Overview

Frequency domain stereo subband Kalman filter

Discussion

Subband Kalman based stereo echo canceler with support for multiple input channels.

I/O Pins

This module works in conjunction with the WOLA filterbank modules and accepts complex data. The default module has two input pins, and one output pin. The first input can have any number of microphone signals, while the second is the stereo reference signal. The reference signals and states are shared among the microphones, saving memory and MIPS over using several instances. The error output pin will have one channel per microphone input.

User Parameters

At instantiation time the maximum number of taps must be defined. The number of taps actually used is the parameter numTaps, found in the module inspector or the module properties view. numTaps allows you to vary the filter length at runtime to verify performance and CPU load.

The module has optional residual noise suppression which is enabled via the enableRNS variable. By default, residual noise suppression is on and the output of the echo canceler (the error signal) is further reduced in level. The residual noise suppression is configured by a 7 element interpolation table. This table operates just like the Table X-Y module with configurable (X, Y) lookup points. The lookup table specifies values in dB for the input and output function

The p0 variable is used to initialize the P and PXTerm arrays at startup time and during any reset events triggered by the optional input pins. p0 defaults to 1.0. Changing this value during runtime will automatically reset P and PXTerm to the new value.

The smoothingTime variable, with units of seconds, is a state transition factor analogous to the leakage factor in a leaky lms algorithm. It defaults to 30, which has a convergence time roughly equivalent to a setting of 100 in the deprecated SbStereoAECV1 module.

The FFTSize parameter is set at instantiation and used to calculate an internal adaptation constant. This is normally the FFT size of the WOLA Analysis module upstream of this module. FFTSize is often (blocksize-1) * 2 unless sub bands are split to multiple AECs.

Optional Input Pins

Three additional input pins will be added to the module if the resetPinsEn module argument is set to true. These control pins can be used to modify the behavior of the module during runtime, as described below:

resetP: This pin has one INT32 value per microphone channel. If the value is not zero, then that microphone's P and PXTerm arrays will be reset to the current p0 value each pump.

masterReset: This pin has only a single INT32 value and one channel. If the value is not zero, then all internal states and weight arrays will be reset to their default values for all microphone channels. This includes the P and PXTerm arrays.

freeze: This pin has one INT32 value per microphone channel. If the value is not zero, then the adaptation stage of the filter will not run for that channel. The existing weights will continue to be applied, but they will not be updated again until the freeze is set to zero. This does not include the RNS weights since they are highly dependent on reference levels.

Optional Output Pins

Three module arguments can be used to add additional output pins, which can be useful for debugging purposes.

outputRNSWeights: Enabling this will output the RNS weights (the W2 array) on an output pin. The output pin will have one channel per microphone. The W2 weights are real and there is one value per input bin. If this pin is enabled, then the W2 weights are not applied to the output error signal.

outputAECWeights: This argument will output the AEC weights (the WW1 array) on an output pin if set to true. This output pin will have 2 * maxtaps channels of numBins per mic input. The WW1 weights are complex, and they are ordered [t1r1c1 t1r1c2 ... t1r2c1 t1r2c2 ... tTrRcC tTrRcC], where r1 is ref channel 1, r2 is ref channel 2, t# is tap number, T is maxTaps and c/C is the mic channel number and max mic channels respectively.

outputPVals: If set to true, this argument will enable two additional output pins - one for the P array, and another for the PXTerm array. Both output pins will have one channel per microphone input.

The P array is all real values, and is ordered per channel as [b1t1c1 b1t1c2 b1t2c1 b1t2c2 ... b1tTc2 b2t1c1 b2t1c2 ... bBtTc1 bBtTc2], where c1 is ref channel 1, c2 is ref channel 2, t# is tap number, T is maxTaps, b# is bin number, and B is number of input bins.

The PXTerm array is complex, and is ordered per channel as [b1t1 b1t2 ... b1tT b2t1 b2t2 ... bBtT], where t# is tap number, T is maxTaps, b# is bin number, and B is number of input bins.

