Research On Unified Speech And Audio Coding Algorithm

Recent years various key technologies of high-quality and low bit-rate audio and speech coding algorithms have been developed. Most of them focus on either audio or speech, separately, but in contrast, unified speech and audio coding is a little far behind. With the development of internet and the emergence of high quality multimedia applications like video conference, however, requirements of efficiently encode both speech and audio signals with one unified codec are keep increasing. So it is of great significance to do some researches on these areas.This paper presents a unified speech and audio algorithm framework. Based on the best audio coder EAAC+, some key modules are modified to meet the need of processing speech at low bit-rate. A new kind of high frequency reconstruction algorithm High Frequency Replication (HFR) is presented, which replace the SBR modules of original EAAC+coder. For audio signal, the original EAAC+coder is used without modifications.The performances of the innovation schemes have shown some improvements and advantages for unified speech and audio coding scheme. Even not always good in any cases, it still has some potential feathers, so further study and investigations about these issues are expected.