Apparatus and method for synthesizing an audio signal, decoder, encoder, system and computer program
A method and an apparatus for synthesizing an audio signal are described. A spectral tilt is applied to the code of a codebook used for synthesizing a current frame of the audio signal. The spectral tilt is based on the spectral tilt of the current frame of the audio signal. Further, an audio decoder operating in accordance with the inventive approach is described.
1. An apparatus for synthesizing an audio signal, comprising:
a processing unit configured to apply a spectral tilt to the code of a codebook used for synthesizing a current frame of the audio signal,
wherein the spectral tilt is based on the spectral tilt of the current frame of the audio signal,
wherein the apparatus is configured to determine the spectral tilt of the current frame of the audio signal on the basis of spectral envelope information for the current frame of the audio signal,
wherein the processing unit is configured to apply the spectral tilt by filtering the code from the codebook based on a transfer function modeling the spectral tilt, and
wherein the processing unit comprises a hardware implementation.
2. The apparatus of claim 1 , wherein the spectral envelope information is defined by LPC coefficients, and wherein the spectral tilt of the current frame of the audio signal is defined as follows:
γ
=
-
∑
n
=
0
N
f
s
(
n
+
1
)
f
s
(
n
)
f
s
2
(
n
)
with:
f s (n) the infinite impulse response of a LPC synthesis filter comprising the transfer function F s (z)=1/A(z), and
N the size of the truncation of the infinite impulse response f s (n).
3. The apparatus of claim 1 , wherein the spectral envelope information is defined by LPC coefficients, and wherein the spectral tilt of the current frame of the audio signal is defined as follows:
γ
=
-
∑
n
=
0
N
f
e
(
n
+
1
)
f
e
(
n
)
f
e
2
(
n
)
with:
f e (n) the infinite impulse response of a LPC synthesis filter comprising the transfer function
F
e
(
z
)
=
A
(
1
/
w
1
)
A
(
1
/
w
2
)
,
N the size of the truncation of the infinite impulse response f s (n), and
w1, w2 weighting constants for defining the formantic structure of the transfer function F e (z).
4. The apparatus of claim 2 , wherein N is equal to the number of codes in the codebook.
5. The apparatus of claim 1 , wherein the transfer function comprising the spectral tilt is defined as follows:
F t1 ( z )=1 −γz −1 ,
with:
γ spectral tilt.
6. The apparatus of claim 1 , wherein the processing unit is further configured to combine the determined spectral tilt of the current frame of the audio signal with a factor related to the voicing of the previous frame of the audio signal.
7. The apparatus of claim 6 , wherein the factor related to the voicing of the previous frame of the audio signal is defined as follows:
β
=
constant
·
(
1
+
voicing
)
with
:
voicing
=
energy
(
contribution
of
adaptive
codebook
)
-
energy
(
contribution
of
fixed
codebook
)
energy
(
sum
of
contributions
)
.
8. The apparatus of claim 6 , wherein the processing unit is configured to apply the spectral tilt by filtering the code from the codebook based on a transfer function comprising the spectral tilt and the factor related to the voicing of the previous frame of the audio signal.
9. The apparatus of claim 8 , wherein the transfer function comprising the spectral tilt is defined as follows:
F t2 ( z )=1−( a·β+b ·γ) z −1 ,
with:
a, b constants.
10. The apparatus of claim 1 , wherein the audio signal is a speech signal, wherein the processing unit for applying the spectral tilt comprises a filter, and wherein the apparatus further comprises:
an adaptive codebook,
a fixed codebook,
the filter coupled to the fixed codebook, the filter being configured to apply the determined spectral tilt to the code of the fixed codebook for acquiring a filtered code of the fixed codebook,
a summer coupled to the adaptive codebook and to the filter, the summer configured to combine a code from the adaptive codebook and the filtered code of the fixed codebook for acquiring a combined code, and
a LPC synthesis filter coupled to the summer.
11. The apparatus of claim 10 , further comprising:
a pitch gain amplifier coupled between the adaptive codebook and the summer, the pitch gain amplifier configured to multiply the code from the adaptive codebook with a pitch gain, and
a code gain amplifier coupled between the filter and the summer, the code gain amplifier configured to multiply the filtered code of the fixed codebook with a code gain.
12. The apparatus of claim 10 , further comprising:
a voicing estimator coupled to the adaptive codebook and to the summer, the voicing estimator configured to output a factor related to the voicing of the previous frame of the audio signal to the filter, and
a storage configured to store LPC coefficients describing spectral envelope information for the current frame of the audio signal, the storage being coupled to the filter.
