3.6 LOOKING AHEAD

Chapter 3 has examined how information sources—particularly speech—may be represented efficiently by exploiting redundancy in the source itself. Source coding removes statistical and perceptual redundancy so that the transmitted bit stream contains as little superfluous information as possible. In doing so, it reduces the required bit rate for a given level of intelligibility or quality.

However, this very efficiency introduces a new vulnerability. By stripping away redundancy, source coding leaves little margin for error: even a small number of bit errors introduced by noise, fading, interference, or jamming can produce disproportionate degradation in reconstructed speech. A highly compressed parametric bit stream is often more sensitive to channel impairments than an uncompressed waveform representation. Thus, while source coding minimizes bandwidth, it also increases the importance of protecting the transmitted bits against transmission errors.

This leads directly to the next fundamental building block of a communication system: channel coding. Whereas source coding removes redundancy to improve efficiency, channel coding deliberately adds structured redundancy to enable the detection and correction of errors introduced by the transmission medium. The apparent tension between removing redundancy (for efficiency) and adding redundancy (for reliability) lies at the heart of modern digital communications and reflects a deeper theoretical framework established by Shannon’s separation principle.

Chapter 4 therefore shifts the focus from efficient representation of information to reliable transmission over imperfect channels. It introduces error-detecting and error-correcting codes, examines their performance in noisy environments, and explains how carefully designed redundancy allows communication systems to approach the theoretical limits of reliability. Together, source coding and channel coding form complementary components of the digital communication chain: one shapes the information for efficiency, the other protects it for integrity across the channel.