Tone Synthesis

 

Digital signal processing allows us to sculpt each tone to sound like different instrument. This allows for a variety of sounds to be created for each element.

             

Atom Tones encompasses my varied interests. While the fundamentals of turning atomic spectra into audible tones began as an educational exercise, it also has roots in my physics outreach work as well.

The basis of translating the spectral lines to audible frequencies is a simple mapping technique.

 

The correlation between the spectral lines and the audible frequencies can be obtained by measuring the distance between each segment (i.e., 0 – 125 Hz) and then measuring the distance to the line in question. A simple ratio determines the corresponding audible frequency of the spectral line.

 

I used this method to obtain both the audible frequencies and the original wavelengths which was then compared these to the values listed in the Atomic Spectra Database at the National Institute of Standards. All wavelength intensities were taken from these. Initially, the measured wavelengths were within a 4 nm margin of error. This was primarily due to the aligning each element’s spectra onto the scale shown above. Since there are 1000 Hz per 300 nm, a conversion factor of 3.33 Hz per 1 nm was used. I recalculated all frequencies using this conversion factor. Rather than measuring each bright line on an image, the wavelengths of brightest lines for each element were looked up on the Database and were then converted to an audible Hertz scale.

The following shows a step-by-step sequence of synthesizing the sound of Helium.

 

The spectrum of Helium is superimposed onto the frequency conversion scale for comparison.

Helium data obtained from the Atomic Spectra Database (nist.gov/pml/atomic-spectra-database)

The frequency and relative amplitude of each spectral line can then be entered into a digital audio program. Most will have a waveform generator function built into it. The one I am using is Amadeus. I have listed a few options here.

When entering the values, I scaled the relative amplitude % in half to avoid clipping of the added signal.

These are then combined to form the base raw sound of Helium.

Which sounds noisy.

 

An exponential decay is then applied to this waveform.

and now has a pleasant sound to it.

The waveforms of the tones themselves also create unique patterns. Some of these show simple patterns that we might call musical and some appear very noisy. Here are a sample of some of the waveforms –