Question

我在MATLAB上编写了以下代码：

N = [64 128 256 512];
NumOfRuns = 50;   
PerLength = 2048;  
freq = 0:2/(PerLength-1):2;
for m=1:length(N)
    Px = zeros(NumOfRuns,PerLength);
    for i=1:NumOfRuns
        x = randn(1,N(m));
        Px(:,i) = periodogram(x);
    end
end

当我在MATLAB上运行此代码时，它会给出错误：

订阅的分配维度不匹配。例子13中的错误（第10行）Px（：，I）=周期图（x）。

Answer 1

代码备注：

N = [64 128 256 512];
NumOfRuns = 50;   
PerLength = 2048; 

% freq = 0:2/(PerLength-1):2; % you are not using freq

% You are looping from 1:4, defining Px each time, not storing Px outside of the loop. 
% After this loop, Px will just be in its 4th state. 
for m = 1:length(N)

    Px = zeros(NumOfRuns,PerLength);

    for i = 1:NumOfRuns 

        x = randn(1,N(m));
        Px(:,i) = periodogram(x);

    end

end

<强>周期图：

您假设periodogram返回与输入相同数量的元素。一个简单的测试会告诉你这不是真的！

a = periodogram([1, 2]);
numel(a)
% >> ans = 129

阅读此处的文档：

https://uk.mathworks.com/help/signal/ref/periodogram.html

离散傅里叶变换（DFT）中的点数nfft是最大值256或者下一个2的幂大于信号长度

如上所述，129与2^7 = 128

订阅分配维度不匹配 - 周期图MATLAB

1 个答案: