• Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint
Share this Page URL
Help

Chapter 10. Speech Coding > MPEG-4 CELP Coding

10.3. MPEG-4 CELP Coding

Like the other coding tools in the MPEG-4 standard, MPEG-4 CELP is defined by a normative bitstream syntax and a corresponding decoding process, whereas an example encoder is described in an informative annex [MPEG4-3]. This approach leaves room for application-specific encoder optimizations even after the finalization of the standard. In contrast to this solution, other standard CELP speech codecs—for example, the ITU-T G.723 series [G723]—also specify a normative encoder. MPEG-4 CELP is specified in a modular way and therefore fits well with the MPEG-4 toolbox approach. One benefit of this approach is high flexibility with respect to the range of covered bit rates and sampling frequencies. Thus, it can operate in a narrowband mode, with 8-kHz sampling frequency, and in a wideband mode, with 16-kHz sampling frequency. Furthermore, MPEG-4 CELP provides additional functionalities besides pure data compression, including bit-rate scalability (see Section 10.3.6). In the following section, a short overview on the general MPEG-4 CELP encoder structure is given, followed by a detailed description of the various decoding tools. Finally, two special modes for scalable coding and for the so-called silence compression are presented.

10.3.1. CELP Encoder

The general structure of a CELP encoder [Atal84], as shown in Figure 10.3, can also be applied for MPEG-4 CELP. This structure processes the audio signal in blocks (frames) consisting of a fixed number of samples and is optimized for representing speech, as explained earlier.


PREVIEW

                                                                          

Not a subscriber?

Start A Free Trial


  
  • Creative Edge
  • Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint