I have a low frequency signal, it's a heart beat signal (not an EKG, a pulse ox, its a much simpler shape).

So the heart beats ~.5 Hz to 4.5 Hz

I'm trying to eliminate the wandering baseline through a high pass filter

I used a single pole RC filter (50k Ohm and 220uF) (cut-off is .015 Hz) and found that it works really well except that it takes FOREVERRRRRRRRRRRR for the DC offset to get filtered out.

I don't know an easy way to do a Fourier Tform of the heart beat waveform. I realize that a heart beat is 2 Hz, but is it's frequency spectrum actually composed of much higher frequencies? Could I have a high pass filter with a cut off frequency of .5 Hz and still see the low heart beat?

I hope I explained that well, I found explaining it difficult.

 

A HP filter with a 0.5Hz corner will have a gain reduction of -3dB for a 0.5Hz heart beat. A corner of .175Hz will give a gain reduction of 0.5dB at 0.5Hz. The latter value would give a settling time-constant of .91s so would take about 4-5s for the DC transient to settle out.

 

But the heart beat isn't a typical wave, so even though it's periodic at .5 Hz, is it possible that its mostly comprised of higher frequency components? So even with a .5 Hz corner.......

I need to find good reading, or maybe I'll digitize the whole signal before filtering.