Digital Signal Processing Reference
In-Depth Information
>As there was no release of MPEG-3, the next standard was
MPEG-4. In parallel, the Video Coding Experts Group (VCEG)
e
a committee operating under the ITU organization
published
video compression standards. They developed the H.261 and
H.263 standards, which were optimized for video conferencing, in
which low bit-rates and low latencies are particularly important.
These two groups eventually merged into the Joint Video Team
(JVT). VCEG had started work on an Advanced Video Codec (AVC),
which was similar to the methods used inMPEG-4 version 10. The
merged efforts have become known as H.264 part 10/AVC, or
H.264/AVC, or just H.264. Blu-ray disc video is compressed using
H.264, as is much of the video content streamed from YouTube.
This chapter will begin by highlighting the enhancements
from MPEG-1 to MPEG-2, which help explain the widespread
adoption of MPEG-2. MPEG-4, in its early versions, wasn't as
widely adopted as it did not offer enough improvement over
MPEG-2 to warrant changing. There were also some disputes over
licensing terms. H.264 offers a more substantial level of
improvement and flexibility over MPEG-2. Due to the prevalence
of high definition video, it has been widely adopted in studios,
consumer video players and by internet video-streaming sites.
The latest standard, H.265, has still not been finalized as of this
writing, and is not covered here.
e
15.1 MPEG-2
The techniques of the previous chapter form the basis of all
the subsequent video compression standards. I-, P- and B-frames
are still used. In MPEG-2, it is typical though not required for
every 15
th
frame to be compressed as an I-frame.
Despite the common technical base with MPEG-1, there are
important differences and improvements in MPEG-2, and many
of these will be reviewed here.
15.1.1 MPEG-2 Levels and Pro
les
A level is used to constrain resources in the decoder, which
often must meet strict cost limitations. By establishing different
levels, the decoder manufacturer can choose how much video
processing resources (and cost) to provide to meet a particular
capability. The level setting puts a limit on parameters such as the
maximum frame size, the frame rate, size of decoder video-frame
buffer and howmany previous frames may be used for prediction.
For example, a fairly low level might be used on a mobile device,
while a higher level would be used on home entertainment
