El oficio del maestro de capilla

As I mentioned in the last chapter, DVD production offers fantastic opportunities for the independent producer. Using prosumer equipment, you can affordably shoot, edit, and distribute professional-quality copies of your work in an all-digital format. Editing applications and DVD production packages ship with encoders that convert your DV footage to the DVD-compatible MPEG-2 format. Understanding how these encoders function before you go out and shoot can help you create better quality DVDs.

Not all shots compress as easily, or as well, as you might want them to. Certain backgrounds and some types of movement might look great when you record them but lose significant quality when compressed for DVD. Once you develop a solid understanding of DVD compression, you can compose shots that still look really good after they've been compressed.

As always, your artistic vision is the most important factor. If you're spending the time, money, and effort to make a film, you have to include the shots you want and the images you think people want to see—even if they aren't the most compression friendly. Understanding DVD compression can help you ensure your most ambitious compositions and framings translate to DVD the way you want, with a minimal amount of headache.

Audio and video files take up tremendous amounts of memory. Five minutes of DV footage takes up approximately 1 GB of memory on your computer's hard drive. A single DVD only holds about 4.7 GB, which translates to less than 25 minutes of camera-format DV (which means DV that's been captured from a camera without applying any additional compression). Just about any disc you pick up at the video store plays for more than 25 minutes, and the reason producers can fit more material onto commercial DVDs is the video and audio have been compressed. Compression removes redundant information from a digital signal, which results in a more manageable file size. DVD compression programs, such as Apple's Compressor, use algorithms to analyze digital audio and video and identify information that can be removed without anyone noticing. These algorithms (which are essentially complex mathematic formulas) use two types of compression, spatial and temporal.

The image on top in Figure 3-6 lends itself to easy compression, while the image below it does not. The image of the person standing against a white wall compresses without difficulty, because much of the image, specifically the white wall, contains very little detail. The image of a person standing in front of a textured background is more complicated, therefore more difficult to compress. While good compression is still possible, the results may be more noticeable. The background contains fine areas of detail that make it harder to simplify.

Figure 3-6. Two drawn images: one of a person standing in front of a white wall, other of a person standing in front of a finely textured background.

Spatial compression compares different parts of a single frame of video to determine what information repeats in the image. Spatial compression algorithms break an image down into blocks, and if parts of a frame repeat, the algorithm places the same block in more than one location. If, for example, a person is standing in front of a white wall, the compression algorithm repeats one section of the background in multiple locations to recreate the wall, rather than saving an image of the entire wall, which would require more memory.

If you've saved an image for the World Wide Web as a JPEG (also written jpg, pronounced "Jay Peg"), you've used a form of spatial compression. JPEG compression examines similarities in an image, and, depending on the quality setting, eliminates higher or lower amounts of information. Spatial compression of video clips works essentially the same way: higher compression settings result in video that requires less storage space. As you apply increased amounts of compression, you also lose image quality.

In Figure 3-7, although each frame in the sequence is different, the background remains a constant. Even though the person walking is in a different position in each frame, the background is still a solid white wall.

This enables the compression algorithm to redraw areas of the wall over several frames, instead of creating an entirely new background for each frame of video. This saves space on the disc and reduces the

processing power needed to play the compressed finished product.

Figure 3-7. A series of drawn figures showing a person walking past a white wall.

If material is too difficult to compress, or if a sequence contains too much information for a DVD player to process during playback, the video will break down. Symptoms include large blocks appearing in

seemingly random locations on the screen, or the DVD player may momentarily "hang up" or pause during playback.

Temporal compression compares multiple frames of video to look for information that repeats from one frame to another. If we use the same example of a person standing in front of a white wall, the wall doesn't change from frame to frame. Rather than saving multiple frames of the unchanged wall, temporal

compression would use a single image of the wall repeated over several frames to save information.

Algorithms that use both spatial and temporal compression would save information by repeating small sections of the wall over several frames.3.2.2. Shots that make compression easier

Evenly lit, solid backgrounds make compression much easier. Finely detailed backgrounds, such as textured drapes, are harder to compress, because the detail requires more information. I once worked with a group of students who shot some video in front of a cork bulletin board. It looked great until they tried to compress it. The small differences in color made compression difficult, and the cork looked like a blocky mosaic in the compressed version. You may notice similar problems if you compress video of people wearing striped or patterned shirts—algorithms often have trouble processing fine details that appear close together.

Note: In situations where a shoot requires more mobility than a tripod allows, you can use a camera stabilization device such as a Steadicam or Glidecam. Both are described in Chapter 6.

Pans and tilts also make temporal compression more difficult. Temporal compression compares one frame of video to the next, so if the contents of each frame are different due to camera movement, the material becomes harder to compress. If the camera doesn't move, much of the background will likely be the same from one frame to another, even if there's movement in the foreground. When the camera moves,

everything is slightly different in each frame, because the camera records each frame from a different angle.

Handheld camera work by definition makes compression more difficult because of the motion—even the most experienced and steady camera operator doesn't stand perfectly still when she holds a camera.

Mounting the camera on a tripod creates a more stable shot, which is easier to compress.

For the same reasons that camera movement makes temporal compression more difficult, background motion does also. If the background of your shot contains motion—for example, textured drapes blowing in the breeze, leaves rustling in a tree, or reflected sunlight sparkling on water—the shot becomes more difficult to compress.

The easiest footage to compress would be a static shot of a solid wall, with nothing happening in front of it.

This would also, of course, be a very boring shot that not too many people would watch. When it comes down to it, you have to include the action that makes your film exciting and the backgrounds that make each shot worth looking at—even if they don't compress all that easily. This doesn't mean you should ignore the realities of compression when framing your shots, if you shoot actors in front of the wrong background, compression may become a frustrating experience, especially if the shot involves a pan or a tilt.

Before you assemble your full cast and crew, choose the locations you want, record a few test shots, and then compress some of your footage—you can't tell what a shot will look like until you actually see it on a screen. You may find out everything compresses exactly the way you'd like, or you may decide to make some changes in your composition. Either way, your goal as a director is to focus an audience's attention on your story, not on technical problems. Try a variety of framings, and incorporate a number of different camera angles and backgrounds. This way, you'll be all set when your star tells you she's ready for her close-up.

This chapter introduced different types of shots, the ways they work together, and the idea of juxtaposing a series of different shot types to form a sequence and tell your story. This chapter also explored the way different compositions, visual elements, and framing choices impact the way footage compresses for distribution via DVD (the same principles apply if you stream video over the Internet, which this book explores in Appendix A).

Note: It's important to know about technical details, but don't let them limit your creativity. During the planning stage of my most recent project, Mi Querdia America, I spent a great deal of time discussing compression strategies with my director of photography. The project follows a group of immigrant teens through their first year of high school in New York City and was shot to be available entirely online. As you can imagine, maintaining good visual quality after compression was a top a priority. After some very lengthy discussion, I finally looked at my DP and said, "You know what, just make it look good." He did—

the video he shot was great, and I worked out the compression details later. The results are online at http://www.digitaldocumentary.org/america.

As a director, you have to have a mind that is good at details as well as the bigger picture. If you have the mind for both, you're going to be a good director. If you don't have the mind for both, pick people for your crew that have the mind for the aspects you lack.

The next chapter addresses lighting, which can make your work look very polished and professional if handled well, and is immediately noticeable to an audience if there are any problems. The chapter covers ways to light various situations, and why you should always use the manual settings instead of relying on a camera to automatically make adjustments for you (as I always tell students, talented people are always much more valuable to a production than any piece of equipment on its own—that's what enables you and me to make a living).