Voice Methods | Voice Methods Explained

Speech Recognition

Our speech recognition systems may be integrated into other technology in a variety of ways: from voice-operation in handheld devices to public information systems. Low memory and electricity consumption requirements were achieved using the latest mathematical apparatus and linguistic know-how of Ectaco, Inc. Linguistic research shows that human beings are capable of producing not more than approximately 1000 different sounds. The quantity of sounds used varies from language to language. Using this knowledge allowed us to create a universal, speaker (and language) independent system with a recognition accuracy of approximately 98%.
Our SR solution is compact and was developed and optimized for implementation in portable devices and embedded systems.

Our technology is totally speaker-independent, i.e. it recognizes anyone’s voice, irrespective of sex, age and accent
Quality recognition of the following is stipulated	Phrases Names Numbers Separate letters of the alphabet (spelling)
Quality attributes	Accuracy not less than 90% (for less than 300 names)

Please, contact us for more information