Image Processing Reference
In-Depth Information
There is a hidden benefit to using the newer codecs once the problems of rolling
them out are solved. The additional compression efficiencies will increase the broadcast
real estate by a factor of at least 2:1. This will provide twice as much capacity for adver-
tising revenues to be earned through additional channels. That alone is a very attractive
option for commercial TV executives to consider.
12.5.1
DCT Improvements
A number of rather esoteric additions to the coding process improve its efficiency. These
are in the area of how the coefficients of the discrete cosine transform are dealt with when
processing P-frames. Quantization and variable-length encoding are also improved, so
there are coding efficiency gains to be had simply by deploying the MPEG-4 part 2 codec.
The big problem is that there are a lot of MPEG-2 codecs already deployed in hard-
ware and it is impossible to upgrade them so they will understand the MPEG-4 part
2 syntax.
12.5.2
Global Motion Compensation
The global motion compensation is useful because when it is applied to the object as a
whole, the effects of motion compensation at the macroblock level are reduced, therefore
reducing the encoding complexity.
12.5.3
Sub-Pixel Motion Compensation
By allowing motion to be calculated at the sub-pixel resolution, slow-moving objects are
coded more accurately and the prediction frames are described with fewer errors and
residuals.
Figure 12-1 shows the result of analyzing the motion vectors on a Tektronix MTS4EA.
12.5.4
Scaled Video
Recall that MPEG-2 introduces scalable video. The concept revolves around delivering a
basic-quality and -resolution signal with a difference signal that is added to augment the
bit stream to provide sufficient data for a higher-quality output. MPEG-2 is scalable in
size, pixel density, and frame rate using this technique. MPEG-4 supports this, but at the
individual-object level and it allows more than one enhancement layer if necessary. There
are potentially many different permutations allowed and the spatial scaling is delegated
to the decoder if necessary.
12.6
Shape-Coded Video Applications
The alpha shape-coded video support provides some interesting possibilities for deploy-
ing signing overlays for those who are hard of hearing. These would be carried alongside
Search WWH ::
Custom Search