signal processing using windowing

I am doing random vibration and starting with the raw data from either test or finite element modeling.

I am using MATLAB to process time domain data into the frequency domain. First I want to apply windowing (hamming to start) to the data and then convert to frequency domain.

My data is in table form ie, x, y, z, time. Column 1 is x, column 2 is y, column 3 is z and column 4 is time.

My table is <2413851x4 double>

I need to reduce the number of samples down to 796 and wanted to use the hamming window to do this. However, I have never done this using mathcad and I keep getting matrices must be of same size. My window size is 1344 with an overlap of 1000.

In essence I need to take the data from 2.4 million plus data points down to 796 data points in a manner that won't lose the integrity of the data and filter it to get as much detail as possible that is burried within the raw data.

Use the oversampling algorithm ; Take the geometric average of 8 samples. This becoms your new sample. You will loose high frequency artifacts.

Ok. I am as new to matlab as they come. Where is this routine located?

You will have to scan through the array, calculate the average and store the down samples points in a new array