicate or fit into preexisting live-action motion picture footage, the Film Gate and associated attributes should be carefully selected.

As you choose an aspect ratio and a film back for an animation project, it is recommended that you use the camera's various gates. Choosing View > Camera Settings > Resolution Gate through the camera workspace view menu displays the Resolution Gate frame in the perspective view. The Resolution Gate represents the maximum render area (see Figure 10.4). The renderer ignores anything outside the Resolution Gate. For extra security, you can toggle on Safe Action and Safe Title gates by choosing View > Camera Settings > Safe Action and View > Camera Settings > Safe Title. Since the average television set (whether SDTV or HDTV) cuts off a portion of the frame between the Resolution Gate and Safe Action gate, keep critical animation within the Safe Action gate. Keep critical text and titles within the Safe Title gate.

Figure 10.4. The Resolution Gate with Safe Action (middle box) and Safe Title (inner box) toggled on

The Resolution Gate aspect ratio, determined by the Device Aspect Ratio attribute of the Render Settings window, may be different from the Film Aspect Ratio. (See the section "Mastering the Render Settings Window" later in this chapter for more information on Device Aspect Ratio.) Hence, it's recommended that the Fit Resolution Gate attribute, in the Film Back section of the camera's attribute Editor tab, be adjusted. If Fit Resolution Gate is set to Fill, the Film Gate is scaled so that it covers the entire Resolution Gate. Depending on the render resolution, a portion of the Film Gate may be lost outside the Resolution Gate. To see the results of Fit Resolution Gate, you can display the Resolution Gate and Film Gate simultaneously by checking Display Resolution and Display Film Gate in the Display Options section of the camera's Attribute Editor tab. If Fit Resolution Gate is set to Horizontal or Vertical, the Film Gate is scaled in one direction so that it matches the Resolution Gate's dimension in that direction; any gap that appears in the opposite direction will render as normal unless Ignore Film Gate is unchecked in the Render Options section of the Maya Software tab in the Render Settings window. If Fit Resolution Gate is set to Overscan, the Film Gate is fit within the Resolution Gate so that no part is left unrendered; a gap between the gates may appear at the top and bottom. The only way to avoid gaps is to provide the Device Aspect Ratio and Film Aspect Ratio with identical values.

Focal length is the distance from the optical center of a lens to its focal point (the point at which the light rays are focused). Once specific shots are determined within an animation project, it's wise to select specific focal lengths. Motion picture and television productions regularly change focal lengths (and even the entire lens) from shot to shot. Different focal lengths can have a significant impact on camera placement, composition, and object distortion. (See Chapter 7 for a demonstration.) Focal length is controlled in the Camera Attributes section of the camera's Attribute Editor tab, which includes Angle Of View, Focal Length, and Camera Scale attributes (see Figure 10.5).

Figure 10.5. The Camera Attributes section of a camera's Attribute Editor tab

Figure 10.6. Clockwise, from top left: 20 mm, 35 mm, 135 mm, and 50 mm focal lengths in Maya with a 35 mm Academy Film Gate

A proper frame rate, or frames per second (fps), is critical for smooth animation. To set the rate, choose Window > Setting/Preferences > Preferences, switch to the Settings section, and choose an option from the Time attribute drop-down menu (see Figure 10.7). Maya provides the most common frame rates, including 24, 25, and 30.

Figure 10.7. A portion of the Maya Software tab in the Render Settings window

Standard television transmission requires the use of interlaced fields. Thus, in reality, PAL runs at 50 interlaced fps and NTSC runs at 60 interlaced fps (or, more accurately, 59.94 interlaced fps). The interlacing process splits any given frame into interlaced upper and lower fields, with one field drawn first and the other field drawn second. Whether the upper field or lower field is drawn first is dependent on the video's field dominance. This varies with video format. You can render interlaced frames in Maya by choosing PAL Field or NTSC Field from the Time drop-down menu in the Settings section of the Preferences window. Compositing programs, such as Adobe After Effects, can also convert noninterlaced frames to interlaced frames at the point of render.

