手机语音识别应用中dsp的选择策略（Selection strategy of DSP in mobile phone speech recognition applications）.docVIP

手机语音识别应用中dsp的选择策略（Selection strategy of DSP in mobile phone speech recognition applications） With the development of DSP technology, more computing power, lower power consumption and smaller size of DSP have appeared, which makes it possible to implant more accurate and complex automatic speech recognition (ASR) functions on 3G handsets. At present, the basic ASR applications can be divided into three categories: 1., text to text conversion (voice input); 2. speaker identification; 3. voice command control (voice control). These three types of functions include the many ASR capabilities required by the 3G. Typical examples of voice to text conversion are voice dialing and e-mail dictation. Speaker recognition enables secure reading of personal data in memory by voice recognition, which meets the need for high security applications such as credit card ordering and banking services. Voice command control features include voice extensions, markup language (VXML), and voice interfaces for web content, which support services such as financial services and directory assistants. Currently, VXML is used to standardize speech tags for web content. The two methods of speech recognition 3G mobile phone ASR application design can be divided into two categories, namely terminal centric and client / server centric applications. As shown in Figure 1, for a terminal centric design approach, the 3G mobile phone (terminal) performs the whole speech recognition process and sends the identification results. In Figure 2 the client / server method, terminal only performs preprocessing feature extraction, and then through a data channel error protected these parameters will be sent to a central server, the server center completed the final speech recognition. If the design method of the center of the client / server, 3G mobile phone use data channel rather than the mobile channel will be sent to a server for speech recognition, because of low bit rate speech encoding used by the mobile ch