Digital Signal Processing, in simple words, means the digitalization of signals. It can be broadly defined as the mathematical manipulation of digital signals. Digital signal processing is brought about with the help of microprocessor chips called Digital Signal Processors. With the use of these chips Digital signal processing can be implemented in a variety of applications, the size of the application not withstanding.
Telecommunication, speech processing are some of the areas where DSP is employed.
The implementation of DSP for speech processing has been carried out in the following fields: Speech analysis and synthesis, Speech coding, Speech and Speaker recognition and speech enhancement. Speech analysis involves extracting the parameters of speech signal and their storage. Speech synthesis uses the parameters obtained through speech analysis and generates speech. Speech coding is the actual digitalization of speech signal. The signal is sampled using the sampling theorem and quantised. In order to obtain the original the signal is decoded back. Speech recognition and speaker recognition uses both speech analysis techniques as well as speech coding techniques. Speech is used in applications where the speaker's words are to be interpreted and an operation is to be performed. Speaker recognition entails matching the speech patterns of the speaker with a database so as to decide whether the speaker's voice is familiar or not. Speech enhancement techniques are used to improve certain aspects of speech like quality, naturalness etc.
DSP applications in speech processing cover a wide area from telephony to public announcements and from business applications to personal communication at home. An example of how DSP applications can make life easier and simpler can be seen from the concept of a "Home Genie". The genie, i.e. a central processing unit with a DSP chip within, makes your home, work for you at your command enabling you to communicate with your home.
During the last few years, there have been sporadic and isolated attempts to apply DSP In several fields of areas of Speech processing. Here I present a conglomeration of several applications in an attempt to develop a home, which will literally be "at your command".
You wake up to the sweet voice of your alarm clock. The sweet voice is a result of advanced speech synthesis technique. You could select the voice that you liked best from a database of voices collected from volunteers all over the world.
On your way to the bathroom you give a voice command to the central processing unit placed conveniently in the center of your home. The message is relayed to your geyser, which heats up your water so that you can have a refreshing bath. Speaker recognition technique ensures that the unit accepts a command from its master only, as it recognizes your voice. You had previously trained your system so that it recognizes not only your voice but also the voices of your household members. The system has already stored your voices in a database. This database is then used to verify the speaker's identity. A microphone array around the unit implements speech enhancement so that the right order goes to the right appliance.
You would like to know what's going on around in the world. A simple command to your central processing unit switches on your television set. That's not all you can even listen to news from various websites on your computer. This is possible using text to speech synthesis technique. Next on your agenda is the checking of e-mails. But you have to arrange your briefcase before you reach your office. Here, again text to speech synthesis comes handy. Your computer reads out your mails for you.
You are ready to leave your home to begin your day at work. Before leaving for work you would like to know all your appointments and important meetings for the day. Your digital diary doubles up as your personal assistant and reads out your schedule. You are worried about the safety of your home. But a single command to your DSP chip inside the central processing unit locks all your windows and safety locks on your cupboards. After you step out of your home the front door is locked on your command. The command is interpreted using speech recognition and speech analysis. You leave for your office peacefully as you know that the system will fail to recognize any other person's voice and would not open the door for any one other than yourself. As mentioned earlier it could be trained to recognize the voices of your family members.
You are back home after a day's hard work. You are so exhausted that you will feel glad if someone switches on the lights and fans for you. Thanks to Speech coding methods the fans and lights can hear your wish and immediately turn on for your convenience.
The above applications are just few of the possibilities presented by the use of speech processing. The various speech processing techniques can be used to make your life easier and simpler whether it be at your home or at workplace. It can provide a strong platform for the personal communicator of the future.