Temporal interpolation of adjacent spectra
View Patent ↗Embodiments of the present invention exploit redundancy of succeeding FFT spectra and use this redundancy for computing interpolated temporal supporting points. An analysis filter bank converts overlapped sequences of an audio (ex. loudspeaker) signal from a time domain to a frequency domain to obtain a time series of short-time loudspeaker spectra. An interpolator temporally interpolates this time series. The interpolation is fed to an echo canceller, which computes an estimated echo spectrum. A microphone analysis filter bank converts overlapped sequences of an audio microphone signal from the time domain to the frequency domain to obtain a time series of short-time microphone spectra. The estimated echo spectrum is subtracted from the microphone spectrum. Further signal enhancement (filtration) may be applied. A synthesis filter bank converts the filtered microphone spectra to the time domain to generate an echo compensated audio microphone signal. Computational complexity of signal processing systems can, therefore, be reduced.
1. A method for echo compensation of at least one audio microphone signal that includes an echo signal contribution due to an audio loudspeaker signal in a loudspeaker-microphone system, the method comprising:
converting overlapped sequences of the audio loudspeaker signal from a time domain to a frequency domain and obtaining a time series of short-time loudspeaker spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a loudspeaker sub-sampling rate;
temporally interpolating the time series of short-time loudspeaker spectra, including, for each pair of temporally adjacent short-time loudspeaker spectra, calculating an interpolated short-time loudspeaker spectrum by weighted addition of the temporally adjacent short-time loudspeaker spectra;
computing an estimated echo spectrum with its sub-band components for at least one current loudspeaker spectrum by weighted adding of a current short-time loudspeaker spectrum and previous short-time loudspeaker spectra, up to a predetermined maximum time delay, wherein:
first filter coefficients are used for weighting the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra with increasing time delay;
second filter coefficients are used for weighting the interpolated short-time loudspeaker spectra temporally adjacent to the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra; and
the first and second filter coefficients are estimated by an adaptive algorithm; converting overlapped sequences of the audio microphone signal from the time domain to the frequency domain and obtaining a time series of short-time microphone spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a microphone sub-sampling rate;
adaptively filtering the time series of short-time microphone spectra of the microphone signal by at least subtracting a corresponding estimated echo spectrum from a corresponding microphone spectrum, where the first and second filter coefficients are applied and sub-band components of the spectra are used for the subtraction;
converting the filtered time series of short-time spectra of the microphone signal to overlapped sequences of a filtered audio microphone signal; and
overlapping the sequences of the filtered audio microphone signal to generate an echo compensated audio microphone signal.
2. The method according to claim 1 , where the step of temporally interpolating the time series of short-time loudspeaker spectra is made by applying an interpolation matrix P, wherein:
P
=
THH
1
H
2
+
T
~
+
with
H
~
1
=
[
H
0
N
×
r
]
,
H
~
2
=
[
0
N
×
r
H
]
,
and
T
~
=
[
T
0
N
/
2
+
1
×
N
0
N
/
2
+
1
×
N
T
]
.
wherein T is a transformation matrix, H is a diagonal matrix containing window coefficients, H 1 is a first extended matrix of the filter coefficients, H 2 is a second extended matrix, r is the sub-sampling rate, and N is the number of samples.
3. A method according to claim 1 , wherein the adaptively filtering comprises suppressing a residual echo, after subtracting the estimated echo spectrum.
4. A method according claim 1 , wherein the adaptively filtering comprises reducing noise, after subtracting the estimated echo spectrum.
5. A method according claim 1 , wherein the loudspeaker sub-sampling rate is not greater than about 0.75 times the sequence length and greater than about 0.35 times the sequence length.
6. A method according to claim 5 , where the loudspeaker sub-sampling rate is about 0.6 times the sequence length.
7. A method according to claim 1 , wherein converting the overlapped sequences of the audio microphone signal from the time domain to the frequency domain, the adaptively filtering the time series of short-time microphone spectra of the microphone signal, the converting the filtered time series of short-time spectra of the microphone signal and the overlapping the sequences of the filtered audio microphone signal are performed for each of a plurality of audio microphone signals.
