chenminxin
2008-10-08, 13:47
在R语言中用wavelet包中的dwt函数：
X=c(1.,2.,3.,4.,5.,6.,7.,8.)
dwt(X, filter="d4", n.levels=3, boundary="periodic", fast=FALSE)
得各级高频小波系数：
An object of class “dwt”
Slot "W":
$W1
[,1]
[1,] 2.828427e+00
[2,] 0.000000e+00
[3,] -2.220446e-16
[4,] 0.000000e+00
$W2
[,1]
[1,] 3.8301270
[2,] -0.3660254
$W3
[,1]
[1,] 4.381341
在matlab中用
dwtmode('per')
[c,l]=wavedec(1:8,3,'db2')
得结果：
c =
Columns 1 through 5
12.727922061357855 -0.000000000002692 -2.464101615135696 4.464101615132895 -1.035276180408362
Columns 6 through 8
-0.000000000000990 -0.000000000000990 3.863703305156531
l =
1 1 2 4 8
显然高频系数是不一样的，我发现matlab做的塔式算法中卷积后作偶数点采样，R语言是做奇数点采样，好像是R错了，是这样么？请高手指点！！
X=c(1.,2.,3.,4.,5.,6.,7.,8.)
dwt(X, filter="d4", n.levels=3, boundary="periodic", fast=FALSE)
得各级高频小波系数：
An object of class “dwt”
Slot "W":
$W1
[,1]
[1,] 2.828427e+00
[2,] 0.000000e+00
[3,] -2.220446e-16
[4,] 0.000000e+00
$W2
[,1]
[1,] 3.8301270
[2,] -0.3660254
$W3
[,1]
[1,] 4.381341
在matlab中用
dwtmode('per')
[c,l]=wavedec(1:8,3,'db2')
得结果：
c =
Columns 1 through 5
12.727922061357855 -0.000000000002692 -2.464101615135696 4.464101615132895 -1.035276180408362
Columns 6 through 8
-0.000000000000990 -0.000000000000990 3.863703305156531
l =
1 1 2 4 8
显然高频系数是不一样的，我发现matlab做的塔式算法中卷积后作偶数点采样，R语言是做奇数点采样，好像是R错了，是这样么？请高手指点！！