[time-nuts] Visual clock comparison

Tom Van Baak tvb at LeapSecond.com
Sat Apr 18 01:11:47 EDT 2015

That would make a fun time experiment. One that you should do and report back to us.
I see two experiments: one using LED flashes and one using speaker ticks.
In each case output a main and a delayed pulse. Try it yourself, and with a number of friends.
The goal is to find at what level people can discern the difference between, or the order of, two events.
>From the data you can determine both a human average and an average human response.

Use identical LEDs, and identical speakers.
Also, use a common timing source. Do not compare CHU with NTP with PC monitor with PC speakers; all that does is introduce layers of unknown offsets into your experiment. This is would be an easy Arduino / Raspberry-Pi sort of project.
For higher precision, see pd26.asm and pd27.asm ( http://leapsecond.com/pic/src/ ) as an example of precise pulse stepping.

If this turns out to give reliable data, you can then try different color LEDs and different frequency speaker ticks to see what effect that might have on resolution. Also try different pulse durations to see if human leading edge detection is affected by pulse width.

For those of you with a Stanford Research DG535 and two spare LEDs, give it a try right now.


----- Original Message ----- 
From: "d0ct0r" <time at patoka.org>
To: "Discussion of precise time and frequency measurement" <time-nuts at febo.com>
Sent: Friday, April 17, 2015 11:16 AM
Subject: [time-nuts] Visual clock comparison

> Hello, Netizens !
> I am wandering what is the average human ability to visually compare two 
> clocks ? Let say I have XClock application running on one machine 
> (stratum 1 NTP) and I have my project clock close by. And I would like 
> to match the reading. If I'll see the difference, which range it will be 
> ? 100ms or so ?
> I also tried to use my ears (CHU radio signals and clock display, NRC 
> phone line). However NRC "Talking Clock" could be routed via Satelite 
> which will compromise the "reading" a little bit. Thanks !
> -- 
> WBW,
> V.P.

More information about the time-nuts mailing list