本文選題：水下通信 + 語音編碼��； 參考：《哈爾濱工程大學(xué)》2014年碩士論文

【摘要】：語音作為一種有效的信息溝通手段,是通信傳輸中的主要信源。水下語音通信由于受到水聲信道的限制成為了水聲通信應(yīng)用中最富挑戰(zhàn)性的研究課題之一。在可利用的頻帶資源內(nèi)最大限度挖掘頻帶資源,并最大限度的降低語音信號(hào)的通信速率,以獲得高質(zhì)量的語音通信成為了當(dāng)前人們的研究熱點(diǎn)�；旌霞�(lì)線性預(yù)測(cè)(Mixed Excitation Liner Prediction, MELP)編碼具有較高的語音壓縮率和較好的保密性能,并具有很高的實(shí)用價(jià)值。本文應(yīng)用混合激勵(lì)線性預(yù)測(cè)編碼算法作為水下通信系統(tǒng)的語音信號(hào)處理模塊。并將重點(diǎn)放在了如何在現(xiàn)有的傳輸數(shù)據(jù)率下提高合成語音的質(zhì)量。本文首先對(duì)混合激勵(lì)線性預(yù)測(cè)低速率語音編碼算法的流程進(jìn)行了分析,給出了包括基音周期、碼本、線性預(yù)測(cè)系數(shù)等的仿真結(jié)果。介紹了從編碼端語音信號(hào)的參數(shù)提取到解碼端合成語音的過程。進(jìn)行了混合激勵(lì)線性預(yù)測(cè)編碼的仿真,重點(diǎn)研究了在不同場(chǎng)景下算法對(duì)干擾的容錯(cuò)能力,分析結(jié)果表明算法自身對(duì)干擾具有一定的容忍能力。驗(yàn)證了該算法在通信中,存在誤碼的條件下依然具有一定的穩(wěn)健性,明確了編碼端發(fā)送幀中各參數(shù)的重要性。據(jù)此可以對(duì)編碼端發(fā)送幀的各參數(shù)的重要程度進(jìn)行排序,對(duì)更為關(guān)鍵的信息予以保證。其次,針對(duì)水下語音通信的特點(diǎn),設(shè)計(jì)采用預(yù)加重技術(shù)和子空間語音增強(qiáng)技術(shù)對(duì)混合激勵(lì)線性預(yù)測(cè)編碼算法進(jìn)行改進(jìn)以提高語音合成效果。再次,為進(jìn)一步降低語音編碼數(shù)據(jù)率,利用自適應(yīng)差分思想對(duì)算法的編碼和量化部分進(jìn)行改進(jìn),仿真結(jié)果表明該算法可以在保證部分語音恢復(fù)質(zhì)量基礎(chǔ)上降低編碼數(shù)據(jù)率。最后,根據(jù)DSP實(shí)時(shí)編碼需求以及所需的資源條件對(duì)算法進(jìn)行了配置與優(yōu)化。根據(jù) TMS320C6713 DSK 中的 DSP/BIOS,EDMA 和 McBSP 等硬件資源,采用 AIC23實(shí)現(xiàn)對(duì)語音信號(hào)的數(shù)據(jù)采集部分以完成硬件平臺(tái)的搭建。調(diào)整程序中寄存器的分配及占用大小等,以實(shí)現(xiàn)對(duì)該算法的優(yōu)化。最終將C語言算法移植到TMS320C6713DSK平臺(tái)上,實(shí)現(xiàn)混合激勵(lì)線性預(yù)測(cè)語音低速率編碼。
[Abstract]:As an effective means of communication, voice is the main source of communication.Underwater voice communication has become one of the most challenging research topics in underwater acoustic communication applications due to the limitation of underwater acoustic channels.In order to obtain high quality voice communication, it has become a hot research topic to mine the frequency band resource and reduce the communication rate of speech signal to obtain high quality speech communication in the available frequency band resources.Mixed Excitation Liner prediction (MELPPC) coding with mixed excitation has higher speech compression ratio and better security performance, and has high practical value.In this paper, the hybrid excitation linear predictive coding algorithm is used as the speech signal processing module of underwater communication system.Emphasis is placed on how to improve the quality of synthetic speech at the existing data rate.In this paper, the flow of mixed excitation linear prediction low rate speech coding algorithm is analyzed, and the simulation results including pitch period, codebook and linear prediction coefficient are given.This paper introduces the process of extracting the parameters of speech signal from the coding end to synthesizing the speech at the decoding end.The simulation of mixed excitation linear predictive coding is carried out, and the fault-tolerant ability of the algorithm to interference in different scenarios is studied. The analysis results show that the algorithm has a certain tolerance for interference.It is verified that the algorithm is robust under the condition of error code in communication, and the importance of the parameters in the transmission frame at the coding end is clear.Based on this, the importance of each parameter of the frame can be sorted, and the more critical information can be guaranteed.Secondly, according to the characteristics of underwater speech communication, the preweighting and subspace speech enhancement techniques are designed to improve the hybrid excited linear predictive coding algorithm to improve the effect of speech synthesis.Thirdly, in order to further reduce the speech coding data rate, the adaptive difference idea is used to improve the coding and quantization part of the algorithm. The simulation results show that the algorithm can reduce the coding data rate on the basis of guaranteeing the partial speech recovery quality.Finally, the algorithm is configured and optimized according to the requirement of DSP real-time coding and the necessary resource conditions.According to the hardware resources such as DSP / BIOSMA and McBSP in TMS320C6713 DSK, the data acquisition part of speech signal is realized by AIC23 in order to build the hardware platform.In order to optimize the algorithm, the allocation and occupation size of registers in the program are adjusted.Finally, C language algorithm is transplanted to TMS320C6713DSK platform to realize mixed excitation linear predictive speech low rate coding.
【學(xué)位授予單位】：哈爾濱工程大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TN912.3

