NTP Time Servers for VideoIQ Cameras

[drvnbysound]

I use these:
 
http://tf.nist.gov/tf-cgi/servers.cgi

 

[pete_c]

Yup noticed this one in your link above. 
 

NIST WWV, Fort Collins, Colorado

 
I used WWV in the 1960's; what a PITA. 
 
I had to wait for the radio to warm up then tune it to WWV; then adjust the crystal oscillator in the radio to be in sync with WWV.
 
I was a kid and time only mattered to me at the time and a bit of a hobby I suppose.
 
Today I have a few of those radio time clocks around the house (low on the WAF though; except for one of them).
 
In the 60's though the WWV station was synonymous with the use of the Atomic Cesium Clock. 
 

Cesium Atomic Clock

The current time standard for the United States is a cesium atomic frequency standard at the National Institute of Standards and Technology in Boulder, Colorado. In 1967 a standard second was adopted based on the frequency of a transition in the Cs-133 atom:

1 second = 9,192, 631,770 cycles of the standard Cs-133 transition

Prior to 1964 the international standard second had been based upon the orbital period of the Earth, but the cesium clock period was found to be much more stable than the Earth's orbit! The SI unit of time, the second, is now defined by this transition in cesium.

 

Atomic Clocks
Very accurate clocks can be constructed by locking an electronic oscillator to the frequency of an atomic transition. The frequencies associated with such transitions are so reproducible that the definition of the second is now tied to the frequency associated with a transition in cesium-133:
1 second = 9,192, 631,770 cycles of the standard Cs-133 transition
The two most widely used atomic clocks in recent years have been the cesium beam atomic clock and the rubidium clock. Such clocks have provided the accuracy necessary to test general relativity and to track variations in the frequencies of pulsars. Atomic clocks are integral parts of the Global Positioning System since extreme accuracy in timing is necessary for the triangulation involved.

 
http://hyperphysics.phy-astr.gsu.edu/hbase/acloc.html

 

GPS clocks

Many modern radio clocks use the Global Positioning System to provide more accurate time than can be obtained from these terrestrial radio stations. These GPS clocks combine time estimates from multiple satellite atomic clocks with error estimates maintained by a network of ground stations. Due to effects inherent in radio propagation and ionospheric spread and delay, GPS timing requires averaging of these phenomena over several periods. No GPS receiver directly computes time or frequency, rather they use GPS to discipline an oscillator that may range from a quartz crystal in a low-end navigation receiver, through oven-controlled crystal oscillators (OCXO) in specialized units, to atomic oscillators (rubidium) in some receivers used for synchronization in telecommunications. For this reason, these devices are technically referred to as GPS-disciplined oscillators.

GPS units intended primarily for time measurement as opposed to navigation can be set to assume the antenna position is fixed. In this mode, the device will average its position fixes. After approximately a day of operation, it will know its position to within a few meters. Once it has averaged its position, it can determine accurate time even if it can pick up signals from only one or two satellites. GPS clocks provide the precise time needed for synchrophasor measurement of voltage and current on the commercial power grid to determine the health of the system.

 
http://en.wikipedia.org/wiki/Radio_clock
 
It would be nice though if "internet travel" was as fast as the speed of light. 

 

[tmbrown97]

Doesn't matter if his clocks are as accurate as an atomic clock - as long as they're in sync with each other.  That's a reason I like running my own NTP server.  I have an installation or two that just run it on a cheap mikrotik box and they work well; they also mean you don't have to worry about having DNS right on a device (camera) that otherwise would have no use for DNS;  If Pete's device is still around for $30, that's pretty slick too.
 
Have the cameras helped much?  My sister's neighborhood is dealing with a string of burglaries and car breakins right now - with how their community is set up, they'd be a prime candidate for that exact type of camera setup (only one way in/out) but we're concerned that more of their problems could be coming from foot traffic; a license plate cam won't do much for that.  Just curious if it's cut down on your problems at all.

 

[BraveSirRobbin]

It's hard to say if they helped a lot, we still get the occasional car break-ins, some stuff stolen from back yards (of unoccupied homes).  I'd say a couple a year, so not a lot of statistics to go off of.
 
It was more a matter of 'keeping up with the Jones’s' as all the other gated communities in our area were putting them in and I didn’t want ours chosen on some burglars drive around because we were the only ones without them.
 
Be aware that we have an overall camera viewing our front gate area, which includes pedestrian gates, and it can pickup people on a movement capture.  Why it misses the occasional vehicle is beyond my comprehension.  We ALSO have a license plate capture that captures license plates of entering vehicles (this works well at night, hit or miss during the day depending on sun position it seems).
 
We also have a park camera, and this one picks up people very well also.
 
The main problem is, by the time you are notified of a break-in, many days usually have elapsed.  Sure we can look-up all the license plates, but what do you do with this info?  Police are not much help here beyond filling out a reporting form (heck, our police even have stopped going to non-injury vehicle accidents)!
 
