在时间与频率的分析领域中,有不少的讯号的单纯使用频域或时域表示,而是同时使用时域与频域来表示。
有几种方法或转换被里昂·柯恩统整组织被称为"时频分析",[ 1] [ 2] [ 3] 最常被使用的方法称为“二次”或“双线性时频分析”,而此类方法中,最被广泛使用的方法中以韦格纳分布为其中之一,其他的时频分布则被称为维格纳分布 的折积版。另一个被广泛使用的方法为频谱图 ,为“短时距傅立叶转换 ”的平方,频谱图有着平方必为正的优点,容易由图理解,但有着不可逆的缺点,如短时距傅立叶转换不可逆计算,无法从频谱图找回原信号。而验证这些理论与定义验证可以参考“二次式时频分布理论”。[ 4]
本文主题虽是讯号处理领域,但是借由量子力学的相空间来推导某些分布从A分布转换至B分布的过程。一个信号在相同的状况下,给与不同的时频分布表示方式,透过简单的平滑器或滤波器,计算出其他分布。
一般化
如果我们用变数ω =2πf ,然后,借用量子力学领域中使用的符号,就可以显示该时间-频率表示,如维格纳分布函数和其它双线性时间-频率分布,可表示为
C
(
t
,
ω
)
=
1
4
π
2
∭
s
∗
(
u
−
1
2
τ
)
s
(
u
+
1
2
τ
)
ϕ
(
θ
,
τ
)
e
−
j
θ
t
−
j
τ
ω
+
j
θ
u
d
u
d
τ
d
θ
,
{\displaystyle C(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iiint s^{*}\left(u-{\dfrac {1}{2}}\tau \right)s\left(u+{\dfrac {1}{2}}\tau \right)\phi (\theta ,\tau )e^{-j\theta t-j\tau \omega +j\theta u}\,du\,d\tau \,d\theta ,}
(1)
ϕ
(
θ
,
τ
)
{\displaystyle \phi (\theta ,\tau )}
为一定义其分布及特性之二维函数。
维格纳分布的核为一。但在一般型式里任何分布的核为一没有任何的意义,在其他状况下维格纳分布的核应为其他结果。
特征方程式
特征方程式为双傅立叶转换,从方程式(1)可以得到
C
(
t
,
ω
)
=
1
4
π
2
∬
M
(
θ
,
τ
)
e
−
j
θ
t
−
j
τ
ω
d
θ
d
τ
{\displaystyle C(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iint M(\theta ,\tau )e^{-j\theta t-j\tau \omega }\,d\theta \,d\tau }
(2)
M
(
θ
,
τ
)
=
ϕ
(
θ
,
τ
)
∫
s
∗
(
u
−
1
2
τ
)
s
(
u
+
1
2
τ
)
e
j
θ
u
d
u
=
ϕ
(
θ
,
τ
)
A
(
θ
,
τ
)
{\displaystyle {\begin{alignedat}{2}M(\theta ,\tau )&=\phi (\theta ,\tau )\int s^{*}\left(u-{\dfrac {1}{2}}\tau \right)s\left(u+{\dfrac {1}{2}}\tau \right)e^{j\theta u}\,du\\&=\phi (\theta ,\tau )A(\theta ,\tau )\\\end{alignedat}}}
(3)
A
(
θ
,
τ
)
{\displaystyle A(\theta ,\tau )}
为对称模糊函数,特征方程式也可易被称为广义模糊函式。
分布之间转换关系
假设有两个分布
C
1
{\displaystyle C_{1}}
and
C
2
{\displaystyle C_{2}}
,个别对应核为
ϕ
1
{\displaystyle \phi _{1}}
and
ϕ
2
{\displaystyle \phi _{2}}
,特征方程式为
M
1
(
ϕ
,
τ
)
=
ϕ
1
(
θ
,
τ
)
∫
s
∗
(
u
−
1
2
τ
)
s
(
u
+
1
2
τ
)
e
j
θ
u
d
u
{\displaystyle M_{1}(\phi ,\tau )=\phi _{1}(\theta ,\tau )\int s^{*}\left(u-{\dfrac {1}{2}}\tau \right)s\left(u+{\dfrac {1}{2}}\tau \right)e^{j\theta u}\,du}
(4)
M
2
(
ϕ
,
τ
)
=
ϕ
2
(
θ
,
τ
)
∫
s
∗
(
u
−
1
2
τ
)
s
(
u
+
1
2
τ
)
e
j
θ
u
d
u
{\displaystyle M_{2}(\phi ,\tau )=\phi _{2}(\theta ,\tau )\int s^{*}\left(u-{\dfrac {1}{2}}\tau \right)s\left(u+{\dfrac {1}{2}}\tau \right)e^{j\theta u}\,du}
(5)
方程式(4)、(5)相除得
M
1
(
ϕ
,
τ
)
=
ϕ
1
(
θ
,
τ
)
ϕ
2
(
θ
,
τ
)
M
2
(
ϕ
,
τ
)
{\displaystyle M_{1}(\phi ,\tau )={\dfrac {\phi _{1}(\theta ,\tau )}{\phi _{2}(\theta ,\tau )}}M_{2}(\phi ,\tau )}
