# What statistics test to use for signal analysis

#### corinne12345

##### New Member
Hi! I am confused about which stats test to use for this problem.

I have measured the flow (continuous variable) over a minute of a certain device A using two different testing machines B and C for a verification procedure. Each machine B and C has an error consideration of +- 0.05. The flow wave is characterised by a repeated morphology that repeats according to a frequency f. So essentially, I have two continuous signals in the time domain that should be the same but due to error are not. I want to compare the similarity of these datasets (from B and C) to see if they are significantly different. I am confused about how many variables I technically have. Do I use correlation coefficients? Do I use t tests, or perhaps Fishers z? Or something different?

Help is appreciated.

#### Miner

##### TS Contributor
I am not familiar with statistical signal processing, but have used FFT (Fast Fourier transform) to convert time series signals into the frequency domain for analysis. This allows you to evaluate the frequencies, harmonics and side-bands. Have you considered or tried FFT?

#### corinne12345

##### New Member
I am not familiar with statistical signal processing, but have used FFT (Fast Fourier transform) to convert time series signals into the frequency domain for analysis. This allows you to evaluate the frequencies, harmonics and side-bands. Have you considered or tried FFT?
Hi Miner! I've seen suggestions for using FFT for signal analysis and while I think I understand what it does, I'm not sure how it would be useful for my application. Given that the signals share the same frequency and start simultaneously, all I really need to do (I think) is get a measure of how related the curves are in terms of the y axis values (in this case, flow rate). I come from a psych stats background and am struggling to convert psych stats into this engineering application as it doesn't involve people groups as a variable.

#### Miner

##### TS Contributor
I think I understand part of your scenario, but have a few questions:
• Are testing machines B & C measuring device A simultaneously?
• What do you mean by a repeated morphology? Do you mean a pattern in the signal that repeats itself such as that a heartbeat would display on a monitor? And the frequency being analogous to the heart rate?