【參考文獻(xiàn)】

中國(guó)期刊全文數(shù)據(jù)庫(kù) 前10條

1 曹芳;趙繼勇;劉亞峰;;2.4kbps MELP低速率語音編碼算法的DSP實(shí)現(xiàn)[J];科技信息;2011年04期

2 李洋;吳玉成;;MELP語音編解碼算法在DSP上的優(yōu)化[J];電聲技術(shù);2010年05期

3 楊光宇;高曉蓉;王黎;王澤勇;;基于TI C6000系列DSP的C/C++程序優(yōu)化技術(shù)[J];現(xiàn)代電子技術(shù);2009年08期

4 張博;朱杰;;基于MELP的變速率語音編碼器算法[J];電子測(cè)量技術(shù);2009年02期

5 郭立;王妙鋒;劉璐;郁理;李琳;;1.6Kb/s類MELP語音壓縮編碼器的FPGA實(shí)現(xiàn)[J];小型微型計(jì)算機(jī)系統(tǒng);2008年08期

6 范晶;和應(yīng)民;;一種改進(jìn)的MELP算法[J];應(yīng)用科技;2008年05期

7 戚銀城;張巍;苑津莎;;超低速率MELP語音編碼算法研究[J];聲學(xué)技術(shù);2007年06期

8 陳華;關(guān)宇東;王健;;一種短波通信中基于DSP的低速率語音編碼技術(shù)[J];應(yīng)用科技;2007年08期

9 郝軟層;徐金甫;;基于DSP芯片的MELP聲碼器的算法實(shí)現(xiàn)[J];微計(jì)算機(jī)信息;2006年29期

10 蘇桃;;淺談?wù)Z音壓縮編碼的發(fā)展和應(yīng)用[J];科技情報(bào)開發(fā)與經(jīng)濟(jì);2006年22期

中國(guó)碩士學(xué)位論文全文數(shù)據(jù)庫(kù) 前9條

1 白夜;蛙人語音編碼技術(shù)及其DSP實(shí)現(xiàn)[D];哈爾濱工程大學(xué);2013年

2 石亞楠;基于MELP的低速率語音編碼算法研究[D];西安電子科技大學(xué);2010年

3 許明;低速率語音編碼參數(shù)高效量化算法研究[D];清華大學(xué);2009年

4 陳曉紅;基于Ad Hoc網(wǎng)絡(luò)的海上VHF通信系統(tǒng)研究[D];大連海事大學(xué);2008年

5 劉媛媛;基于DSP的MELP語音編碼算法研究及實(shí)現(xiàn)[D];哈爾濱工程大學(xué);2008年

6 陳輝;MELP語音編碼的研究及其在DSP上的優(yōu)化[D];浙江大學(xué);2006年

7 舒小華;TETRA協(xié)議中語音壓縮編譯碼算法研究及其DSP實(shí)現(xiàn)[D];西北工業(yè)大學(xué);2005年

8 鄭魯杰;低速率語音編碼MELP算法的研究[D];電子科技大學(xué);2003年

9 劉強(qiáng)國(guó);語音壓縮編解碼系統(tǒng)的硬件設(shè)計(jì)與實(shí)現(xiàn)[D];電子科技大學(xué);2003年

，

本文編號(hào)：1733098