Signal statistics check: oscillations and jumps about mean

I have a signal recorded for 200ms and total 20 blocks are recorded.

1. For one of the sensor signal oscillates about the mean for all the blocks. in this case the mean and sigma between the blocks is nearly same and signal can be treated as stationary
2. for other sensor (located in transonic shock zone) The signal for each block (200ms) shows oscillations around almost a fixed DC value(mean) for most of time, but in between (say about 10times) signal *jumps to high value for few milliseconds * and comes back and again oscillates about DC(mean) value. The signal is also not clipped (i.e. signal within sensor threshold)

     In 2nd case also the mean as well as sigma is nearly same from one block to other, but can this be considered as stationary signal, because as i mentioned above that within the 200ms (one block) the signal shows a kind of jumps around the mean many times.

Can i use spectrogram.m for computing the Power spectral density and OASPL subsequently.

Answer 1

[Wayne King]

Yes, I think you can. Use the optional output argument, P, for the short-time PSD estimates.

 [S,F,T,P] = spectrogram(...)

[Vinod]

I am still not clear about the stationary aspect of the signal (2nd bullet). How can i upload the figure here to show you the variation of signal with time, so that it is more clear.