US 12,236,976 B2
Voice activity detection system and acoustic feature extraction circuit thereof
Meng-Ju Chiang, Tainan (TW); and Soon-Jyh Chang, Tainan (TW)
Assigned to NCKU Research and Development Foundation, Tainan (TW); and Himax Technologies Limited, Tainan (TW)
Filed by NCKU Research and Development Foundation, Tainan (TW); and Himax Technologies Limited, Tainan (TW)
Filed on May 2, 2024, as Appl. No. 18/653,676.
Application 18/653,676 is a division of application No. 17/411,735, filed on Aug. 25, 2021, granted, now 12,020,725.
Prior Publication US 2024/0282332 A1, Aug. 22, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. H03F 3/45 (2006.01); G10L 15/02 (2006.01); G10L 25/84 (2013.01)
CPC G10L 25/84 (2013.01) [G10L 15/02 (2013.01); H03F 3/45475 (2013.01)] 5 Claims
OG exemplary drawing
 
1. A voice activity detection (VAD) system, comprising:
an amplifier coupled to receive a sound signal, according to which an amplified signal is generated;
an acoustic feature extraction (AFE) circuit coupled to receive the amplified signal, and configured to generate a feature signal representing a feature extracted from the amplified signal;
a classifier configured to identify the feature signal as a voice or a noise;
wherein the AFE circuit comprises a single operational amplifier that is shared among a plurality of channels;
a plurality of band pass filters (BPFs) adaptable to the plurality of channels with different band-pass frequency ranges respectively for switchably receiving the amplified signal, thereby generating corresponding filtered signals; and a rectifier switchably coupled to receive the filtered signals, thereby generating a rectified signal; wherein the amplified signal is time-division demultiplexed onto the BPFs in different phases, and the filtered signals are time-division multiplexed onto the rectifier in different phases.