Product Overview

When using a device that uses voice, ambient noise may reduce the speech intelligibility and the recognition rate in voice recognition. There is some solution to use multiple microphones to emphasize the desired voice, but from the viewpoint of cost and installation space, there is a strong need to reduce the noise with a single microphone.

TMC has developed and commercialized a software IP that estimates the constant sound as noise and reduce it, improves the voice recognition rate by improving SNR (signal-to-noise ratio), and improves the voice intelligibility in conversation.

Model Number

• TM1819

Features

• No patent fees are required by taking advantages of using an original voice processing technology.
• Improvement of the voice recognition rate and the voice intelligibility in conversation, by the improved SNR.
• Processing with one microphone, by estimating the constant sound as noise and reducing the noise.
• Parameterized noise suppression, that enables to adjust according to the equipment.

Specification

Sampling Frequency	8kHz, 16kHz, 24kHz, 32kHz, 48kHz
Input / Output Data Format	16-bit linear PCM (monaural)
Strength of Noise Suppression	Adjustable in 9 steps

Processing Example