In order to alleviate the boredom when walking on the treadmill I listen to different podcasts. The other day I got to listen to a Science Friday discussion on “sonfication” which is the process whereby some observational property is transformed into a sound. Perhaps the most recent example of this is the audio presentation of the gravitational waves detected by LIGO. There is also a rather haunting sonification of solar wind data into a “sun song” that you can hear if you go to the link above.
According to research, hearing is more sensitive to variations than the other senses including vision. Even non-technical people can tell when their car is not sounding “right”. By coincidence our own Max Maxfield recently wrote a blog “Introducing the Floppotron 2.0” and several of the comments to that blog detail quite how sensitive the human ear can be. Back in the day of long playing records, there was a discussion that although human hearing couldn’t hear above 20KHz at maximum, the frequency response of the hi-fi system should go to ~25KHz because although the ear couldn’t hear that the sound was there, it could detect that it was missing. But let me return to sonification.
Since the ear is so sensitive, it is much easier to discern changes in the phenomenon being monitored by converting the measured data into sound. There are probably several algorithms to do this and no doubt, one approach may be better than another for a given set of circumstances. Apparently Fast Fourier Transforms loom large in this field.
I must admit that I found the evolution of the word “sonification” from a noun to a verb disconcerting, but at the risk of tooting my own horn (blowing my own trumpet, depending on which version of English you learned) I must confess to have sonified data well in advance of hearing this episode of Science Friday. Those of you who read my ramblings (hi mom) will know that I design analog signal conditioners that need to be calibrated. In an attempt to help the calibration technician from having to keep a small screwdriver located in the slot of a tiny screw whilst monitoring an ammeter, I came up with a technique to calibrate 4-20mA current loops without having to look from screwdriver to meter and back. If the reading was above the target value the test system produced a tone that increased in frequency as the output deviated and conversely dropped as it approached the desired output. If the reading was below the target value, the system worked in the same way, only a couple of octaves lower. The idea was to tune for a null.
The project was implemented on an early PC using C and especially since I am not musically talented, the sounds were rather harsh. I argued that it would motivate the technician to complete the calibration quickly. Co-workers hated the continuous sound and so we provided headphones, but they were not particularly resilient and kept breaking - perhaps it wasn’t normal wear and tear that contributed to their demise! I must admit that it wasn’t as successful as I had hoped. I actually published it as a design idea in EDN as “Calibration technique uses sound” in September 1, 1998, but good luck finding it.
From experience and the tales related in Max’s blog it is obvious there are many instances of passive sonification that we become aware of through conditioning. However, now that I know, it seems that there are several common uses of active sonification, where a designer has made a conscious decision to transform some other signal into sound - for instance in higher end cars the proximity detectors generate sounds to indicate when you are too close to an object and if I am not mistaken they use tones to indicate side and back as well. There must be others and you must have some ideas you would like to implement. Please let me know in the comments below.