Commitment to excellence
the common choice of users all around the world
Introduction
Cimon acoustics software provides customers with intelligent interaction capability of "listening, understanding, and speaking" in a variety of real application scenarios based on industry-leading and mass-produced proven technologies, and is widely used in smart homes, smart conference, smart education, smart healthcare, intelligent manufacturing, intelligent vehicle, security surveilliance, smart energy, smart courts, smart transportation and robotics, etc. to improve quality and experience for far field voice services.
Core Technology
Far-Field Acoustic Processing
Cutting-edge acoustic technologies, with completely independent intellectual property rights;Including echo cancellation, noise suppression, sound source localization, reverberation cancellation, beamforming, etc
Far-Field Keywrod Spotting
One of the top wake-up technologies, real-time response, make the machine hear human commands; Support Dual wake, offline voice wake-up, anti-noise voice wake-up, voiceprint authentication wake-up, Open AEC wake-up, etc., wake-up words customization, shortcut command words, etc
Far-Field ASR
One of the top ASR technologies, accurate, fast and stable, make the machine to understand human language; Cover a variety of scenarios such as local field, near-far-field, super far-field, and distributed field, match the vertical scenario interaction needs of office, home, travel, etc., support Free-ask, One-shot, VAN and other technologies
TTS
One of the top TTS technologies, provide the machine with the ability to speak; Provide TTS service that are more natural, more professional, and more in line with the needs of the scenarios. Support multiple voice customizations, customizable volume, speech rate, etc
Unique Technology
Single Microphone Noise Reduction
Microphone Array Noise Reduction
Self-defined Keywrod Spotting
Offline Keywrod Spotting
Offline Command Word Identification
InverseArray
Dualwake
Penetrated DOA
VAN
(Vertical Anti-Noise)
Anti-Noise Keywords Spotting
SoundAI Space Perception
SoundAI Space Awareness
OneShot
Free-cut
Free-ASK
TVNS
OpenAEC
VibratedAEC
Demo
Reverberation Suppression
De-reverberation Algorithm
For reverberation in room, De-reverberation algorithm that eliminates voice trailing
Effectively remove the phenomenon of speech tailing caused by room reverberation, enhance phonetic intelligibility, and elevate far-field speech recognition rate.
  • BEFORE
  • AFTER
  • Frequency

    HZ
    Time(s)
  • Frequency

    HZ
    Time(s)
  • Echo Cancellation
    Acoustic Echo Cancellation Algorithm
    For wake engine, deeply optimized echo cancellation algorithm
    Convergence time is less than 0.5s and echo suppression exceeds 35dB. Though under DoubleTalk condition, the algorithm still maintain its robustness and capability to optimize the wake engine of SoundAI which can effectively improve the recognition rate on music or TTS mode.
    • BEFORE
    • AFTER
  • Frequency

    HZ
    Time(s)
  • Frequency

    HZ
    Time(s)
  • Beamforming
    Beamforming Algorithem 1
    Beamforming algorithm for removing stationary noise like air conditioner and tire noise
    While the main direction voice remains undistorted, noises of air conditioner and tire are eliminated. The stationary noise suppression is more than 30dB.
    • BEFORE
    • AFTER
  • Frequency

    HZ
    Time(s)
  • Frequency

    HZ
    Time(s)
  • Beamforming
    Beamforming Algorithem 2
    Beamforming algorithm for removing non-stationary noise such as natural sounds and TV sound
    While the main direction voice remains undistorted, non-stationary noises as natural sounds and TV are eliminated. The non-stationary noise suppression is more than 20dB which means higher recognition rate under noisy condition.
    • BEFORE
    • AFTER
  • Frequency

    HZ
    Time(s)
  • Frequency

    HZ
    Time(s)
  • Effect Contrast of Duplex Talk Algorithm
    Suppress the scattering background noises and eliminate the echoes when both parties speak at the same time to make sure the other party can clearly hear you and can talk during that time, so as to maintain natural and clear auditory perception.
    • BEFORE
    • AFTER
  • Frequency

    HZ
    Time(s)
  • Frequency

    HZ
    Time(s)
  • ASR Demo
    With the help of algorithm of upfront-arranged microphone array which provides effective noise reduction and enhanced signal input, far-field voice recognition technology works to make sure that smart home devices, as well as intelligent hardware and robots, accurately capture user’s voice. The far-field voice recognition technology by SoundAI guarantees accuracy within a range of three to ten meters. The rate of precise keyword spotting is 95% and above, functionally supporting Dual-wake, Free-cut, Open AEC, and AKS. At a distance of five meters, the precise recognition is at least 96%, with reduced disturbances from any disordered voice wake-up. Moreover, our technology has been comprehensively programmed for improved compatibility and inclusion in specific scenarios, such as offices, households, and travelling. Leading technological revolutions in the field, SoundAI brings optimized user-experience and smoother voice interactions between humans and machines.
    Voice Print Recognition Demo
    Like your fingerprint, voiceprint is also biologically unique to every individual. Voice recognition is sometimes considered as the recognition of the person who makes the speech. As technologies developing, we expect to recognize more than one speech-makers / speakers in the future, which would not only secure and protect user privacy, but also enables personalized service to meet different preferences. SoundAI expands our technology scope for enriched informational analyses and in-depth services, supportive to recognitions based on voiceprint, age, mood, gender, humming, or even any abnormalities. Looking into the future, as acoustic technologies indeed will evolve to be more sophisticated and advanced, SoundAI aims to provide key solutions to challenges that users still encounter today in complicated scenarios.