Source coding enhancement using spectral-band replication
View Patent ↗The present invention proposes a new method and apparatus for the enhancement of source coding systems. The invention employs bandwidth reduction ( 101 ) prior to or in the encoder ( 103 ), followed by spectral-band replication ( 105 ) at the decoder ( 107 ). This is accomplished by the use of new transposition methods, in combination with spectral envelope adjustments. Reduced bitrate at a given perceptual quality or an improved perceptual quality at a given bitrate is offered. The invention is preferably integrated in a hardware or software codec, but can also be implemented as a separate processor in combination with a codec. The invention offers substantial improvements practically independent of codec type and technological progress.
1. A method for providing a transposed signal which is transposed by a factor M, comprising the following steps:
filtering an input signal through a parallel bank of L filters with impulse responses as
h
k
(
n
)
=
K
p
0
(
n
)
exp
[
j
π
2
L
(
2
k
+
1
)
(
n
-
N
-
1
2
)
+
j
(
-
1
)
k
π
4
M
]
,
where k=0, 1, . . . , L−1, K is a constant, and p 0 (n) is a lowpass prototype filter of length N, producing a set of L complex-valued signals;
downsampling said set of L signals with a factor L/M, producing a set of L complex-valued subband signals;
multiplying phase-angles of said set of L complex-valued subband signals by M, giving a new set of subband signals;
selecting real parts of said new set of subband signals, resulting in a set of L real-valued subband signals;
upsampling a subset of said set of L real-valued subband signals with a factor L′, producing a set of real-valued signals;
filtering said set of real-valued signals through a parallel bank of L′ filters with impulse responses as
f
k
(
n
)
=
K
′
p
0
′
(
n
)
cos
[
π
2
L
′
(
2
k
+
1
)
(
n
-
N
′
-
1
2
)
-
(
-
1
)
k
′
π
4
]
,
where k′=0, 1, . . . , L′−1, K′ is a constant and p 0 (n) is a lowpass prototype filter of length N′, forming a set of L′ filtered signals; and
adding said set of L′ filtered signals and the input signal to produce a transposed signal.
2. A method according to claim 1 , wherein the step of multiplying said phase-angles and the step of selecting said real parts, is computed by the following steps:
providing said set of complex-valued subband signals as
Z k ( n )= R k ( n )+ jI k ( n ),
where R k (n) and I k (n) are real and imaginary parts of Z k (n), respectively;
calculating said set of real-valued subband signals W k (n) as
W
k
(
n
)
=
Z
k
(
n
)
cos
{
M
arctan
(
I
k
(
n
)
R
k
(
n
)
)
}
,
where |Z k (n)|=sqrt{R k (n) 2 +I k (n) 2 } and M is a positive integer transposition factor, using a following trigonometric identity
cos( Ma )=cos M ( a )−( 2 M )sin 2 ( a )cos M-2 ( a )+( 4 M )sin 4 ( a )cos M-4 ( a ) . . .
where a=arctan{I k (n)/R k (n)}, and following relations
cos
(
α
)
=
R
k
(
n
)
Z
k
(
n
)
and
sin
(
a
)
=
I
k
(
n
)
Z
k
(
n
)
;
whereby computational complexity is reduced by elimination of all trigonometric calculations.
3. A method according to claim 1 , further including the following steps:
on a block basis, extracting information conveyed by the phase-difference of an adjacent pair of said complex-valued subband signals;
performing said multiplying of said phase-angles by M, forming a pair of said new subband signals; and
negating one of said new subband signals, on a condition provided by said information; whereby 180° phaseshifts of the subband signals are retained when employing an even integer-valued transposition factor M.
4. A method according to claim 3 , in which said information is given by the dot-product of said complex-valued subband signals Z k (n) and Z k+1 (n) according to
Z k ( n )∘Z k+1 ( n )= R k ( n ) R k+1 ( n )+ I k ( n ) I k+1 ( n ),
where R i (n) and I i (n) are real and imaginary parts of Z i (n) respectively, i=k, k+1, and one of said new subband signals is negated provided said dot-product is negative.
5. An apparatus for providing a transposed signal which is transposed by a factor M, comprising:
a filter for filtering an input signal through a parallel bank of L filters with impulse responses as
h
k
(
n
)
=
K
p
0
(
n
)
exp
[
j
π
2
L
(
2
k
+
1
)
(
n
-
N
-
1
2
)
+
j
(
-
1
)
k
π
4
M
]
,
where k=0, 1, . . . , L−1, K is a constant, p 0 (n) is a lowpass prototype filter of length N, and M is the factor, producing a set of L complex-valued signals;
a downsampler for downsampling said set of L signals with a factor L/M, producing a set of L complex-valued subband signals;
a multiplier for multiplying phase-angles of said set of complex-valued subband signals by M, giving a new set of subband signals;
a selector for selecting real parts of said new set of subband signals, resulting in a set of L real-valued subband signals;
an upsampler for upsampling a subset of said set of L real-valued subband signals with a factor L′, producing a set of real-valued signals;
a filter for filtering said set of real-valued signals through a parallel bank of L′ filters with impulse responses as
f
k
(
n
)
=
K
′
p
0
′
(
n
)
cos
[
π
2
L
′
(
2
k
+
1
)
(
n
-
N
′
-
1
2
)
-
(
-
1
)
k
′
π
4
]
,
where k′=0, 1, . . . , L′−1, K′, is a constant and p 0 (n) is a lowpass prototype filter of length N′, forming a set of L′ filtered signals; and
an adder for adding said set of L′ filtered signals and the input signal to produce the transposed signal.