Multi-state echo suppressor

Abstract

An echo suppressor uses a multi-state operating sequence to provide non-disruptive monitoring of a remotely broadcast program during a performer's live performance at a local site. The operating sequence is actuated by a voice activated switch that initiates signal processing for correlating a locally originating audio (e.g., the performer's voice) to its echo in the broadcast program. During correlation states in the operating sequence, the echo suppressor switches monitoring to a local microphone signal to allow the performer to hear his or her own voice while preventing echo from being heard before correlation is achieved and echo suppression can begin. The correlation states include an initial correlation at a reduced sampling rate to detect echo within a wide delay time window and determine the locally originating audio is on-the-air. A subsequent state then performs a second correlation at a higher sampling rate to correlate to the echo in a narrower delay time window displaced according to the approximate delay determined in the initial correlation.

1 . An echo suppressor in an audio signal transmission system in which an audio signal originating at a local site is transmitted on the system and a return audio signal that potentially contains echo of the local audio signal is received from the system, the echo suppressor operating at the local site and comprising: 
 a first input for said locally originating audio signal;    a second input for said return audio signal received at the local site;    an output for audio monitoring at the local site;    a signal processor operating to process said locally originating audio signal and said return audio signal for correlating said locally originating audio signal to said echo in said return audio signal and suppressing said echo to produce an echo suppressed return audio signal, and to provide the echo suppressed return audio signal at the output for monitoring at the local site; and    a voice activated switch operating in response to voice activity in the locally originating audio signal to switch from providing said return audio signal at the output to providing said locally originating audio signal at the output until correlation is achieved.    
     
     
         2 . An echo suppressor in an audio signal transmission system in which an audio signal originating at a local site is transmitted on the system and a return audio signal that potentially contains an echo of the locally originating audio signal is received from the system at the local site, the echo suppressor operating at the local site and comprising: 
 a first input for said locally originating audio signal;    a second input for said return audio signal received at the local site;    an output for audio monitoring at the local site; and    a signal processor operating to process said locally originating audio signal and said return audio signal for correlating said locally originating audio signal to said echo in said return audio signal and suppressing said echo to produce an echo suppressed return audio signal, and to provide the echo suppressed return audio signal at the output for monitoring at the local site;    the signal processor further operating to initially perform a reduced sampling rate correlation of said locally originating audio signal to said echo in said return audio signal to identify an approximate delay of said echo from said locally originating audio signal within a wider window of time, and to perform a full sampling rate correlation of said locally originating audio signal to said echo in said return audio signal within a narrower window of time defined according to the approximate delay for use in suppressing said echo from said return audio signal.    
     
     
         3 . The echo suppressor of  claim 2  wherein the signal processor additionally operates to store parameters of a prior correlation, and to perform a second full sampling rate correlation based on the stored parameters concurrently with the reduced sampling rate correlation, so as to more quickly correlate said locally originating audio signal to said echo when the audio signal transmission system causes a consistent delay of said echo.  
     
     
         4 . The echo suppressor of  claim 3  further comprising: 
 data storage for retaining parameters of a plurality of prior correlations of said locally originating audio signal to said echo as plural correlation presets;  
 controls for actuating by a user to select from the correlation presets; and  
 the signal processor being operative to perform the second full sampling rate correlation based on the parameters of the selected correlation preset.  
 
     
     
         5 . The echo suppressor of  claim 3  further comprising: 
 data storage for retaining parameters of an immediately preceding successful correlation of said locally originating audio signal to said echo; and  
 the signal processor being operative to perform the second full sampling rate correlation based on the retained parameters of the immediately preceding successful correlation.  
 
     
     
         6 . An echo suppressor in an audio signal transmission system in which an audio signal originating at a local site is transmitted on the system and a return audio signal that potentially contains an echo of the locally originating audio signal is received from the system at the local site, the echo suppressor operating at the local site and comprising: 
 a first input for said locally originating audio signal;    a second input for said return audio signal received at the local site;    an output for audio monitoring at the local site;    a signal processor operating to process said locally originating audio signal and said return audio signal for correlating said locally originating audio signal to said echo in said return audio signal and suppressing said echo to produce an echo suppressed return audio signal, and to provide the echo suppressed return audio signal at the output for monitoring at the local site;    the signal processor further operating to simultaneously perform a reduced sampling rate correlation of said locally originating audio signal to said echo in said return audio signal to identify an approximate delay of said echo from said locally originating audio signal within a wider window of time, and to perform a full sampling rate correlation of said locally originating audio signal to said echo in said return audio signal within a narrower window of time, the full sampling rate correlation being preset according to stored parameters of a prior successful correlation of said locally originating audio signal to said echo in said return audio signal.    
     