Same thing with attempted breakins in front of my house.  I have good footage, even posted YouTube videos, but not much else you can do.  I honestly believe video surveillance is only good as evidence if you happen to apprehend suspected thieves presently in the act (court convictions).
 
If you have a contractor install a system, it is an expensive endeavor.  Just erecting a pole seems to be a big deal (cost wise).  Plus (as seemed to be the case with us) you get a lot of truck slammers that know just surface basics about their technology that is being deployed. 
 
Real shame too as some of the equipment out there is top notch.
 
I don't know if the training is inadequate, or they just want the quick bucks as it really takes some investigation by someone with some technical expertise to call them on some of these problems.
 
I will give you one piece of advice…hide the fact that you know a bit about systems, networks, and such.  Just don’t get involved.  Believe me, you will thank me later!

 

[BraveSirRobbin]

I *think* part of the problem you're having is just basic understanding of analytics/object classification vs. "rules", and how the system really works and makes decisions.  If you email a couple of video clips of scenes/events that you're having problems with to [email protected] I can have one of my techs take a look.  I'll also be at ISC West next month, if you want to come by our booth we can chat a bit there as well.  I assume you already signed up for one of the free exhibit hall passes, but if not let me know and I can get a free exhibit hall pass.

 
I would definitely like to meet you at ISC West.  I'm in the process of enabling your PM's and will send you one soon so I can get your contact specifics.
 
I would like to give this system a chance and get it working.  I want to give this one more good shot, then I'm done.  I don't know what we will do, as there is nobody else here willing to investigate problems and interface with the installer/vendor, but I'm just tiring of this (non productive) process.

 

[pete_c]

If Pete's device is still around for $30, that's pretty slick too.

 
Still same price.  Very sensitive.  The one here is in the attic getting 8-12 GPS satellites.  Outside it does better.
 
You can also do an all in one little Rasberry Pi these days; with PPS and all for little money. 
 
The price you pay for the accuracy/technology is really very little today versus just a few years back.
 
Yup here the IP cams are all time sync'd to each, along with the CCTV DVRs, automation servers et al.  I just pulled a streaming video from the DVR which pulls in the HD IP streaming video; each with time and each the same.  One device off time sync can mess other stuff; domino like creating much havoc.
 
Clocks are a hobby and sort of picky about time in general with my automation. 
 
I have a couple which are from the 1800's.  Nicest clock is a regulator newer but very old design.

 

[pete_c]

Unrelated to OP; noticed that my Grandstream IP HD cams did change time last Saturday; but the Aircam's did not.  Both utilize the internal NTP server.  Looks to be a bug in the Aircam firmware as it is configured to get the time from the NTP server; but appears to be doing its own thing.
 
Its interesting cuz I cannot find a manual time change in the Aircam configuration pages; only an NTP reference.  I am running most current firmware.
 
Its not a big deal cuz the camera was only some $100. 
 
That said though if it was $300-$800 IP HD camera being sold with half ass (half baked) firmware on it I would be sort of upset.

 

[wkearney99]

Unrelated to OP; noticed that my Grandstream IP HD cams did change time last Saturday; but the Aircam's did not.  Both utilize the internal NTP server.  Looks to be a bug in the Aircam firmware as it is configured to get the time from the NTP server; but appears to be doing its own thing.

Time does NOT equal time ZONE. NTP knows nothing of your location and how you interpret time in relation to it.

As in, NTP essentially provides just a counter based on UTC (which is not the 'same' as GMT). How determine the offset from that counter to your local time is entirely up to your device. If your device doesn't have time zone tables and schedules (or it's not configured properly) then it won't adjust accordingly.

Time zones are FAR messier than people first expect. The tables behind using them are quite a lot more complex than most programmers fully grasp.

 

[wkearney99]

Its interesting cuz I cannot find a manual time change in the Aircam configuration pages; only an NTP reference.  I am running most current firmware.

Right, so presumably the firmware presents a selection somewhere to chose your time zone? And perhaps there's also an option to dis/enable daylight savings time? Some devices don't have the ability to alter their use of DST automatically. Mainly because the programming necessary to do it would exceed either the firmware's size or the programmer's ability. This is because not everywhere treats the DST change in the same way. So lots of times the device makers 'give up' and just present a check box or something that allows selecting DST or standard.

 

[pete_c]

Time does NOT equal time ZONE. NTP knows nothing of your location and how you interpret time in relation to it.

 

Right, so presumably the firmware presents a selection somewhere to chose your time zone?

 
Thank you wkearney99.
 
Yup; totally aware of this (over 20 years now).  The Grandstream lets you put it in manually.  The Aircam doesn't ("faux pas" in the firmware).
 
That I have to do this is another half backed "faux pas" in the firmware which sort of pisses me off and I am glad that I am using a CCTV DVR with this stuff.
 
I have been playing much with Grandstream OS boards.  (pieces all over my workbench).  They are little marvel HD IP boards.  I'm thinking the adds for DST auto changes would fit with their current firmware.  Not sure on the Ubiquti Aircams.  Just testing beta stuff; but not commenting lately on their forums.
 

Time zones are FAR messier than people first expect. The tables behind using them are quite a lot more complex than most programmers fully grasp.

