Home  /  Partners  /  CEVA 3rd Party Partners  /  Malaspina Labs

Malaspina LabsPrint this page

CEVA Software / Algorithm Partner

VoiceBoost™ Noise Reduction and ASR Assist


Malaspina Labs advances the state of the art in real-time speech enhancement for ultra-low power devices. Malaspina Labs offers fully supported commercial implementations of our proprietary Model Based Speech Discrimination algorithms which improve both person-to-person and person-to-device voice communication in high-noise environments. Malaspina Labs focuses on the development of speech processing algorithms which operate deterministically at bone-conduction, telecom, and voice interface latencies within the resource constraints of milliwatt-class processors designed for mobile, embedded, medical and consumer devices.

CEVA DSP Cores Supported

Product Offering

Malaspina’s VoiceBoost™ signal processing blocks go beyond traditional filter-based noise reduction approaches by employing a mathematical model of the human voice to allow the detection, discrimination and isolation of speech-of-interest from high-volume competing background noise including high volume babble speech, traffic and impulse noise. Malaspina’s VoiceBoost™ signal processing blocks are rotation & orientation robust and are effective in enhancing speech within near- field and arms-length audio signals from single omni-directional microphones or directional microphones, as well as spatially separated input signals from multiple microphones. Configurations can be built which operate in as little as 8 msec latency and under 5 MIPS CPU. Malaspina's speech discrimination technologies are available in source code or binary form for integration into existing audio processing pipelines and have been shown to increase speech intelligibility by hearing impaired users in noisy environments, as well as to increase automatic speech recognition performance and mobile voice call quality in noisy environments