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

11.7. Summary

This chapter described the technology the MPEG-4 Audio standard provides for the coding of general audio signals (general audio coding), as opposed to specific signal classes, such as speech signals. Most of the MPEG-4 coding tools for general audio coding adhere to the paradigm of filterbank-based (T/F) coding of spectral coefficients in general and MPEG AAC in particular. Introduced within MPEG-2 Audio, the AAC coder is a powerful multichannel coder with a rich set of coding tools that allow the coder to adapt flexibly to arbitrary input signals. Building on this core technology, MPEG-4 defines a number of extensions to further enhance compression performance and support new modes of operation and functionalities.

The PNS tool achieves a compact parametric coding of noiselike signal components. LTP exploits the redundancy of stationary signals and effectively replaces the MPEG-2 AAC prediction tool at about one-half of its complexity. The TwinVQ coding kernel can be used to replace the standard AAC quantization/coding tools to enable operation at extremely low bit rates (down to 6 kbit/s/ch), which is of particular interest in the context of scalable audio coding. The AAC-LD coder enables the coding of multichannel general audio signals with algorithmic delays as low as 20 ms, thus fulfilling the requirements for high-quality two-way communication applications. Improved performance under error-prone transmission conditions is achieved by using the ER tools.


PREVIEW

                                                                          

Not a subscriber?

Start A Free Trial


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