(6)
方程式(6)相当重要,其结果使其连接特征方程式在有线区域内之核不为零。
欲获得两分布之间的关系,需使用双傅立叶转换并使用方程式(2)
C
1
(
t
,
ω
)
=
1
4
π
2
∬
ϕ
1
(
θ
,
τ
)
ϕ
2
(
θ
,
τ
)
M
2
(
θ
,
τ
)
e
−
j
θ
t
−
j
τ
ω
d
θ
d
τ
{\displaystyle C_{1}(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iint {\dfrac {\phi _{1}(\theta ,\tau )}{\phi _{2}(\theta ,\tau )}}M_{2}(\theta ,\tau )e^{-j\theta t-j\tau \omega }\,d\theta \,d\tau }
(7)
用
C
2
{\displaystyle C_{2}}
来表示
M
2
{\displaystyle M_{2}}
C
1
(
t
,
ω
)
=
1
4
π
2
⨌
ϕ
1
(
θ
,
τ
)
ϕ
2
(
θ
,
τ
)
C
2
(
t
,
ω
′
)
e
j
θ
(
t
′
−
t
)
+
j
τ
(
ω
′
−
ω
)
d
θ
d
τ
d
t
′
d
ω
′
{\displaystyle C_{1}(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iiiint {\dfrac {\phi _{1}(\theta ,\tau )}{\phi _{2}(\theta ,\tau )}}C_{2}(t,\omega ^{'})e^{j\theta (t^{'}-t)+j\tau (\omega ^{'}-\omega )}\,d\theta \,d\tau \,dt^{'}\,d\omega ^{'}}
(8)
可改写成
C
1
(
t
,
ω
)
=
∬
g
12
(
t
′
−
t
,
ω
′
−
ω
)
C
2
(
t
,
ω
′
)
d
t
′
d
ω
′
{\displaystyle C_{1}(t,\omega )=\iint g_{12}(t^{'}-t,\omega '-\omega )C_{2}(t,\omega ')\,dt^{'}\,d\omega '}
(9)
其中,
g
12
(
t
,
ω
)
=
1
4
π
2
∬
ϕ
1
(
θ
,
τ
)
ϕ
2
(
θ
,
τ
)
e
j
θ
t
+
j
τ
ω
d
θ
d
τ
{\displaystyle g_{12}(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iint {\dfrac {\phi _{1}(\theta ,\tau )}{\phi _{2}(\theta ,\tau )}}e^{j\theta t+j\tau \omega }\,d\theta \,d\tau }
(10)
频谱与其他双线性相互关系
我们专注于其中一个从任意代表性的频谱转换的情况,在方程式(9)中,
C
1
{\displaystyle C_{1}}
为频谱图而
C
2
{\displaystyle C_{2}}
为任意数,为了简化符号使用以下表示,
ϕ
S
P
=
ϕ
1
{\displaystyle \phi _{SP}=\phi _{1}}
,
ϕ
=
ϕ
2
{\displaystyle \phi =\phi _{2}}
,
g
S
P
=
g
12
{\displaystyle g_{SP}=g_{12}}
,可被表示为
C
S
P
(
t
,
ω
)
=
∬
g
S
P
(
t
′
−
t
,
ω
′
−
ω
)
C
(
t
,
ω
′
)
d
t
′
d
ω
′
{\displaystyle C_{SP}(t,\omega )=\iint g_{SP}(t^{'}-t,\omega ^{'}-\omega )C(t,\omega ^{'})\,dt^{'}\,d\omega ^{'}}
(11)
频谱图的核为
g
S
P
(
t
,
ω
)
=
1
4
π
2
∬
A
h
(
−
θ
,
τ
)
ϕ
(
θ
,
τ
)
e
j
θ
t
+
j
τ
ω
d
θ
d
τ
=
1
4
π
2
∭
1
ϕ
(
θ
,
τ
)
h
∗
(
u
−
1
2
τ
)
h
(
u
+
1
2
τ
)
e
j
θ
t
+
j
τ
ω
−
j
θ
u
d
u
d
τ
d
θ
=
1
4
π
2
∭
h
∗
(
u
−
1
2
τ
)
h
(
u
+
1
2
τ
)
ϕ
(
θ
,
τ
)
ϕ
(
θ
,
τ
)
ϕ
(
−
θ
,
τ
)
e
−
j
θ
t
+
j
τ
ω
+
j
θ
u
d
u
d
τ
d
θ