13. An audio decoder comprising an apparatus for synthesizing an audio signal according to claim 1 .
14. A system, comprising:
an audio decoder according to claim 13 , and
an audio encoder configured to determine from a spectral tilt of a current frame of the audio signal a spectral tilt for a code of a codebook representing a current frame of the audio signal.
15. A method for synthesizing an audio signal, the method comprising:
applying, by a processing unit, a spectral tilt to the code of a codebook used for synthesizing a current frame of the audio signal,
wherein the spectral tilt is determined on the basis of the spectral tilt of the current frame of the audio signal,
wherein the spectral tilt of the current frame of the audio signal is determined on the basis of spectral envelope information for the current frame of the audio signal, and
wherein applying the spectral tilt comprises filtering the code from the codebook based on a transfer function modeling the spectral tilt
wherein the processing unit comprises a hardware implementation.
16. The method of claim 15 , wherein the spectral envelope information is defined by LPC coefficients, and wherein the spectral tilt of the current frame of the audio signal is determined as follows:
γ
=
-
∑
n
=
0
N
f
s
(
n
+
1
)
f
s
(
n
)
f
s
2
(
n
)
with:
f s (n) the infinite impulse response of a LPC synthesis filter comprising the transfer function F s (z)=1/A(z), and
N the size of the truncation of the infinite impulse response f s (n).
17. The method of claim 15 , wherein the spectral envelope information is defined by LPC coefficients, and wherein the spectral tilt of the current frame of the audio signal is determined as follows:
γ
=
-
∑
n
=
0
N
f
e
(
n
+
1
)
f
e
(
n
)
f
e
2
(
n
)
with:
f e (n) the infinite impulse response of a LPC synthesis filter comprising the transfer function
F
e
(
z
)
=
A
(
1
/
w
1
)
A
(
1
/
w
2
)
,
N the size of the truncation of the infinite impulse response f s (n), and
w1, w2 weighting constants for defining the formantic structure of the transfer function F e (z).
18. The method of claim 16 , wherein N is equal to the number of codes in the codebook.
19. The method of claim 15 , wherein the transfer function comprising the spectral tilt is determined as follows:
F t1 ( z )=1 −γz −1 ,
with:
γ spectral tilt.
20. The method of claim 15 , further comprising combining the determined spectral tilt of the current frame of the audio signal with a factor related to the voicing of the previous frame of the audio signal.
21. The method of claim 20 , wherein the factor related to the voicing of the previous frame of the audio signal is determined as follows:
β
=
constant
·
(
1
+
voicing
)
with
:
voicing
=
energy
(
contribution
of
adaptive
codebook
)
-
energy
(
contribution
of
fixed
codebook
)
energy
(
sum
of
contributions
)
.
22. The method of claim 20 , wherein applying the spectral tilt comprises filtering the code from the codebook based on a transfer function comprising the spectral tilt and the factor related to the voicing of the previous frame of the audio signal.
23. The method of claim 22 , wherein the transfer function comprising the spectral tilt is determined as follows:
F t2 ( z )=1−( a·β+b ·γ) z −1 ,
with:
a, b constants.
24. The method of claim 15 , wherein the audio signal is a speech signal, and wherein synthesizing the audio signal comprises for a frame of the audio signal:
applying the determined spectral tilt to the code of a fixed codebook for acquiring a filtered code of the fixed codebook,
combining a code from an adaptive codebook and the filtered code of the fixed codebook to acquire a combined code, and
filtering the combined code by a LPC synthesis filter.
25. The method of claim 24 , further comprising multiplying the code from the adaptive codebook with a pitch gain, and multiplying the filtered code of the fixed codebook with a code gain.
26. The method of claim 24 , further comprising:
based on the code from the adaptive codebook and the combined code, generating a factor related to the voicing of the previous frame of the audio signal, and
storing LPC coefficients describing spectral envelope information for the current frame of the audio signal.
27. A non-transitory computer medium storing instructions for carrying out, when run on a computer, a method for synthesizing an audio signal, the method comprising:
applying a spectral tilt to the code of a codebook used for synthesizing a current frame of the audio signal,
wherein the spectral tilt is determined on the basis of the spectral tilt of the current frame of the audio signal,
wherein the spectral tilt of the current frame of the audio signal is determined on the basis of spectral envelope information for the current frame of the audio signal, and
wherein applying the spectral tilt comprises filtering the code from the codebook based on a transfer function modeling the spectral tilt.