heart rhythm visualization and fft analysis

pagarreton's icon

I am working on a heart rhythm visualization in realtime project and I need to make differences from one person to other. I am trying to do an fft analisys with zsa.freqpeak~ seeking for differences between different hearts. I'm trying to compare the list of frequencies, and trying to determine which is noisier than the other but I don't get consistent results over the time. I am using a some filters for trying to isolate a part of the spectrum, supposing that what we call a noisier heart is located on the upper harmonics. My goal would be to get at the end only one parameter, something like "noisiness", for each heart, maybe with the comparison of an "ideal" heart with no noise.
Any ideas? any help is welcome

Zancudo's icon

Hi Pagarreton,

I am a little confused by your project. There are two different sound waves which sound a bit processed. As either an acoustic or electrical (ECG) signal, the waveform of heart function is very slow. Noisiness is not a heart-related variable as far as I know.

The differences in the sound qualities of the example files is more related to how the sounds were recorded or wether the files were processed afterwards than to the characteristics of the hearts themselves.

Hope my comments are useful

pagarreton's icon

Thanks for your comments, unfortunately, this are the sound I will be working with. They are recorded with a doppler ultrasound system and yes it has a lot of noise. I would not have purchased this, but I was invited to a project that was already running.
Anyway I found a very nice object called noisiness from http://web.media.mit.edu/~tristan/.