     
         7 . A method of suppressing echo of a locally originating audio signal from a return transmission audio signal, the method comprising the steps of: 
 detecting voice activity in the locally originating audio signal;    processing the locally originating audio signal and return transmission audio signal according to a multi-state operating sequence having at least one correlate state during which the processing correlates the locally originating audio signal to echo in the return transmission audio signal, and an echo suppress state in which the processing suppresses echo in the return transmission audio signal to produce an echo suppressed audio signal;    selectively switching among at least the locally originating audio signal and the echo suppressed audio signal for output to a monitoring device depending on a current state in the multi-state operating sequence;    causing a transition to a first of the at least one correlate state in response to the presence of voice activity in the locally originating audio signal being detected; and    switching to output of the locally originating audio signal in the first of the at least one correlate state.    
     
     
         8 . The method of  claim 7  further comprising the step of: 
 switching to output of the echo suppressed audio signal in the echo suppress state.  
 
     
     
         9 . The method of  claim 8  further comprising the step of: 
 switching to output of a combination of the locally originating audio signal and the echo suppressed audio signal in the echo suppress state.  
 
     
     
         10 . The method of  claim 9  further comprising the step of: 
 reinforcing the amplitude of the locally originating audio signal relative to said echo suppressed audio signal in said combination to thereby further mask said echo.  
 
     
     
         11 . The method of  claim 10  wherein the amplitude of the locally originating audio signal is reinforced according to a user-selectable setting.  
     
     
         12 . The method of  claim 9  further comprising the step of: 
 matching the amplitude of the locally originating audio signal to that of said echo suppressed audio signal in said combination.  
 
     
     
         13 . The method of  claim 7  further comprising the step of: 
 switching to output of the return transmission audio program in an initial state prior to the at least one correlate state.  
 
     
     
         14 . The method of  claim 13  further comprising the step of: 
 switching to output of a combination of the locally originating audio signal and the return transmission audio signal in the initial state.  
 
     
     
         15 . The method of  claim 7  wherein the step of selectively switching comprises the steps of: 
 continuously passing the locally originating audio signal for output to the monitoring device;  
 selectively adding the echo suppressed audio signal to the locally originating audio signal for output to the monitoring device depending on the current state in the sequence.  
 
     
     
         16 . The method of  claim 15  further comprising the step of: 
 amplifying the added locally originating audio signal relative to said echo suppressed audio signal for said output to the monitoring device to thereby further mask said echo.  
 
     
     
         17 . The method of  claim 16  wherein the added locally originating audio signal is amplified by a user-adjustable gain ratio relative to the echo suppressed audio signal.  
     
     
         18 . The method of  claim 15  further comprising the steps of: 
 adding the return transmission audio signal to the locally originating audio signal in an initial state prior to the at least one correlate state for output to the monitoring device;  
 passing only the locally originating audio signal for output to the monitoring device in the at least one correlate state; and  
 adding the echo suppressed audio signal to the locally originating audio signal for output to the monitoring device in the echo suppress state.  
 
     
     
         19 . The method of  claim 15  further comprising the steps of: 
 attenuating the return transmission audio signal; and  
 adding the attenuated, return transmission audio signal to the locally originating audio signal for output to the monitoring device in the at least one correlate state.  
 
     
     
         20 . The method of  claim 7  further comprising the steps of: 
 in the first of the at least one correlate state, performing a reduced sampling rate correlation of the locally originating audio signal to the echo in the return transmission audio signal to determine whether the locally originating audio signal is on-the-air and to approximately determine delay of the echo from the locally originating audio signal within a wider window of time;  
 causing a transition to a second of the at least one correlate state in response to determining that the locally originating audio signal is on-the-air in the first of the at least one correlate state; and  
 in the second of the at least one correlate state, performing a first full sampling rate correlation of said locally originating audio signal to said echo in said return transmission audio signal within a narrower window of time defined according to the approximate delay so as to more exactly converge on the delay of the echo from the locally originating audio signal for use in suppressing said echo from said return transmission audio signal.  
 
     
     
         21 . The method of  claim 20  further comprising the steps of: 
 simultaneous with at least one the steps of performing the reduced sampling rate correlation within the wider window and performing the first full sampling rate correlation within the narrower window, separately performing a second full sampling rate correlation of the locally originating audio signal to the echo in the return transmission audio signal, the second full sampling rate correlation being preset using stored parameters of a prior successful correlation.  
 
