Segmenting, windowing, framing with 50%-75% overlap

Question

Mona 2014-11-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/162788-segmenting-windowing-framing-with-50-75-overlap

评论： Mona 2014-11-17

Hi everyone,

I am very new to matlab so please be patient with me if I sound dumb!!!

I have a project where I am suppose to read a noisy speech and filter it, and I am stuck in the first part. I was able to read the speech and plot it using the following

[x, fs, nbits]= wavread('noisy.wav'); sound(x, fs); timex=(1: length(x))/fs; plot(timex, x); xlabel ('Seconds') ; ylabel ('Amplitude');

where fs = 16000 and length(x) = 159999

Now, I am suppose to do Segmentation , windowing, and framing with a 50%~ 75% overlap in order to take STFT, but I am not able to find a matlab function that will section my matrix x into overlapped frames?!! Can someone help me please!!!!

Thanks

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Star Strider 2014-11-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/162788-segmenting-windowing-framing-with-50-75-overlap#answer_159033

If you have the Signal Processing Toolbox, use the spectrogram function. It will do what you want.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

Mona 2014-11-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/162788-segmenting-windowing-framing-with-50-75-overlap#answer_159096

编辑：Star Strider 2014-11-15

在 MATLAB Online 中打开

Thank you Star Strider,

I've used the spectrogram function as follows:

flength=256;
foverlap=flength/2;
fftlength=flength;
wnd=hamming(flength,'periodic');
[S, F, T, P] = spectrogram (yf, wnd,foverlap, fftlength, fs);

But I still have questions about its return. It returns four things S, F, T, and P . Is S a matrix of rows and columns where the number of columns matches the number of frames? For example, my original yf is 159999 and I took the overlap to be 50% and my fs is 16000. I am not sure how to figure out the length of S and number of frames?

Please help

Thanks

2 个评论
显示无隐藏无

Star Strider 2014-11-15

My pleasure!

The ‘S’ matrix according to the documentation:

Each column of S contains an estimate of the short-term, time-localized frequency content of x. Time increases across the columns of S and frequency increases down the rows.

It is difficult to answer in any detail. You probably need to plot ‘S’ to see what your spectrogram looks like. That may tell you everything you need to know.

In the ‘Examples’ section of the documentation (about the power spectral densities of chirps) is a reasonably detailed discussion of the various variables the spectrogram function returns and how to plot and interpret them.

Mona 2014-11-17