About Speech

Research work in the area of speech recognition has been carried out for many years, based around complex statistical models and predictive algorithms. These algorithms and models have been increasing in recognition accuracy at a rapid rate. In the past, extensive user training was required to "tune" the recognition engine to the individual voice of the user, and while this is still true today (albeit much reduced) of speech to text dictation systems, advances have meant that in many applications, this training element can be dispensed with.

Constant increases in microprocessor processing capabilities over the last 20 years, brought about as a result of Moores Law, have meant that the processing power available in todays desktop/laptop/handheld device, is thousands of times that which was available 20 years ago.

The convergence of these advances means that high accuracy levels are now attainable on a reliable basis opening up countless opportunities for greater/more cost effective forms of customer interaction.

To make an enquiry or to find out more click here