Cell Phone Clock Accuracy

Just how accurate are the clocks on smart phones?
A quick Google comes up with all kinds of problems.
Has anyone done a study?
We are looking at an app that downloads a sound track to people's smart phones and starts the music at a specific time using the phones clock.
Fireworks are programmed to a resolution of 1/100th of a second.
Do you think a 1000 phones would be within 1/100 of a second of each other?

4pyros said:

Just how accurate are the clocks on smart phones?
A quick Google comes up with all kinds of problems.
Has anyone done a study?
We are looking at an app that downloads a sound track to people's smart phones and starts the music at a specific time using the phones clock.
Fireworks are programmed to a resolution of 1/100th of a second.
Do you think a 1000 phones would be within 1/100 of a second of each other?

Click to expand...

Sounds HIGHLY unlikely

If they all had GPS activated, time would be more precise, since GPS is uses atomic clocks.

Its worth noting that GPS clocks actually run very slightly faster than down here on earth, and have to be compensated as such.

The problem with clocks on smart phones and similar clocks is They rely on periodically re-synchronize with NIST or other time sources. Some apps sync every 10 minuets and some maybe once daily. The clock application on my phone is typically within 4 seconds of UTC(NIST). This is more than adequate for my alarm clock functions which I have plenty of. NIST has a few apps out there for smart phones like this one that does update every 10 seconds and allows for the offset. Now with that said and fireworks being times to 0.01 second do we mean 0.01 second with respect to the actual time of day or 0.01 second just as an uncertainty of time, as in a clock signals accuracy?

Ron

Reloadron said:

The problem with clocks on smart phones and similar clocks is They rely on periodically re-synchronize with NIST or other time sources. Some apps sync every 10 minuets and some maybe once daily. The clock application on my phone is typically within 4 seconds of UTC(NIST). This is more than adequate for my alarm clock functions which I have plenty of. NIST has a few apps out there for smart phones like this one that does update every 10 seconds and allows for the offset. Now with that said and fireworks being times to 0.01 second do we mean 0.01 second with respect to the actual time of day or 0.01 second just as an uncertainty of time, as in a clock signals accuracy?

Ron

Click to expand...

We can program a shoot every .01 second, the music needs to be within .01 of a second of the program, other wise the hole thing will look out of sync.

I have a company where everyone has a phone that used the same calender program, same service, same type of phone, etc. At 8:50 and again at 9:00 every phone will ding to announce 9:00 meeting. There is no 0.01 second syncing of the phones. The time might be exactly the same but the "dings" happens over 30 seconds. Then one phone might be a good minute or more off. (usually mine) There are many times where the dings are within 5 seconds.

I do think you could make a application that is much better than the calendar program. I question if you can get under one second.

Overclocked said:

If they all had GPS activated, time would be more precise, since GPS is uses atomic clocks.

Its worth noting that GPS clocks actually run very slightly faster than down here on earth, and have to be compensated as such.

Click to expand...

Now do those clocks run faster or does the planet run slower?

Ron

4pyros said:

We can program a shoot every .01 second, the music needs to be within .01 of a second of the program, other wise the hole thing will look out of sync.

Click to expand...

OK now I see where this is going. We don't care what time it is but we care about making sure the display is in sync to the music within 0.01 second. I mean how would it look (and sound) if the explosions in the 1812 Overture were out of sync? That would suck for the cannons.

So I get to thinking about the software used for Christmas Lighting Displays. Where the light show is in sync with the music. Christmas lighting design software which ranges from Free to I can't believe it cost that much. Software like this actually allows you to see your sound tracks and program the lighting to the sound track(s). You fire a string of lights or fire a launch tube, there has to be a similarity in there. A laptop driving a digital out shouldn't care what is out there. You look at a sound track with respect to time and use the timeline. Wouldn't something like this work?

Ron

Reloadron said:

OK now I see where this is going. We don't care what time it is but we care about making sure the display is in sync to the music within 0.01 second. I mean how would it look (and sound) if the explosions in the 1812 Overture were out of sync? That would suck for the cannons.

So I get to thinking about the software used for Christmas Lighting Displays. Where the light show is in sync with the music. Christmas lighting design software which ranges from Free to I can't believe it cost that much. Software like this actually allows you to see your sound tracks and program the lighting to the sound track(s). You fire a string of lights or fire a launch tube, there has to be a similarity in there. A laptop driving a digital out shouldn't care what is out there. You look at a sound track with respect to time and use the timeline. Wouldn't something like this work?

