Audio Compression and Codecs

Unlike video data, audio and speech data is hard to compress effectively. This is because audio signals do not contain as much redundant information as video, namely temporal correlation between macroblocks in consecutive frames. Commonly used speech coding techniques may be classified into three classes - waveform, source and hybrid coding. Waveform coding is used at high bit rates. The resulting encoded signal has very good quality. Source coding, on the other hand, operates at very low bit rates. Thus, as expected, the resulting signal quality is degraded but intelligible, the signal sounds synthetic. A mixture between these two coding techniques is generally used which is the hybrid coding. Hybrid codecs give good quality at moderate bit rates. We discuss all these coding schemes as well as the most used audio codecs in this Section.