The AI-powered noise cancellation is trained through deep neural network (DNN), the model is generated with more than 500 million data, simulate the way human brain works. When a sound is heard, the model automatically recognizes and separate vocal sound from background noises; hence, filtering out all the unwanted Noise. This feature greatly improves the user’s communication experience.
For "Microphone (MSI Sound Tune)" – Talking without noise
Remove background noise from the microphone, so the other call participants can hear your voice clearly
For "Speaker (MSI Sound Tune)" – Listening without noise
Remove background noise from incoming audio, so you can always hear their voice clearly
The human brain can identify if a sound is vocal or non-vocal even if it is a language unknown to the
brain, the hearing system can also concentrate on vocal sounds and ignore the surrounding noise. MSI
Sound Tune is based on this concept to break the traditional boundaries on noise reduction function.
The traditional noise reduction need to assume the state of the noise it is reducing, causing the
noise reduction on unstable noise to be poor, this would include the noise from people typing on
keyboard, the sound from an construction site, or the sound when people whispers in a coffee shop,
sometimes the noise may even surpass the vocal sounds. This is because, the traditional noise
reduction technology is based on the knowledge of signals, using manually compiled algorithm, can
only suppress noise at a stable level or when the SNR is relatively high. The MSI Sound Tune on the
other hand can extract the vocal sound under complicated noise environment, allowing the user to
have a smooth call in different environment, and make sure the message are delivered clearly.