Ron

Click to expand...

Yes, it's only the phones that's the problem.

Yes we already have firing systems that sync using SMP time code. Its getting the music to the masses that's the problem.
One guy said he made an app to send the music to the phones but I don't know if cell phone clocks are accurate enough.

If you start synchronizing with some sort of a central server (NTP might be ok) few minutes before the time and do it well, then you probably can go to 0.2-0.3s accuracy, may be even better.

But then again, if you start playing a sound, there's no guarantee of immediacy.

Reloadron said:

Now do those clocks run faster or does the planet run slower?

Click to expand...

It's all relative to the observer. To the observer on the ground the satellite clock runs slower.

Reloadron said:

Now do those clocks run faster or does the planet run slower?

Ron

Click to expand...

There are two things that affect the clocks, gravity(GR) and speed (SR). The GPS clock moves faster IRT the clock on earth so there is a slowing per day (~7us) but it's in weaker gravity so there is a speedup(~45us) for a net speedup IRT the ground clock.

Reloadron said:

Now do those clocks run faster or does the planet run slower?

Click to expand...

Guys, that was said in humor. I mentioned it because the earth's rotation is slowing down. Enter the occasional leap second added to time. Amazing how a hundred years ago, when I was younger, we weren't as concerned with the measurement of time as we are today. The evolution of the clock is interesting stuff and how the time is actually kept. Anyway, my quote was in humor.

Ron

When I was in school, I adjusted my mechanical watch to be exactly in sync with the school bell clock.
I stood up in class and walked to the door as the teacher and classmates wondered what I was doing. The bell rang the moment I touched the door handle then I opened the door and walked out waving.

But could this work using the phones clock?

4pyros said:

But could this work using the phones clock?

Click to expand...

Over a short time yes, but you would need to synchronise all phones at exactly the same time - just as you would with wrist watches.

But I would suggest your requirements are FAR too strict for synchronising?.

I just want to reinforce what Nigel said. The clocks in cell phones are almost certainly accurate enough for a performance. If the performance is 5 minutes long, that is 300 seconds, so 0.01 seconds is 1 part in 30000 or about 30 ppm, which most if not all cellphones will meet. They are made like that because it is quite easy and so that the clocks keep good time. If you ran a 30 ppm clock for a month, it could be out by a bit over a minute, which is fine for a cellphone.

The problem you have is the initial synchronisation. There is no other reason for cellphones to be synchronised to within 0.01 seconds, and with internet delays etc you are going to get a few seconds variation like that.

I suggest that you have some local synchronisation method. That could be bluetooth streaming to everyone, or it could be an app that uses the microphone to align the timing, like choirs do when singing.

0.01 seconds is very tight. Sound travels 3.44 m in that time. If you need to keep the sounds synchronised within that, and you used conventional speakers, you would need to have the entire audience in a 3.44 m diameter area? If so, one speaker would do. There must be around 1/2 second delays at most large concerts.

Diver300 said:

I just want to reinforce what Nigel said. The clocks in cell phones are almost certainly accurate enough for a performance. If the performance is 5 minutes long, that is 300 seconds, so 0.01 seconds is 1 part in 30000 or about 30 ppm, which most if not all cellphones will meet. They are made like that because it is quite easy and so that the clocks keep good time. If you ran a 30 ppm clock for a month, it could be out by a bit over a minute, which is fine for a cellphone.

Click to expand...

The shows are about 20 minutes long.

Diver300 said:

The problem you have is the initial synchronisation. There is no other reason for cellphones to be synchronised to within 0.01 seconds, and with internet delays etc you are going to get a few seconds variation like that.

I suggest that you have some local synchronisation method. That could be bluetooth streaming to everyone, or it could be an app that uses the microphone to align the timing, like choirs do when singing.

Click to expand...

The app would start playing the sound that has already been downloaded as soon as the phones clock gets to a predetermined time.

Diver300 said:

0.01 seconds is very tight. Sound travels 3.44 m in that time. If you need to keep the sounds synchronised within that, and you used conventional speakers, you would need to have the entire audience in a 3.44 m diameter area? If so, one speaker would do. There must be around 1/2 second delays at most large concerts.

Click to expand...

Big sound setups like that use delays to remote speaker stacks.
For the most part we just put speakers in front of a long line of people that may not be very deep but it is wide.

Wounder if anyone has ever done a study on this?