Home > Solutions > Edge AI Speech Algo > Front-end acoustic processing > NR

NR (Noise Reduction)

Definition:

Noise Reduction filters ambient noise through hardware or algorithms to preserve target audio (such as vocals, musical instruments) and improve signal-to-noise ratio.

Classification and technical characteristics:

Type	Principle	Typical Scene
passive noise reduction	Physical sound insulation (such as headphone sound insulation cotton)	Fixed noise scenes such as airplanes and subways
Active Noise Reduction (ANC)	Generate inverse sound waves to counteract ambient sounds (e. g. sine wave noise)	Headphones, industrial noise-canceling earmuffs
algorithm noise reduction	Filter noise bands by Fourier transform etc.	Recording studio, voice call pre-processing
Application cases:

Wireless microphone: In noisy environments such as exhibitions and stages, filter background vocals and equipment current sounds to ensure that the speaker's voice is clear.

Online education: eliminate the fan sound and pet sound of students through algorithm noise reduction, and optimize the classroom audio quality.

Technical value:

Reduce the impact of environmental interference on the target audio, improve the accuracy of speech recognition (such as the false trigger rate of wake-up words) and the intelligibility of audio content.

NR

NR (Noise Reduction)

Definition:

Application cases:

Technical value: