The core of Fortemedia's voice enhancement technology is called AMBIN ( Array Microphone Beamforming Integrated with Noise cancellation). Through using multiple microphones with advanced digital signal processing, AMBIN achieves unprecedented SNR enhancement with minimum voice signal distortion.

Figure 1 illustrates a noisy signal before and after AMBIN processing.

AMBIN technology leverages the strength of both small array adaptive beamforming, noise cancellation and none-linear frequency-domain filtering technologies, as illustrated in Figure 2. The front-end beamforming create two beams: the main beam focuses on the voice, and the reference beam senses the noise only. The reference signal is first used to adaptively cancel the linearly correlated noise in the main beam. The noise post-filter further suppresses the remaining noise. In this way, AMBIN can achieve higher noise suppression.

AMBIN Key Features

• Small array size. This makes it especially suitable for mobile communication applications.
• Narrow beam width, and therefore is capable to achieve high SNR enhancement.
• Strong echo rejection when the speaker is placed at the different direction of the voice. This makes it possible to achieve much high echo cancellation performance.
• Suppressing both stationary ambient noise such as car engine and road noise, and non-stationary noise such as radio or other people voices.
• Flexible configuration with one or multiple microphones for different applications.
• Good voice quality and small processing delay, making it suitable for wireless phone and speech recognition application.