短时傅里叶变换的matlab实现代码,有效完成时频分析

以下是短时傅里叶变换的MATLAB实现代码，可以有效完成时频分析：

function [S, F, T] = stft(x, fs, window, noverlap, nfft)
% x: input signal
% fs: sampling frequency
% window: window function (default: hamming)
% noverlap: number of samples to overlap (default: 50%)
% nfft: number of FFT points (default: nextpow2 of window length)

if nargin < 3 || isempty(window)
    window = hamming(round(fs/50));
end

if nargin < 4 || isempty(noverlap)
    noverlap = round(length(window)/2);
end

if nargin < 5 || isempty(nfft)
    nfft = 2^nextpow2(length(window));
end

[S, F, T] = spectrogram(x, window, noverlap, nfft, fs);

S = abs(S)/length(window); % normalize by window length
S = 20*log10(S); % convert to dB scale
S = max(-120, S); % clip values below -120 dB

使用方法：

% load example signal
load gong.mat;

% calculate STFT
[S, F, T] = stft(y, Fs);

% plot spectrogram
imagesc(T, F, S);
set(gca, 'YDir', 'normal');
xlabel('Time (s)');
ylabel('Frequency (Hz)');
colorbar;

这里使用了MATLAB内置的`spectrogram`函数来计算STFT，并将结果转换为dB单位。代码中还可以自定义窗口函数、重叠数量和FFT点数等参数。