Automatic speech recognition system
The speech recognition is very widely used across for various applications in various electronic gadgets and systems. The major application is to support self service database look up using IVRS(Integrated Voice Response System). IVRS help callers search the information either via touch screen selection of menus or via speech commands. The other applications are as follows:
• Operators providing directory assistance is no longer needed and hence company can save on salaries.
• Employees providing extension number guidance is not needed with the incorporation of directories.
• Encouragement of using 'self service' option and provide speak with agent option with paid service.
• Provide feature such as speed assisted dial where in caller just speaks the name of the person whose number to be dialed while driving.
• Automated attendant feature.
How it works?
Figure depicts simple speech recognition system block diagram. As shown user speaks which is stored and decoded to the text based on acoustic,language models and gender.
• Speech recognition first detects and captures spoken words.
• It converts these into digital representation after noise removal. This is done using DSP algorithms.
• It breaks up sounds into smaller chunks if it has very large patterns.
• Now speech is assigned with particular phonemes based on records in the databases and based on probabilities.
Automatic speech recognition is basically independent computer driven conversion of spoken language into readable text in realtime environment.
Following are benefits of automativ speech recognition system.
• Cost reduction due to automating this system.
• Searchable text capability
• Provision to use by deaf.
Protocols
G.711,G.723.1,G.729,H.323,RTP,SIP are protocols and services used for speech recognition and Voice over IP(VOIP).