Specifications
9
A Digital Video Primer
Color space issues
When producing video, knowledge of color sampling is a plus, but you will rarely have to
think about it. Typically, the only time you’ll run into problems is when converting or crossing
between color spaces. In most situations, conversion happens automatically and the result is
acceptable or unnoticed. However, you should be aware of one situation in particular that can
signicantly reduce color delity.
e DV video format, discussed in the next section, uses 4:1:1 color, while DVDs use 4:2:0 color.
Quite oen, producers shoot on DV to reduce costs but distribute on DVD, because of its wide
availability. e problem arises when converting from DV (4:1:1) to DVD (4:2:0). Here’s why: e
color components in 4:1:1 are reduced to 1/4 resolution in the horizontal domain, and the color
components in 4:2:0 are reduced to 1/4 resolution by going to 1/2 resolution in both the hori-
zontal and vertical domains. When you convert directly from 4:1:1 to 4:2:0, a great deal of color
resolution is lost. To avoid the loss of resolution when video is destined for DVD, make sure your
source video uses the 4:2:2 or 4:2:0 color space.
Video compression
Whether you use a capture card or a digital camcorder, in most cases, your digitized video will
be compressed. Compression is necessary because of the enormous amount of data required for
uncompressed video.
A single frame of uncompressed video takes about 1MB of space to store. You can calculate
this by multiplying the horizontal resolution (720 pixels) by the vertical resolution (486 pixels),
and then multiplying that by 3 bytes for the RGB color information. At the standard video rate
of 29.97 fps, uncompressed video consumes about 30MB of storage for each and every second
of video and over 1.5 gigabytes (GB) to hold a minute of video. In order to view and work with
uncompressed video, you would need a very expensive disk array, a very fast CPU, and a whole
lot of RAM to move and process all that data in real time.
e goal of compression is to reduce the data rate while keeping the image quality high. e
amount of compression depends on how the video will be used. e popular DV25 format
compresses at a 5:1 ratio. In other words, the video is compressed to one-h of its original size.
Video you access on the web might be compressed at 50:1 or even more. Generally, the higher the
compression ratio, the lower the quality.
How compression works
Before applying compression, there are a number of ways to reduce the size and bit rate of a video
le or stream. One method is to simply reduce the dimensions of each video frame. A 320 x 240
image has only one fourth the number of pixels of a 640 x 480 image. Reducing the frame rate
will also reduce the data rate. An uncompressed 15 fps video has only half the data of a 30 fps
video. ese simple methods won’t work, however, if a video is to be displayed on a television
monitor at full resolution and frame rate.
Beyond reducing the dimensions and frame rate, compression is also most oen required to
reduce the size of video. To get the compression needed to work with audio and video, a codec is
used to compress and then decompress the content. Codecs may be found in hardware (for
example in DV camcorders and capture cards), or in soware. Some codecs have a xed com-
pression ratio that compresses video at a xed data rate. Others can compress each frame
dierently depending on the content, resulting in a data rate that varies over time. Many codecs
enable you to select a quality setting that controls the data rate, or a data rate that controls the
quality. Such settings can be useful for editing. For example, you may want to capture a large
quantity of video at a low-quality setting to edit a rough cut of your program, and then recapture
just the portions that will go into the nal edit at a high-quality setting. is process enables you
to edit large quantities of video with a smaller hard disk, because you do not need to store the
high-data-rate video that will not be used.
Color detail is lost when converting from DV25 com-
pression, which uses 4:1:1 sampling, to MPEG-2, which
uses 4:2:0.