The Voice Navigator (VN) combines software and technologies to create Interactive Voice Response (IVR) telephone information systems that automatically recognize and synthesize Russian speech.

Speech is our friendliest and most natural interface. Furthermore, it's useful in environments where eyes and hands are otherwise occupied, like driving or working on a computer.

Telephone information services are often limited by the number of operators available at a given call center. The best way to improve call center services, while cutting costs, is to integrate an IVR system that handles calls automatically. An IVR system is a human-computer dialogue system that answers customer inquiries using automatic speech recognition (ASR) and text-to-speech (TTS) capabilities.

IVR can be used to provide:

• Interactive ticket sales
• Scheduling information
• Weather forecasts
• Banking services
• Personal account information

Main Features

• Automatic speech recognition (Russian language) regardless of gender/individual voice characteristics
• Virtually unlimited recognizable vocabulary
• Robust and reliable speech recognition, natural-sounding speech synthesis
• Operates in realistic sound environments (can handle office or street noise)

How It Works

All modern IVR systems comply with the MRCP standard. This standard protocol integrates IVR system operation with ASR and TTS resources. Automatic Speaker Identity Verification (SIV) can also be implemented with version 2 of the MRCP. Depending on the task at hand, the resources can be used separately or jointly.

The IVR system handles all incoming calls according to the speech menu. The speech menu is a scripted dialogue that is used to accomplish a customer inquiry (e.g. find information on available flights for a certain date and destination). The system's responses are generated using a TTS resource (the MRCP server receives the query in SSML format). The customer's inquiries are recognized by an ASR resource in SRGS format. The generated speech output, as well as customer speech input for recognition/verification, are transferred with the RTP protocol.

TTS Features

• Natural-sounding synthesized speech (Russian language)
• Correct "reading" of numbers, acronyms, abbreviations and special symbols
• Transliteration of words in foreign languages
• Large morpho-grammatical dictionary (6 million word forms)
• Tailor-made voices can be developed by request

ASR features

• Takes into account special features of the Russian language
• Expanding dictionary of recognizable words
• Robust to noise

SIV Features

• 3-5 second speech sample is sufficient for verification
• Reliable verification over telephone

"Hear" for yourself! To test the quality of the speech recognition service for Russian, dial our demo +7 (812) 325-88-48 ext. 6227 in touch-tone mode.

Listen to a sample of Russian text-to-speech synthesis