 
You are right.  I personally see this as a bit neglected. 
 
The issues cropping up have been related to a "catch up" to the whole time (DST) changes which occurred a few years back. Stupid stuff...that actually caused many issues...some idiot tried again just in the last month to change it again....to two hours instead of one...
 

Changing clocks and DST rules has a direct economic cost, entailing extra work to support remote meetings, computer applications and the like. For example, a 2007 North American rule change cost an estimated $500 million to $1 billion,and Utah State University economist William F. Shughart II has estimated the lost opportunity cost at around $1.7 billion USD.Although it has been argued that clock shifts correlate with decreased economic effienciency, and that in 2000 the daylight-saving effect implied an estimated one-day loss of $31 billion on US stock exchanges,the estimated numbers depend on the methodology and the results have been disputed.
 
In 2005, the Sporting Goods Manufacturers Association and the National Association of Convenience Stores successfully lobbied for the 2007 extension to US DST. In December 2008, the Daylight Saving for South East Queensland political party was officially registered in Queensland, advocating the implementation of a dual-time zone arrangement for Daylight Saving in South East Queensland while the rest of the state maintains standard time. DS4SEQ contested the March 2009 Queensland State election with 32 candidates and received one percent of the state-wide primary vote, equating to around 2.5% across the 32 electorates contested. After a three-year trial, more than 55% of Western Australians voted against DST in 2009, with rural areas strongly opposed. On 14 April 2010, after being approached by the DS4SEQ political party, Queensland Independent member Peter Wellington, introduced the Daylight Saving for South East Queensland Referendum Bill 2010 into Queensland Parliament, calling for a referendum to be held at the next State election on the introduction of daylight saving into South East Queensland under a dual-time zone arrangement.[66] The Bill was defeated in Queensland Parliament on 15 June 2011.

In the UK the Royal Society for the Prevention of Accidents supports a proposal to observe SDST's additional hour year-round, but is opposed in some industries, such as postal workers and farmers, and particularly by those living in the northern regions of the UK.

 
http://en.wikipedia.org/wiki/Daylight_saving_time
 
This indeed did cause some major issues.  That said now its been a few years since these changes and folks are still selling products which have never been fixed relating to these.  That issue was what I was sort of bringing up in the above post.  I mean its been a few years now; way more than enough time to fix stuff (firmware?).
 
Paying a bit closer attention noticed that the Grandstream cameras are off too.
 
What a PITA this is turning out to be.
 
Just noticed though that the time change is still a manual thing on the Grandstream and I will need to change the following:
 
This parameter allows the users to define their own time zone.  What is good is that the CCTV DVR does overlay the time on  the videos and this is correct.
 
 

The syntax is: std offset dst [offset], start [/time], end [/time]
Default is set to: MTZ+6MDT+5,M4.1.0,M11.1.0MTZ+6MDT+5,
This indicates a time zone with 6 hours offset with 1 hour ahead which is U.S central
time. If it is positive (+) if the local time zone is west of the Prime Meridian (A.K.A:
International or Greenwich Meridian) and negative (-) if it is east.
M4.1.0,M11.1.0
 
The 1st number indicates Month: 1,2,3.., 12 (for Jan, Feb, .., Dec)
The 2nd number indicates the nth iteration of the weekday: (1st Sunday, 3
Tuesday...)
The 3rd number indicates weekday: 0,1,2,..,6( for Sun, Mon, Tues, ... ,Sat)
Therefore, this example is the DST which starts from the first Sunday of April to the
1st Sunday of November

 
Guessing the best way to deal with this is just a manual DST time entry that the camera remembers in the firmware in case it ever happens again. (a spring and fall one).  Well similar to what HAI / Elk do with their alarm panels (I am only familiar with the HAI though).

 

[pete_c]

BTW noticed that the default date change (spring ahead) on the Grandstream cameras is incorrect. 
 
So if you change it as so; it will automatically change.
 
You can change it on the Ubiquti Aircam such that it automatically changes by editing the /etc/TZ file to read the same by SSHing to the Aircam and saving TZ change.  (there is no menu option for this)
 
This is not documented and Ubiquiti support answer is not to use the clock on the Aircam as a reference but rather the Ubiquti DVR software. 
 
Tangent note:  Noticed the same issue with the Ubiquiti AP.  (No GUI item for autochanging time and a manual TZ thing needs to be done). 
 
You can't really see the local time (just GMT on main page) in any of the pages; you have to look at the logs for the time.
 
MTZ = Mountain Time Zone or same as MST/MDT
 
Auto time zone change format for self defined timezone entry under system tab.
 
Look here for your timezone. http://famschmid.net/timezones.html
Look here for your timezone date change: http://www.timeandda...e/dst/2014.html
 
Example: CST+6CDT+5,M3.1.0,M11.1.0
CST = Central Standard Time + 6 (GMT)
CDT = Central Daylight Time + 5 (GMT)
M3.1.0 = 3rd Month, 1st Sunday
M11.1.0= 11th Month, 1st Sunday