One of the most difficult aspects of rendering is the conversion of one frame rate to another. The conversion of motion picture footage to NTSC video, for instance, requires the 3:2 pulldown process. A 3:2 pulldown converts four film frames into ten interlaced video frames. Two of the frames are repeated three times and two of the frames are repeated twice. 3:2 pulldowns are normally created with telecine machines but can be created with compositing programs such as Adobe After Effects or Autodesk Combustion. Compositing plug-ins, such as RE:Vision Effects Twixtor, offer additional techniques for interpolating and smoothing out frame rate conversions.

To avoid potential fps conversion difficulties, determine the primary presentation format of an animation project early on. If the work is destined for 35 mm transfer and a theatrical release or the film festival circuit, 24 fps would make the most sense. If an animation is created for a television commercial in the United States, 30 fps is necessary. If an animation needs to go to multiple outlets at multiple points around the globe, conversion artifacts should be expected. Even though many postproduction houses can electronically or digitally convert between frame rates, the result is never as smooth as the original. For instance, converting from PAL to NTSC will leave a "judder" in the animation where a slight hesitation appears every few frames.

The Maya Software renderer is a general-purpose renderer that is suitable for most projects. Critical attributes include Edge Anti-Aliasing, Shading, and Max Shading. Important sections include Multi-Pixel Filtering and Contrast Threshold (see Figure 10.10).

The Maya Hardware renderer provides a quick method of rendering tests and other projects that do not require a high degree of refinement (see Figure 10.11). The Hardware renderer uses the built-in capabilities of the system graphics card.

Figure 10.11. (Top) A model rendered with Maya Software. (Bottom) The same model rendered with Maya Hardware via an entry-level graphics card.

The easiest way to set the quality of the Hardware renderer is to use one of the four options of the Presets attribute (Preview Quality, Intermediate Quality, Production Quality, and Production Quality With Transparency). Nevertheless, many of the corresponding attributes are unique and are worth a closer look (see Figure 10.12).

Figure 10.12. A portion of render Help, as shown in the Command Prompt window

The Maya Vector renderer can create stylized cartoon and wireframe renders (see Figure 10.13). Although the majority of options are straightforward, a few warrant a more detailed description.

Figure 10.13. Top, Left to Right) Maya Vector renderer with Single Color and Entire Mesh, Vector with Single Color and Outlines, Vector with Four Color. (Bottom) The Maya Vector tab in the Render Settings window.

The speed of any given Maya render depends on the quality of the scene file. If a scene file contains unnecessary construction history, broken nodes, and unneeded geometry, the render will suffer. A quick solution to this problem is to choose File > Optimize Scene Size. By default, Optimize Scene Size deletes unused curves, unassigned materials, orphaned group nodes (those without children), and empty layers. By opening the Optimize Scene Size Options window, you can optimize specific categories by checking or unchecking the category buttons (see Figure 10.15). Use caution when dealing with complex scenes since it is possible to unintentionally delete critical components of character rigs and other advanced setups.

If you are unable to determine why a particular scene is rendering slowly, switch to the Rendering menu set and choose Render > Run Render Diagnostics. The Script Editor opens and displays suggestions for optimizing the scene in question (see Figure 10.16). Although these suggestions can be quite helpful, they are by no means mandatory.

During the modeling process, it is also important to choose Edit > Delete By Type > History when construction history is no longer needed. If construction history remains on a rigged character, for instance, the render time can be significantly increased.

