Method for extending the spectral bandwidth of a speech signal
View Patent ↗A method for extending the spectral bandwidth of an excitation signal of a speech signal includes determining a bandwidth limited excitation signal of the speech signal, and applying a nonlinear function to the excitation signal for generating a bandwidth extended excitation signal.
1. A method for extending the spectral bandwidth of an excitation signal of a speech signal, comprising:
determining a bandwidth limited excitation signal of the speech signal; and
generating a bandwidth extended excitation signal based on the bandwidth limited excitation signal by applying a quadratic function to the bandwidth limited excitation signal where the quadratic function is:
{tilde over (x)} Anr,i ( n )= c 2 ( n ) x 2 p,i ( n )+ c 1 ( n ) x p,i ( n ),
where c 1 and c 2 are determined according to the following relations:
c
1
(
n
)
=
K
1
-
x
max
(
n
)
c
2
(
n
)
=
K
1
-
x
max
(
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
)
,
and
c
2
(
n
)
=
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
,
where
x max =Maximum value of input signal vector x p ,
x min =Minimum value of input signal vector x p ,
ε>0,
n=time,
K 1 =Constant for determining maximum value after applying quadratic function to speech signal,
K 2 =Constant for determining minimum value after applying quadratic function to speech signal,
i =segment of signal, and
x p,i (n)=Portion of i of spectrally flat excitation signal at time n.
2. The method of claim 1 , where x max and x min are determined according to the following relations:
x max ( n )=max{ x p,0 ( n ), x p,1 ( n ), . . . x p,N−1 ( n )},
x min ( n )=min{ x p,0 ( n ), x p,1 ( n ), . . . , x p,N−1 ( n )},
K 1 =1.2, and
K 2 =0.2.
3. The method of claim 1 , further including determining a bandwidth limited spectral envelope of the speech signal, and removing the bandwidth limited spectral envelope from the speech signal by applying an inverse spectral envelope to the speech signal.
4. The method of claim 3 , where determining the bandwidth limited spectral envelope of the speech signal includes utilizing a linear predictive coding analysis.
5. The method of claim 3 , where removing the spectral envelope from the speech signal includes multiplying the inverse spectral envelope with the speech signal in the frequency domain of the speech signal.
6. The method of claim 3 , where removing the spectral envelope from the speech signal includes convolving the inverse spectral envelope with the speech signal in the time domain of the speech signal.
7. The method of claim 1 , further including dividing the speech signal into overlapping segments, each segment being described by the following vector, with the spectral envelope of the speech signal being removed: x p (n)=[x p,0 (n), x p,1 (n), . . . , x p,N−1 (n)] T where N=length of the segment.
8. The method of claim 1 , further including high pass filtering the extended excitation signal for removing frequency components around 0 Hz.
9. The method of claim 1 , further including utilizing extended parts of the excitation signal for replacing noisy parts of the bandwidth limited excitation signal, the bandwidth limited excitation signal corresponding to a speech signal recorded in a noisy environment.
10. The method of claim 1 , further including utilizing extended parts of the excitation signal for replacing the corresponding parts of a bandwidth limited excitation signal corresponding to a bandwidth limited speech signal transmitted via a transmission unit of a telecommunication system, the spectral parts of the speech signal suppressed by the transmission line being generated on the basis of the extended spectral bandwidth parts of the excitation signal.
