Image Processing Reference
In-Depth Information
The comparison process needs to be somewhat “fuzzy.” because there is some diffi-
culty in defining what exactly is a good enough quality level. A completely non-trained
person who has never seen very high quality displays in the laboratories might be satis-
fied with a quality level that is approximately the same as a VHS recording. Someone who
has worked as a TV diagnostic or research engineer and has seen the very-high-definition
TV displays at the NAB or IBC conferences will find that same VHS-quality level very
unsatisfactory.
It is difficult to do this kind of measurement and comparison automatically, even
when some artificial intelligence is used. There is no substitute for human intervention
when running quality assurance on your compression farm.
Then there is the technical side of the process. You may have mandated certain bit-
rate requirements, and you must verify that the video is compressed to comply with them.
For example, if you cap the bit rate of a video channel in a multiplex or transponder, you
cannot afford for the video to burst at a greater bit rate unless a saving is made elsewhere
in the multiplex. If the transport stream running within the multiplex exceeds the bit-rate
capacity available, the multiplex will break and you end up with an interrupted transmis-
sion. The receiver will resynchronize but your viewers will be displeased.
38.7
Summary: How to Achieve Good Quality Every Time
There are aspects of video processing that have a material effect on the quality of our com-
pressed-video output. Having some insights into them will help as we move toward the
goal of compressing some footage. We should also bear them in mind as we plan how to
build a system to do this compression work.
These last few chapters are what it is all about. Everything in the topic up to this
point prepared you for that journey, step by step through the encoding process. You might
not have thought some of the earlier stuff was relevant, but now you understand why
video compression works the way that it does. Recall my earlier assertion, at the begin-
ning of Chapter 5, about going back in time to reinvent some of those technologies so that
they would compress more easily.
By now you can appreciate 1) why compression has become so complex and 2) that
that is because of technology choices made in the early days of the television evolutionary
process. Those decisions were perfectly reasonable at the time, but we are paying a high
price in terms of processing power and throughput in our compression systems. With the
advent of worldwide adoption of HDTV systems, which are progressively scanned and
operate at similar display formats everywhere, we may be able to eliminate some of this
complexity over the next 10 to 20 years.
In the last chapter we will look at the future—where digital video might be going
over the next few years and the implications for video compression.
Search WWH ::
Custom Search