Custom synth waveform using heart rhythms


Pill Pusher
My brain scares me sometimes...

I'm a pharmacist, and work in a hospital. I see electrocardiograms (EKGs or more appropriately ECGs) all the time. The healthcare worker in me sees normal sinus rhythm...


...the musician in me sees a synth waveform.

I just think it would be really cool to use a single cycle of an ECG waveform as a synth waveform. It would be an awesome Easter egg for anyone that, say, imports a song with a synthesizer solo using this waveform into a DAW and sees a heart rhythm.

Am I completely off my rocker, or is this possible? Would be really sweet with Torsades de Pointes!



It should certainly be possible with a synth or soft synth that lets you use custom waveforms. It's also possible to analyze the waveform to see what the harmonics are.


Okay, I apparently had nothing better to do today, so I spent several hours making a waveform based on the ECG graph you posted. (Feel free to pay me handsomely for my efforts, preferably in cash-- although I'll settle for a cold beer as long as it isn't PBR!)

What I did was try to create a reasonably "focused" graph from the one you posted (which was kind of blurry), restricted to one cycle and scaled down to a 256-by-255-pixel grid. Then I used a nifty program I found that takes a bitmap file and converts it into an Excel spreadsheet. I used the Excel spreadsheet to come up with 255 samples or slices of said waveform, and wrote a quickie FreeBASIC program to read the 255 samples and write them to a WAV file. I also pulled the WAV file into WavePad so I could do a FFT analysis of it.

Attached are three files:
-- an MP3 conversion of the original WAV file;
-- the waveform as displayed by WavePad; and
-- the FFT analysis of the waveform.

The WAV file is programmed to play an A4 (440 Hz) with the waveform, for 5 seconds, so you can hear what it sounds like. It sounds a lot higher than A4 because of all the harmonics. Obviously, there's no ADSR envelope applied to it.

The values of the samples are unsigned 8-bit integers ranging from 0 to 255, whereas the original graph has a gray line near the bottom-- so the samples I generated extend equally above and below the "middle line," whereas in the original graph about three-fourths of the waveform is above the gray line and about one-fourth is below it. I assume that the waveform would sound a bit different if I were to shift the sampled values upward, since the harmonics would presumably be different. Maybe I'll try doing that later just to see.

View attachment Heartbeat.mp3




I don't know about flatlining, but it's very "rich" in harmonics, both odd and even. It's definitely not a "soft" tone-- more a high-pitched, somewhat shrill tone.


Well-known member
DC offset shouldn't make it sound any different unless it causes clipping. That offset is definitely stripped out coming out the DAC, if not at some point before.