We present a scaleable audio coder, suitable for both speech and music compression. Scalability has been implemented on both the bandwidth and bit rate. The encoder accepts 16 and 32 kHz input sampling frequencies and operates at 1 bit/sample. In addition to these rates, at bandwidths of 8, 12 and 16 kHz a multi-stage embedded quantisation is employed to produce a higher output quality at the expense of an increased bit rate. The design of the algorithm combines subband decomposition and a pulse excitation search offering low complexity, low algorithmic delay and good output quality when compared to existent coding standards
Published in:
Audio and Music Technology: The Challenge of Creative DSP (Ref. No. 1998/470), IEE Colloquium on
Date of Conference: 18 Nov 1998