Factors to Consider when Choosing an MPEG-2 Encoder

This paper intends leading the audience through some of the important criteria to consider when selecting a video encoder. In addition to discussing features, the presentation will outline the practical tests that users can perform to assess the relative merits of MPEG-2 encoder products. — MPEG-2 video encoders have been on the market now for nearly five years. By now buyers may have expected to be able to call on a strong pool of expertise to assist in picking through the hype and so reach a rational conclusion. This dream scenario is far from complete. — Many people forget that MPEG defines the standard characteristics of the stream such that it can be decoded. This successful strategy for standardization permits a degree of freedom in the design of encoding systems. Subsequently, there are a variety of encoder designs but also a correspondingly wide deviation in the picture quality for a given bit rate. Performing a feature check and determining whether an encoder can produce syntactically correct MPEG streams is easy. The difficulty arises in determining whether an encoder is efficient. — Although some users are now very comfortable with the application and selection of compression systems these users remain a strong minority. One of the fundamental problems with video encoding is that the underlying compression efficiency is very hard to measure. This issue is even more acute when running the encoder at very low data rates where the difference between encoders could mean substantially reduced bandwidth costs if the most efficient design had been selected. Although we are now seeing some admirable measurement devices that go some way to measuring picture quality, they can only tell part of the story. — This paper will explain how to augment objective measurements with additional tests and feature analysis. This should place the buyer in a better position to unzip the hype and fully evaluate the strengths and weaknesses of an encoder.

Published

1998-10

Content type

Original Research

DOI

10.5594/M00268

ISBN

978-1-61482-929-4