11. A method for enhancing the quality of a speech signal, comprising:
determining a spectral envelope of the speech signal based on the speech signal having a limited spectral bandwidth;
generating a bandwidth limited excitation signal of the speech signal;
extending the spectral bandwidth of the generated excitation signal by applying a quadratic function to the bandwidth limited excitation signal; and
applying the bandwidth extended excitation signal to the spectral envelope for generating the enhanced speech signal where the quadratic function is:
{tilde over (x)} Anr,i ( n )= c 2 ( n ) x 2 p,i ( n )+ c 1 ( n ) x p,i ( n ),
where c 1 and c 2 are determined according to the following relations:
c
1
(
n
)
=
K
1
-
x
max
(
n
)
c
2
(
n
)
=
K
1
-
x
max
(
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
)
,
and
c
2
(
n
)
=
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
,
where
x max =Maximum value of input signal vector x p ,
x min =Minimum value of input signal vector x p ,
ε>0,
n=time,
K 1 =Constant for determining maximum value after applying quadratic function to speech signal,
K 2 =Constant for determining minimum value after applying quadratic function to speech signal,
i=segment of signal, and
x p,i (n) =Portion of i of spectrally flat excitation signal at time n.
12. The method of claim 11 , where the speech signal is one transmitted by a bandwidth limited transmission system, and generating the enhanced speech signal extends the spectral bandwidth of the speech signal and causes signal reconstruction of noisy parts of the speech signal recorded in a noisy environment.
13. The method of claim 11 , including removing the determined spectral envelope from the bandwidth limited speech signal for generating the bandwidth limited excitation signal.
14. The method of claim 11 , including multiplying the extended excitation signal with the spectral envelope in the frequency domain of the speech signal for generating the enhanced speech signal.
15. The method of claim 11 , including increasing the sampling frequency before determining the spectral envelope.
16. The method of claim 11 , where the speech signal is a signal transmitted via a transmission unit of a telecommunication system, the spectral parts of the speech signal suppressed by the transmission unit being added by the spectral bandwidth extension.
17. The method of claim 16 , where the frequency components suppressed by the transmission unit of the telecommunication system are the frequency components of the speech signal between 0 and approximately 200 Hz and frequency components larger than approximately 3700 Hz.
18. The method of claim 11 where, for extending the spectral bandwidth, the spectral envelope is determined on the basis of the bandwidth limited speech signal transmitted by a bandwidth limited transmission system, a bandwidth extended spectral envelope is determined by comparing the bandwidth limited spectral envelope to predetermined envelopes stored in a look up table and by selecting the envelope in the look up table that best matches the bandwidth limited spectral envelope of the voice signal, and the extended spectral envelope being applied to the extended excitation signal for generating the enhanced bandwidth extended speech signal.
19. The method of claim 11 , including reconstructing noisy parts of a speech signal by replacing the noisy parts of the speech signal on the basis of the extended parts of the bandwidth extended excitation signal for generating an enhanced speech signal.
20. A system for extending the spectral bandwidth of the speech signal transmitted by a bandwidth limited transmission system and for signal reconstruction for noisy parts of the speech signal recorded in a noisy environment, the system comprising:
a determination unit for determining a spectral envelope based upon a bandwidth limited part of the speech signal;
a generating unit for generating an bandwidth limited excitation signal;
a calculation unit for calculating a bandwidth extended excitation signal by applying a quadratic function to the bandwidth limited excitation signal and for applying the spectral envelope to the bandwidth extended excitation signal for generating an enhanced speech signal where the quadratic function is:
{tilde over (x)} Anr,i ( n )= c 2 ( n ) x 2 p,i ( n )+ c 1 ( n ) x p,i ( n ),
where c 1 and c 2 are determined according to the following relations:
c
1
(
n
)
=
K
1
-
x
max
(
n
)
c
2
(
n
)
=
K
1
-
x
max
(
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
)
,
and
c
2
(
n
)
=
K
1
-
K
2
x
max
(
n
)
-
x
min
(
n
)
+
ɛ
,
where
x max =Maximum value of input signal vector x p ,
x min =Minimum value of input signal vector x p ,
ε>0,
n=time,
K 1 =Constant for determining maximum value after applying quadratic function to speech signal,
K 2 =Constant for determining minimum value after applying quadratic function to speech signal,
i =segment of signal, and
x p,i (n) =Portion of i of spectrally flat excitation signal at time n.