The evolution of processor capabilities used in smartphones has improved dramatically in recent years and many of us now carry far more processing power in our pockets than is necessary to run timing machines like Witschi’s Watch Expert or the classic Vibrografs. Such timing machines are expensive, though, and far from portable.
I am blessed to have a Watch Expert II to use at work, but can’t afford to spend the necessary coin on a decent timing machine to use at home on “hobby projects” and watches for friends or family. When speaking with acquaintances or potential clients about their watches while not at work, I’ve often wished I had timing machine in my pocket to help me diagnose problems and offer better advice on the spot. When working on the precision timing of chronometer grade wristwatches, like Rolexes, I’ve also often wished that Witschi’s Watch Expert displayed timing data to within 1/10 of a second. All of these frustrations, coupled with the powerful tools available through Apple’s Developer Program, are what led me to develop Kello.
The app works by using the microphone of the iPhone, iPad, or iPod Touch to analyze the regularity of the ‘tick-tock’ sound generated by the escapement of a mechanical watch. I designed Kello to automatically detect the frequency of the watch being analyzed as well as the beat error of its escapement. Using the data collected, Kello compares the rate of the watch against the much more precise internal time signal of the iPhone and then outputs how much time, in seconds, the watch is gaining or losing each day. If the watch is running within +/-10 seconds/day, Kello will display this gain or loss to a precision of 1/10 of a second.
I have been testing the app for a little over a month now on a second generation iPod Touch as well as on a kind friend’s iPhone, and have been happy with the results. The only caveat I’ve found in testing it in the wild, is that analysis has to be performed in near perfect silence to yield optimal results. I am working hard to bring honed frequency analysis to the next version of Kello to greatly eliminate timing anomalies due to external noise.
In the mean time, I am really looking forward to seeing how it performs for other watchmakers, as well as watch collectors, particularly on Apple’s faster A4 processor, and I intend to evolve it into a much more sophisticated, pocket-sized timing machine, able to perform on par with a Witschi, as the native iOS APIs and iPhone hardware continue to improve.
In light of the tidal wave of feedback I’ve received since Kello went live on the App Store, I realize that I babied the app too much in testing and that it’s not quite ready for full public consumption. Due to the large number of people who have had trouble getting Kello to work successfully, I’ve decided to pull the app from the store and head back to the drawing board to focus on making it easier for the end user to derive reliable results without having to toy with microphone positioning or eliminating ambient noise.
Update 2: Kello is back on the iTunes App Store, updated with new signal processing algorithms to focus in on the sound of the watch’s escapement and help eliminate ambient noise.