O'Reilly logo

Speech and Audio Signal Processing: Processing and Perception of Speech and Music, Second Edition by Dan Ellis, Nelson Morgan, Ben Gold

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

CHAPTER 34

image

MEDIUM-RATE AND HIGI RAT VOCODERS

34.1 INTRODUCTION

In Chapter 32 we were introduced to vocoders and the technique of analysis–synthesis systems. The channel vocoder was discussed in some detail. In Chapter 33, low-rate vocoders were presented that coded speech at rates below 2400 bps, with an expected sacrifice in sound quality. In this chapter1 we discuss speech coders that operate at bit rates greater than 2400 bps (typically 4800–16000 bps) but that can deliver more robust and higher-quality speech. This always involves some form of waveform coding in addition to the customary modeling of the vocal tract parameters. That is, under a source-filter model, while low-rate vocoders focus on the filter parameters, quality is improved at higher bitrates by improving the source description. We also discuss some pure waveform coding systems including subband and differential coding.

A variety of medium-rate systems have been proposed and implemented, including voice-excited systems, LPC residual-excited systems, split-band systems and, more recently, code-excited systems, employing analysis-by-synthesis methods. It should be noted that analysis by synthesis requires much more computational power than previous systems so that its use in real-time environments is a result of advances in high-speed computing.

34.2 VOICE EXCITATION AND SPECTRAL FLATTENING

Much of the effort ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required