Methods and Voice Activity Detectors for a Speech Encoders

Abstract

Voice activity detectors and related methods are provided. Methods include receiving a frame of the input signal; determining a first SNR of the received frame; comparing the determined first SNR with an adaptive threshold; and detecting whether the received frame comprises voice based on the comparison. The adaptive threshold is at least based on total noise energy of a noise level, an estimate of a second SNR and on energy variation between different frames.

1 . (canceled) 
     
     
         2 . A computer-implemented method, in a voice activity detector, for determining whether frames of an input signal comprise voice comprising:
 receiving a frame of the input signal;   determining a first signal-to-noise-ratio (SNR) of the received frame by:
 obtaining a plurality of subband SNR values of the received frame by dividing energy levels of each of a plurality of subbands of the received frame by its respective subband background energy; 
 applying subband specific significance thresholds to each of the plurality of subband SNR values of the received frame through selectively adjusting the plurality of subband SNR values using a non-linear function; and 
 obtaining the first SNR by summing together all non-linearly adjusted SNR values of each of the plurality of subbands; 
   comparing the determined first SNR with an adaptive threshold,
 wherein the adaptive threshold is at least based on an estimate of a second SNR and an estimate of energy variation between different frames of the input signal, 
 wherein the second SNR is a long-term SNR, and 
 wherein the estimate of energy variation between different frames has limits on how quickly the estimate increases such that the estimate may not increase beyond a fixed constant for each frame; and 
   detecting whether the received frame comprises voice based on the comparison.   
     
     
         3 . The method of  claim 2 , wherein the energy variation between different frames is the energy variation between the received frame and a last received frame which did not comprise voice. 
     
     
         4 . The method of  claim 2 , wherein the second SNR is measured over a plurality of frames of the input signal. 
     
     
         5 . The method of  claim 2  wherein comparing the determined first SNR with the adaptive threshold comprises adjusting the estimate of the second SNR upwards responsive to a determination that the estimate of the second SNR is lower than a smooth input dynamics measure, wherein the smooth input dynamics measure is indicative of energy dynamics of the received frame. 
     
     
         6 . The method of  claim 5 , wherein the estimate of the second SNR is adjusted upwards to a value which is less than or equal to the smooth input dynamics measure. 
     
     
         7 . The method of  claim 6 , wherein the smooth input dynamics measure for the received frame comprises a high/max energy tracker, a low/min energy tracker, and a portion corresponding to a smooth input dynamics measure for prior frames. 
     
     
         8 . The method of  claim 7 , wherein, prior to calculating the smooth input dynamics measure for the received frame, the low/min energy tracker is set to either the total energy of the received frame or an incremental upward adjustment of the low/min energy tracker, whichever is smaller, and
 wherein, prior to calculating the smooth input dynamics measure for the received frame, the high/max energy tracker is set to either the total energy of the received frame or a decremental downward adjustment of the high/max energy tracker, whichever is greater.   
     
     
         9 . The method of  claim 7 , wherein the smooth input dynamics measure is a function of a difference between the high/Max energy tracker based on a highest frame energy value over a plurality of frames and the low/mm n energy tracker based on a lowest frame energy value over the plurality of frames. 
     
     
         10 . A voice activity detector for determining whether frames of an input signal comprise voice, the voice activity detector comprising:
 an input circuit configured to receive a frame of the input signal; and   a processor configured to:
 determine a first signal-to-noise-ratio (SNR) of the received frame by:
 obtaining a plurality of subband SNR values of the received frame by dividing energy levels of each of a plurality of subbands of the received frame by its respective subband background energy; 
 applying subband specific significance thresholds to each of the plurality of subband SNR values of the received frame through selectively adjusting the plurality of subband SNR values using a non-linear function; and 
 obtaining the first SNR by summing together all non-linearly adjusted SNR values of each of the plurality of subbands; 
 
 compare the determined first SNR with an adaptive threshold,
 wherein the adaptive threshold is at least based on an estimate of a second SNR and an estimate of energy variation between different frames of the input signal, 
 wherein the second SNR is a long-term SNR, and 
 wherein the estimate of energy variation between different frames has limits on how quickly the estimate increases such that the estimate may not increase beyond a fixed constant for each frame; and 
 
 detect whether the received frame comprises voice based on the comparison. 
   
     
     
         11 . The voice activity detector of  claim 10 , wherein the energy variation between different frames is the energy variation between the received frame and a last received frame which did not comprise voice. 
     
     
         12 . The voice activity detector of  claim 10 , wherein the second SNR is measured over a plurality of frames of the input signal. 
     
     
         13 . The voice activity detector of  claim 10  wherein comparing the determined first SNR with the adaptive threshold comprises adjusting the estimate of the second SNR upwards responsive to a determination that the estimate of the second SNR is lower than a smooth input dynamics measure, wherein the smooth input dynamics measure is indicative of energy dynamics of the received frame. 
     
     
         14 . The voice activity detector of  claim 13 , wherein the estimate of the second SNR is adjusted upwards to a value which is less than or equal to the smooth input dynamics measure. 
     
     
         15 . The voice activity detector of  claim 13 , wherein the smooth input dynamics measure for the received frame comprises a high/max energy tracker, a low/min energy tracker, and a portion corresponding to a smooth input dynamics measure for prior frames. 
     
     
         16 . The voice activity detector of  claim 15 , wherein, prior to calculating the smooth input dynamics measure for the received frame, the low/mm energy tracker is set to either the total energy of the received frame or an incremental upward adjustment of the low/min energy tracker, whichever is smaller, and
 wherein, prior to calculating the smooth input dynamics measure for the received frame, the high/max energy tracker is set to either the total energy of the received frame or a decremental downward adjustment of the high/max energy tracker, whichever is greater.)   
     
     
         17 . The voice activity detector of  claim 15 , wherein the smooth input dynamics measure is a function of a difference between the high/max energy tracker based on a highest frame energy value over a plurality of frames and the low/min energy tracker based on a lowest frame energy value over the plurality of frames.