DFT

>>8013019
Why don't you just do it and see what happens? Do you actually know which frequency you're expecting the desirable portion of the signal to show up at?

>>8013026
unfortunately, i don't know what I should be looking for. but I figure this might be better than nothing for my project.
i guess i'll try it.

>>8013036
Yeah, just plot it and see what happens. Hopefully you can narrow down the signal to a range of frequencies, and hopefully that range of frequencies is nicely separated from the power spectrum of the noise (probably not). Then you can just do something simple like a high/low/band pass filter. If that doesn't work, you can always just try a sliding average/freq-domain sinc filter.

>>8013036
you should really do some research about what frequencies the signals you're interested in are.
You can always start by modeling the noise as a white noise, and estimate it.

When you know that, try Wiener filtering your signal.

>>8013296
>wiener filter
>to get the white stuff out
Science!

Are you trying to filter out noise OP? Or are you looking for the prescence of specific frequencies?

Perhaps you should use the Fastest Fourier Transform in the West

File: 1415572176868.jpg (21KB, 365x505px) Image search: [Google]

21KB, 365x505px

>>8013547
>yfw your dna sequence is white noise

File: 1453401428825.png (201KB, 331x378px) Image search: [Google]

201KB, 331x378px

>>8013556
>>8013565

the former but I have to do the latter as well first because I don't really know what frequencies I want, so I'll just be operating on the assumption that I'll find the power of the signal frequencies over time will be greater than that of the noise.

anyway, thanks for your help everyone

>>8013961
keep us posted OP, I'm interested in what you find.

One thing I want to mention: you should have studied the signal you're interested in BEFORE using any hardware.

You need to know beforehand:
-frequencies
-power
of the signal you want to acquire.

This allows you to choose adapted hardware to acquire the signal.

Then the processing is straightforward from your observations and your expectations.

>>8013988
A bit more specifically: you need a sampling frequency at least twice that of the highest frequency component you will be looking for. Look up Nyquist.

Skin conductance might be affected by pulse so you might want to sample that too, at least just to make sure it does not impact your experiment significantly.

>>8013988

unfortunately, I'm a poorfag, I just have an arduino, some electrodes, and a circuit board I had printed for the sensor.

I got the schematics for the sensor circuit board from this paper:
http://affect.media.mit.edu/pdfs/10.Poh-etal-TBME-EDA-tests.pdf

I designed it in Eagle and then had it printed.

Now I'm using a Java program to plot the data and store it in an array.

I'll definitely keep you guys posted tho.

>>8014057
>A bit more specifically: you need a sampling frequency at least twice that of the highest frequency component you will be looking for. Look up Nyquist.

hm, that's a good point, I'm not sure what the sampling rate of the analogRead() of the arduino is.