8. A signal processor system for echo compensation of at least one audio microphone signal that includes an echo signal contribution due to an audio loudspeaker signal in a loudspeaker-microphone system, the signal processor system comprising:
a loudspeaker analysis filter bank configured to convert overlapped sequences of the audio loudspeaker signal from a time domain to a frequency domain and to obtain a time series of short-time loudspeaker spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a loudspeaker sub-sampling rate;
a temporal interpolator configured to interpolate the time series of short-time loudspeaker spectra, including, for each pair of temporally adjacent short-time loudspeaker spectra, computing an interpolated short-time loudspeaker spectrum by weighted addition of the temporally adjacent short-time loudspeaker spectra;
an echo spectrum estimator having a computer processor configured to compute an estimated echo spectrum with its sub-band components for at least one current loudspeaker spectrum by weighted addition of a current short-time loudspeaker spectrum and previous short-time loudspeaker spectra, up to a predetermined maximum time delay, wherein:
first filter coefficients are used for weighting the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra with increasing time delay;
second filter coefficients are used for weighting the interpolated short-time loudspeaker spectra temporally adjacent to the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra; and
the first and second filter coefficients are estimated by an adaptive algorithm;
a microphone analysis filter bank configured to convert overlapped sequences of the audio microphone signal from the time domain to the frequency domain and obtain a time series of short-time microphone spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a microphone sub-sampling rate;
a synthesis filter bank configured to convert the filtered time series of short-time spectra of the microphone signal to overlapped sequences of a filtered audio microphone signal;
an adaptive filter configured to adaptively filter the time series of short-time microphone spectra of the microphone signal by at least subtracting a corresponding estimated echo spectrum from a corresponding microphone spectrum, where the first and second filter coefficients are applied and sub-band components of the spectra are used for the subtraction; and
a synthesis filter bank configured to overlap the sequences of the filtered audio microphone signal to generate an echo compensated audio microphone signal.
9. A signal processor system according to claim 8 , wherein the adaptive filter comprises a residual echo suppressor applied after the subtraction of the estimated echo spectrum.
10. A signal processor system according to claim 8 , wherein the adaptive filter comprises a noise reducer applied after the subtraction of the estimated echo spectrum.
11. A signal processor system according to claim 8 , wherein the loudspeaker sub-sampling rate is not greater than about 0.75 times the sequence length and greater than about 0.35 times the sequence length.
12. A signal processor system according to claim 11 , wherein the loudspeaker sub-sampling rate is about 0.6 times the sequence length.
13. A signal processor system according to claim 8 , further comprising a beamformer configured to beamform the adaptively filtered time series of short-time microphone spectra of a plurality of microphone signals to generate a combined filtered time series of short-time spectra of the plurality of microphone signals.
14. A signal processor system according to claim 8 , further comprising a hands-free telephony system.
15. A signal processor system according to claim 8 , further comprising a speech recognition system.
16. A signal processor system according to claim 8 , further comprising a vehicle communication system.
17. A computer program product for providing echo compensation of at least one audio microphone signal that includes an echo signal contribution due to an audio loudspeaker signal in a loudspeaker-microphone system, the computer program product comprising a non-transitory computer-readable medium having computer readable program code stored thereon, the computer readable program configured to:
convert overlapped sequences of the audio loudspeaker signal from a time domain to a frequency domain and obtain a time series of short-time loudspeaker spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a loudspeaker sub-sampling rate;
temporally interpolate the time series of short-time loudspeaker spectra, including, for each pair of temporally adjacent short-time loudspeaker spectra, calculate an interpolated short-time loudspeaker spectrum by weighted addition of the temporally adjacent short-time loudspeaker spectra;
compute an estimated echo spectrum with its sub-band components for at least one current loudspeaker spectrum by weighted addition of a current short-time loudspeaker spectrum and previous short-time loudspeaker spectra, up to a predetermined maximum time delay, wherein:
first filter coefficients are used for weighting the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra with increasing time delay;
second filter coefficients are used for weighting the interpolated short-time loudspeaker spectra temporally adjacent to the current loudspeaker spectrum and the corresponding previous short-time loudspeaker spectra; and
the first and second filter coefficients are estimated by an adaptive algorithm;
convert overlapped sequences of the audio microphone signal from the time domain to the frequency domain and obtain a time series of short-time microphone spectra with a predetermined number of sub-bands, wherein the sequences have a predetermined sequence length and an amount of overlapping of the overlapped sequences predetermined by a microphone sub-sampling rate;
adaptively filter the time series of short-time microphone spectra of the microphone signal by at least subtracting a corresponding estimated echo spectrum from a corresponding microphone spectrum, wherein the first and second filter coefficients are applied and sub-band components of the spectra are used for the subtraction;
convert the filtered time series of short-time spectra of the microphone signal to overlapped sequences of a filtered audio microphone signal; and
overlap the sequences of the filtered audio microphone signal to generate an echo compensated audio microphone signal.