     
     
         22 . A method of suppressing echo of a locally generated audio signal from a return transmission audio signal, comprising the steps of: 
 detecting activity on the locally generated audio signal;    in response to the presence of activity on the locally generated audio signal, processing the locally generated audio signal in relation to the return transmission audio signal using a finite impulse response filter having coefficients adapted according to a stochastic gradient adaptation process at a first sampling rate to detect the presence of the echo of the locally generated audio signal in the return transmission audio signal at an approximate delay time within a first window of time;    in response to the detection of the echo, processing the locally generated audio signal in relation to the return transmission audio signal using the finite impulse response filter at a second sampling rate to correlate the locally generated audio signal to its echo in the return transmission audio signal within a second window of time displaced according to the approximate delay time, the second sampling rate being greater than twice the first sampling rate so that the first window is wider than the second window; and    after achieving correlation, suppressing the echo of the locally generated audio signal from the return transmission audio signal using the finite impulse response filter.    
     
     
         23 . The method of  claim 22  further comprising the step of: 
 passing the return transmission audio signal to an output for monitoring; and  
 attenuating the return transmission audio signal at the output during the steps of processing to detect echo and of processing to correlate to the echo.  
 
     
     
         24 . The method of  claim 22  further comprising the step of: 
 providing output for monitoring the return transmission audio signal;  
 switching output to the locally generated audio signal during the steps of processing to detect echo and of processing to correlate to the echo.  
 
     
     
         25 . The method of  claim 22  further comprising the steps of: 
 pausing the processing using the finite impulse response filter to detect echo if activity is no longer detected on the locally generated audio signal;  
 retaining while paused the coefficients of the finite impulse response filter thus far adapted during the processing; and  
 resuming the processing to detect echo with the retained coefficients in response to again detecting the presence of activity on the locally generated audio signal.  
 
     
     
         26 . The method of  claim 22  further comprising the steps of: 
 pausing the processing using the finite impulse response filter to correlate to the echo if activity is no longer detected on the locally generated audio signal;  
 retaining while paused the coefficients of the finite impulse response filter thus far adapted during the processing; and  
 resuming the processing to correlate to the echo with the retained coefficients in response to again detecting the presence of activity on the locally generated audio signal.  
 
     
     
         27 . The method of  claim 22  further comprising the steps of: 
 pausing the suppressing of the echo using the finite impulse response filter if activity is no longer detected on the locally generated audio signal;  
 retaining while paused the coefficients of the finite impulse response filter thus far adapted during the processing; and  
 resuming the suppressing of the echo using the finite impulse response filter in response to again detecting the presence of activity on the locally generated audio signal.  
 
     
     
         28 . The method of  claim 22  further comprising the step of: 
 determining that correlation is achieved during processing to correlate to the echo if at least one of the coefficients is adapted to a value indicating the echo is greater than a first fraction of the locally generated audio signal.  
 
     
     
         29 . The method of  claim 28  further comprising the steps of: 
 continuing to adapt the coefficients of the finite impulse response filter during suppressing the echo; and  
 repeating the method of  claim 22  if the coefficients decrease to less than a second fraction of the locally generated audio signal, the second fraction being less than the first fraction.  
 
     
     
         30 . The method of  claim 29  further comprising the step of retaining the coefficients of the finite impulse response filter from a prior iteration of the method of  claim 22  for use in a subsequent iteration of the method of  claim 22 .  
     
     
         31 . An echo suppressor comprising: 
 a first input for a local signal;    a second input for a return signal; and    a signal processor for processing the local signal and the return signal using a finite impulse response filter having a plurality of adaptively adjustable coefficients to correlate the local signal to an echo of the local signal in the return signal, the signal processor operating in a first mode to initially process the signals using the finite impulse response filter at a reduced sample rate to detect the echo to an approximate delay within an expanded window, and in a second mode to process the signals using the finite impulse response filter at an increased sample rate to correlate to the echo within a narrower window about the approximate delay, whereby the echo suppressor can more quickly correlate to echo over a wide range of delays.    
     
     
         32 . The echo suppressor of  claim 31  further comprising: 
 a voice activated switch for initiating the first mode upon detection of activity in the local signal; and  
 a correlation detector for initiating the second mode upon the coefficients attaining values indicating echo greater than a predetermined fraction of the local signal.  
 
     
     
         33 . The echo suppressor of  claim 32  further comprising: 
 a signal output; and  
 an output switcher operating in response to the voice activated switch and the correlation detector for switching from passing the return signal to the signal output over to passing the local signal to the signal output while the signal processor operates in the first and second modes until correlation is achieved.