{\displaystyle {\begin{alignedat}{3}g_{SP}(t,\omega )&={\dfrac {1}{4\pi ^{2}}}\iint {\dfrac {A_{h}(-\theta ,\tau )}{\phi (\theta ,\tau )}}e^{j\theta t+j\tau \omega }\,d\theta \,d\tau \\&={\dfrac {1}{4\pi ^{2}}}\iiint {\dfrac {1}{\phi (\theta ,\tau )}}h^{*}(u-{\dfrac {1}{2}}\tau )h(u+{\dfrac {1}{2}}\tau )e^{j\theta t+j\tau \omega -j\theta u}\,du\,d\tau \,d\theta \\&={\dfrac {1}{4\pi ^{2}}}\iiint h^{*}(u-{\dfrac {1}{2}}\tau )h(u+{\dfrac {1}{2}}\tau ){\dfrac {\phi (\theta ,\tau )}{\phi (\theta ,\tau )\phi (-\theta ,\tau )}}e^{-j\theta t+j\tau \omega +j\theta u}\,du\,d\tau \,d\theta \\\end{alignedat}}}
(12)
令
ϕ
(
−
θ
,
τ
)
ϕ
(
θ
,
τ
)
=
1
{\displaystyle \phi (-\theta ,\tau )\phi (\theta ,\tau )=1}
,
g
S
P
(
t
,
ω
)
{\displaystyle g_{SP}(t,\omega )}
为窗函数,然而在
−
ω
{\displaystyle -\omega }
状况下得
g
S
P
(
t
,
ω
)
=
C
h
(
t
,
−
ω
)
{\displaystyle g_{SP}(t,\omega )=C_{h}(t,-\omega )}
(13)
使其核满足
ϕ
(
−
θ
,
τ
)
ϕ
(
θ
,
τ
)
=
1
{\displaystyle \phi (-\theta ,\tau )\phi (\theta ,\tau )=1}
C
S
P
(
t
,
ω
)
=
∬
C
s
(
t
′
,
ω
′
)
C
h
(
t
′
−
t
,
ω
′
−
ω
)
d
t
′
d
ω
′
{\displaystyle C_{SP}(t,\omega )=\iint C_{s}(t^{'},\omega ^{'})C_{h}(t^{'}-t,\omega ^{'}-\omega )\,dt^{'}\,d\omega ^{'}}
(14)
其核亦满足
ϕ
(
−
θ
,
τ
)
ϕ
(
θ
,
τ
)
=
1
{\displaystyle \phi (-\theta ,\tau )\phi (\theta ,\tau )=1}
其证明可见Janssen[4]. 当
ϕ
(
−
θ
,
τ
)
ϕ
(
θ
,
τ
)
{\displaystyle \phi (-\theta ,\tau )\phi (\theta ,\tau )}
不等于1时,
C
S
P
(
t
,
ω
)
=
⨌
G
(
t
″
,
ω
″
)
C
s
(
t
′
,
ω
′
)
C
h
(
t
″
+
t
′
−
t
,
−
ω
″
+
ω
−
ω
′
)
d
t
′
d
t
″
d
ω
d
ω
″
{\displaystyle C_{SP}(t,\omega )=\iiiint G(t^{''},\omega ^{''})C_{s}(t^{'},\omega ^{'})C_{h}(t^{''}+t^{'}-t,-\omega ^{''}+\omega -\omega ^{'})\,dt^{'}\,dt^{''}\,d\omega ^{\,}d\omega ^{''}}
(15)
G
(
t
,
ω
)
=
1
4
π
2
∬
e
−
j
θ
t
−
j
τ
ω
ϕ
(
θ
,
τ
)
ϕ
(
−
θ
,
τ
)
d
θ
d
τ
{\displaystyle G(t,\omega )={\dfrac {1}{4\pi ^{2}}}\iint {\dfrac {e^{-j\theta t-j\tau \omega }}{\phi (\theta ,\tau )\phi (-\theta ,\tau )}}\,d\theta \,d\tau }
(16)
参考资料
^ L. Cohen , "Generalized phase-space distribution functions," Jour. Math. Phys. , vol.7, pp. 781–786, 1966.
^ L. Cohen, "Quantization Problem and Variational Principle in the Phase Space Formulation of Quantum Mechanics," Jour. Math. Phys. , vol.7, pp. 1863–1866, 1976.
^ A. J. E. M. Janssen, "On the locus and spread of pseudo-density functions in the time frequency plane," Philips Journal of Research , vol. 37, pp. 79–110, 1982.
^ B. Boashash, “Theory of Quadratic TFDs”, Chapter 3, pp. 59–82, in B. Boashash, editor, Time-Frequency Signal Analysis & Processing: A Comprehensive Reference, Elsevier, Oxford, 2003; ISBN 0-08-044335-4 .