Figure 10.15. The Optimize Scene Size Options window

Figure 10.16. A sample render diagnostics message displayed in the Script Editor window

Maya .mb and .ma files contain all the elements required for a render—except actual bitmaps. Instead, paths pointing to bitmaps are hard-coded in a Maya file. For example, as shown at the top of Figure 10.17, a .ma file contains the following line:

settAttr ".ftn" -type "string" "C:/3D/logo.tif";

Figure 10.17. (Top) The hard-coded bitmap path of a Maya .ma file, as displayed in a text editor. (Bottom) A truncated path listed by the Image Name attribute of a File texture.

Thus, if the Maya file in question is moved between computers with different drive letters or directory structures, logo.tif will be "lost." In this case, the Multilister and Hypershade windows display a black icon for bitmaps that are missing. This holds true even if the project directory has been set through File > Project > Set. In such a situation, Maya displays a truncated path in the File texture's Image Name field (see the bottom of Figure 10.17). Nevertheless, Maya will be unable to find the texture if the drive letter or directory structure has changed.

Fortunately, you can fix this problem by quickly editing an .ma version of the file, which is simply text. Using the Find and Replace All function of Microsoft Windows WordPad or an equivalent text editor, replace c:/ with d:/ or any other appropriate path. In a similar fashion, you can edit .mb files. However, since .mb files are binary, a hexadecimal editor is required (see Figure 10.18).

Figure 10.18. The hard-coded bitmap path of a Maya .mb file, as displayed in a hexadecimal editor

Maya Software, Hardware, and Vector renderers can output 30 different image formats. You can select the format by switching the Image Format attribute in the Render Settings window. Although any of the formats can be used successfully in the right circumstance, AVI, QuickTime, JPEG, Targa, and TIFF formats are perhaps the most popular. At the same time, Maya IFF, PSD, Adobe Illustrator, EPS, and SWF formats are designed for specialized tasks.

The majority of Maya image formats operate in an 8-bit color space (8 bits in red, 8 bits in green, and 8 bits in blue, totaling 24 bits, or 16,777,216 possible colors). In the realm of consumer electronics, this color space is commonly referred to as True Color. At present, the majority of consumer monitors offer a 32-bit variation of True Color. This is a 24-bit color space with an extra 8 bits set aside as an empty placeholder (necessary for 32-bit architecture) or for alpha information. By comparison, Maya16 IFF, TIFF16, and SGI16 are three available Maya image formats that operate in 16-bit color space (16 bits per channel, totaling 281 trillion possible colors).

The human eye is popularly believed to discern 10 million color variations. As such, 16-bit color may seem like extreme overkill. However, many image-processing filters create superior results when operating at a higher bit depth. Hence, programs such as Adobe Photoshop and Adobe After Effects offer the option to work with 16-bit images. Low bit-depth errors are most commonly seen as banding (posterization), where the color transitions fail to be smooth (see Figure 10.19). Although 8-bit color space is satisfactory for many applications, 16-bit color space is superior for any project in which color and color manipulation is critical.

Figure 10.19. (Top) Color banding caused by a Gaussian blur in 8-bit color space. (Bottom) The result of the same blur in 16-bit color space.

To change the default compression setting of JPEG and TIFF formats within Maya, you must create an environment variable. Follow these steps:

If time permits, it is always best to render larger than the default size necessitated by a particular project. For example, if an animation is to be created for a video, it is not necessary to stick to 720 × 540. Any multiplier of the 1.33 aspect ratio is equally valid. For instance, 798 × 600 or 1197 × 900 would work equally well. The use of an odd size assumes that the rendered frames will be taken into a compositing program where they can be resized. Since professional animations generally require compositing, oversized rendering can be employed quite often.

Oversized rendering guarantees an animation one very important thing—additional anti-aliasing. This anti-aliasing occurs when the oversized image is shrunk. Compositing programs such as Adobe After Effects must average the pixels of an image during a size reduction. Ultimately, this averaging reduces stair-stepping and other common aliasing problems (see Figure 10.20).

Figure 10.20. (Top) Detail of standard 640 × 480 render. (Bottom) The same image rendered 2560 × 1920 and scaled down.