Alango OnlyVoice™ Technology Preserves Clear Voice
Hands-free communication via headphones is convenient, but only if the person on the other end of the call — or the voice assistant — hears and understands what’s communicated, even when speech originates from a noisy environment.
Consumers are attracted to the new breed of hearables and OEM’s are responding with sleek headphone designs such as the increasingly popular True Wireless Stereo “TWS” form-factor. Unfortunately, voice signal acquisition techniques employing microphone beamforming have reached their practical performance limit. Technological advancement of voice acquisition is necessary to improve communication reliability and customer satisfaction for sustainable market growth in the next generation of communication headphones.
Neural network-based noise reduction algorithms have the potential to improve communication performance but are presently unsuitable for implementation in ultra-low power chipsets because of the significant computational resources required. However, there exists the potential to employ advanced algorithmic noise reduction utilizing a combination of a traditional microphone beamforming array with additional signal sources such as an in-ear microphone or accelerometer. Alango has been engaged in developing internal sensor-based voice processing technology for the past three years.
OnlyVoice™ acoustic and bone conduction signal fusion
Alango OnlyVoice improves communication intelligibility in noisy environments by employing in-ear signal sources, which due to their positioning are isolated from ambient noise. User’s speech is acquiredwith either an in-ear microphone via the eustachian tube or an in-ear accelerometer via bone conduction. The in-ear signal is combined with the external microphone’s beamforming signal and intelligently mixed since the frequency limit of internal voice pick-up is effectively 1.5kHz. Hence, the combination of the in-ear sensor and external beamforming array is necessary to capture clear, full-spectrum speech. OnlyVoice breaks through the performance limitations of modern-day headphones. The functionality of OnlyVoice includes acoustic echo-cancellation (AEC), noise reduction (NR), multi-microphone beamforming, gain control (AGC, DRC), equalization (EQ) and special processing combining acoustic microphone(s) with a vibration sensor or in-ear microphone. Processing occurs on both the transmit (Tx) and receive (Rx) channels. OnlyVoice also includes Voice Activity Detection (VAD) functionality allowing low-power always-listening operation of voice-activated devices.