Type Definition

typedef struct _ModuleSbKalmanAECStereoV1
{
    ModuleInstanceDescriptor instance;            // Common Audio Weaver module instance structure
    INT32 maxTaps;                                // Maximum length of the filter
    INT32 numTaps;                                // Current length of the filter
    FLOAT32 V;                                    // Adaptation constant based on num taps and fft size
    FLOAT32 smoothingTime;                        // Time constant of the smoothing process.
    FLOAT32 A;                                    // Smoothing factor
    FLOAT32 p0;                                   // Starting value for covariance estimate
    INT32 enableRNS;                              // Enable residual noise suppression
    INT32 firstRun;                               // First run for setting P from p0 on startup
    FLOAT32 FFTSize;                              // FFT Size used for input bins. Can differ from input block size.
    INT32 enableW2Output;                         // tracker for rns weight output enable
    INT32 enableWOutput;                          // tracker for aec weight output enable
    INT32 enablePVOutput;                         // tracker for rns weight output enable
    INT32 enableControlInputs;                    // tracker for input control pin enable
    INT32 scratchIndex;                           // tracker for which pin scratch is on
    INT32 stateIndex;                             // Circular buffer index. [0 numTaps-1]
    INT32 RNS_numPoints;                          // Number of values in the lookup table. The total table size is [maxPoints 2].
    INT32 RNS_order;                              // Order of the interpolation. This can be either 2 (for linear) or 4 (for pchip).
    FLOAT32* WW1;                                 // Adaptive filter coefficients.
    FLOAT32* P;                                   // Real estimation error covariance matrix.
    FLOAT32* PXTerm;                              // Estimation error covariance matrix - the off-diagonal terms.
    FLOAT32* CC;                                  // Complex state variables (previous inputs).
    FLOAT32* corrC;                               // Stores all the correlation between reference signals per bin and tap - diagonal part. Real number.
    FLOAT32* corrCXTerm;                          // Correlation between ref channels - the off-diagonal terms.
    FLOAT32* W2;                                  // Residual noise suppression weights.
    FLOAT32* RNS_table;                           // Lookup table. The first row is the X values and the second row is the Y values.
    FLOAT32* RNS_polyCoeffs;                      // Interpolation coefficients returned by the grid control.
} ModuleSbKalmanAECStereoV1Class;

Variables

Properties

Name	Type	Usage	isHidden	Default value	Range	Units
maxTaps	int	const	0	16	Unrestricted	samples
numTaps	int	parameter	0	16	1:1:16	samples
V	float	derived	0	1.0302	Unrestricted
smoothingTime	float	parameter	0	30	0.1:1000	seconds
A	float	derived	0	1	0:1
p0	float	parameter	0	1	Unrestricted
enableRNS	int	parameter	0	1	0:1:1
firstRun	int	state	0	0	Unrestricted
FFTSize	float	derived	1	512	Unrestricted
enableW2Output	int	const	1	0	Unrestricted
enableWOutput	int	const	1	0	Unrestricted
enablePVOutput	int	const	1	0	Unrestricted
enableControlInputs	int	const	1	0	Unrestricted
scratchIndex	int	const	1	3	Unrestricted
stateIndex	int	state	1	0	Unrestricted
RNS_numPoints	int	const	1	7	4:1:1000
RNS_order	int	const	1	2	2:2:4
WW1	float*	state	0	[1024 x 1]	Unrestricted
P	float*	state	0	[1024 x 1]	Unrestricted
PXTerm	float*	state	0	[512 x 1]	Unrestricted
CC	float*	state	0	[32 x 32]	Unrestricted
corrC	float*	state	0	[32 x 32]	Unrestricted
corrCXTerm	float*	state	0	[16 x 32]	Unrestricted
W2	float*	state	0	[32 x 1]	Unrestricted
RNS_table	float*	parameter	1	[2 x 7]	Unrestricted
RNS_polyCoeffs	float*	state	1	[4 x 6]	Unrestricted

Pins

Input Pins

Name: dataIn

Description: Mic inputs

Data type: float

Channel range: Unrestricted

Block size range: Unrestricted

Sample rate range: Unrestricted

Complex support: Complex

Name: reference

Description: Stereo reference input

Data type: float

Channel range: 2

Block size range: Unrestricted

Sample rate range: Unrestricted

Complex support: Complex

Output Pins

Name: errorSignal

Description: Error signal = difference between filter output and each channels data input

Data type: float

Scratch Pins

Channel count: 1

Block size: 86

Sample rate: 48000

MATLAB Usage

File Name: sb_kalman_aec_stereo_v1_module.m

 M = sb_kalman_aec_stereo_v1_module(NAME, MAXTAPS, FFTSIZE, RESETPINSENABLE, OUTPUTRNSWEIGHTS, OUTPUTAECWEIGHTS, OUTPUTPVALS)
 Frequency domain subband stereo echo canceler using a Kalman filter based
 algorithm.  Arguments:
    NAME - name of the module.
    MAXTAPS - maximum length of each frequency domain complex adaptive filter
    FFTSIZE - fft size used for input data (typically (# WOLA
              subbands-1)*2
    RESETPINSENABLE - creates 3 optional input pins for controlling behavior of module. 
        - The resetP pin resets all P and PXTerm values to initial p0 value
        - The masterReset pin resets all internal states to initial values (including P) for all channels
        - The freeze pin can freeze the adaptation in the filter. Current weights are still applied
    OUTPUTRNSWEIGHTS - optional output pin for RNS weights
    OUTPUTAECWEIGHTS - optional output pin for AEC weights
    OUTPUTPVALS   - optional output pin for P values
 Copyright 2020.  DSP Concepts, Inc.  All Rights Reserved.
 ----------------------------------------------------------------------
 Set default input arguments
 ----------------------------------------------------------------------