Qualitative Analysis and Enhancement of Sine Transform Coding

In the past several years, significant progress has been made in the STC (sine transform coder) approach in terms of compression gain and implementation complexity [1]–[3]. As a result, it now appears to be a viable alternative to the CELP in the range of 2.4 kbps to 9.6 kbps. In contrast to the CELP, which is a timedomain, search-oriented technique, the STC compresses speech directly in the frequency domain using a harmonic model without decomposing it into excitation and time-varying filter parts. Because of its simplicity in speech modelling, at least in principle, it provides a tractable means of gaining insight into the process of speech compression.