| The embedded Linux operating system has many advantages such as stability, efficiency, wide hardware support and free, open codes provided. DSP has great advantage in subsequent real time voice processing because of its special system framework. With the development of DSP, embedding technique and Internet etc, the embedded voice transmitting technique based on DSP will be widely applied for its low power consumption and great performance.Voice recorder is greatly applied in safety and surveillance aspects of railway, electric power, police and financial systems. Traditional voice recorder was implemented by stimulation, but with behindhand techniques or high cost. In this thesis, we introduce recent voice processing and transmitting techniques based on DSP in an embedded Linux environment. We realize an embedded multi-channel digital voice recorder based on AC48304 digital signal processor, ARM 9 CPU in an embedded Linux operating system. This powerful device has very low cost and many functions such as voice coding and decoding, DTMF generation and detection, voice validity detection, echo cancellation and etc. During the development of the voice recorder, the optimization of the driver and analyses of the ability of real time processing in Linux environment have been well made. We also adopted a RTP protocol during network transmission to guarantee the voice quality.In the rest of the thesis, we study the currently popular protocols of VOIP and SIP, and make a general research on expanding the voice recorder to a voice transmitting and processing platform. We also discuss how to apply the voice transmitting and processing platform to education systems based on SIP and VOIP.The network multi-channel digital voice recorder we developed is useful and has a low cost, thus has high practical value and market potential. The voice transmitting and processing platform is also advanced and has a wide usage. |
PDF Full Text Request