Image Processing Reference
The industry at large seems to be calling it H.264 when describing it in their product-
feature set. If a video codec is referred to as MPEG-4, you should assume that it means the
part 2 variant in the absence of any other qualifying details. The recommended nomen-
clature is H.264 or AVC.
Evolution to a Higher Form
The codecs have evolved gradually starting with MPEG-1, adding improvements for
MPEG-2, and continuing with the enhancements for MPEG-4 part 2. Each one is slightly
more complex than the one before.
A lot of people were involved in the design of this codec. Many of them were special-
ists in very narrow fields. They all set out to achieve the best possible image quality at the
most efficient compression ratio that they could accomplish. At the extremes of the perform-
ance curve, you have very complex, high-quality modes at one end, but low complexity with
some necessary quality trade-offs at the other end. The decoding process is unambiguous.
All decoders should produce exactly the same output. Fine detail, gradients, and motion are
all dealt with far more effectively than they have been in earlier codec designs.
Figure 13-1 shows a comparison of the bit rates achieved by the same quality video for
MPEG-2 and H.264. The H.264 plot includes work done under the H.26L and MPEG-4
part-2 work streams. Note that MPEG-2 is now outperforming the early work on H.264.
Figure 13-1 Relative compression efficiency between MPEG-2 and H.264.