| Summary: | 碩士 === 國立臺灣科技大學 === 工程技術研究所技職專業發展學程 === 84 ===
In general, the speech coder of low bit rate such as CELP, VSELP, etc. But they have coding delay of 50 to 60 ms, also coder don't against channel error. However, they can't fit telecommunication networks. To address this issue, a standard for 16 Kbps LD_CELP (Low_Delay CELP) with low_delay speech coding has been developed in 1992 by ITU.
This paper we focus on the robustness oflow_delay speech coding, example we enhance the robustness of codebook, and the robustness of lattice analysis or synthesis filter, etc. We propose a robust 16Kbps speech coder. The quality of synthesis speech is similar to 16 Kbps LD_CELP in channel noiseless, but which is better in having channel error. Also complexity and coding delay have lower. Which can work in BER (bit error rate) that is equal 0.001, 0.01 and 0.02 with communication channel.
|
|---|
