[FAQ] Numbers, numbers, numbers

Collection of all FAQs and HOW-TOs posted throughout the system. Read only.
Post Reply
User avatar
Harry
Site Admin
Site Admin
Posts: 10518
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

[FAQ] Numbers, numbers, numbers

Post by Harry »

Hi,

Thought I post all the numbers / specs here - instead of having them spread across the forum:

Video
  • Videos are mostly recorded at 30Hz, there are higher rates available to allow slo-mo effects; this part is not covered here
  • Memory used for videos depends on the so called bitrate - the number of bits available per second of a video. For HD video, a typical bitrate is 10.5Mbit/s. Add some additional space for audio and we are at approximately 80MB/min. For FullHD video - supported by the iPhone4S and later, a typical bitrate is 24Mbit/s, summing up to approximately 190MB/min. For 4K UltraHD video - supported by the iPhone6S and later, a typical bitrate is 50Mbit/s, summing up to approximately 400MB/min.
  • HD resolution on an iPhone is 1280x720, Full HD resolution is 1920x1080, and Ultra HD is 3840x2160.
  • The time required to overlay compared to the original video's length depends on video resolution, phone generation, LapTimer version, and OS version. Late iPhones require less time to overlay a video than the video's length - even for FullHD (factor 0.5 to 1.0). As an example, a 20 minute session may require 10 minutes to overlay. Android smartphones supporting hardware accelerated video encoding (Android > 5.1) will require between double and four times the session's time (factor 2.0 to 4.0) while older Android phones may take up zoo a factor 20.0 (software encoding).
  • This actually means LapTimer is able to overlay videos in realtime - in theory. We will see what is possible in the future... The progress in the overlaying is pretty exciting.
Data Recording
  • Starting with v15, LapTimer is capable to record more than 4bill of GPS positions. For both 1Hz and higher update rate GPSes this means there is practically no limitation for GPS recording.
  • Memory is no real limitation for data recordings, as one fix (including acceleration, OBD, everything) requires 78 bytes only. Going back to the memory per minute discussion, this means data recording requires 4.6kB/min for the internal GPS, and 23kB/min for a 5 Hz external GPS.
Timing Accuracy

Timing accuracy has been improved by sophisticated interpolation technics from the beginning; in addition, both the iPhone's internal sensors, Apple's location service implementation, and LapTimer's direct integration of sensors have improved over time; check the table below to get rough numbers on timing accuracy...
  • iPhone3G's and 3GS's internal GPS: better than 0.5 seconds
  • iPhone4s and 4Ss internal GPS: better than 0.1 seconds
  • iPhone5s seem to deliver same accuracy and update rate like the iPhone4s
  • Any iPhone directly integrated with a supported external sensors (Emprum's Ultimate GPS, Dual's XGPS150, TomTom car kit, and G-Fi): around 0.05 seconds, 5 Hz device better than 0.05 seconds
  • Android Smartphones allowing access to their NMEA data deliver accuracy around 0.05 seconds
Please note accuracy reported by GPS sensors are statistic values achieved 95% of the time. You need to expect deviations from the above accuracy accordingly.Furthermore, timing accuracy depends on the speed when passing the trigger. As an example, timing accuracy will be lower for karts as they rare typically slower than cars or motorbikes.

The interpolation algorithms used by LapTimer generate the best possible accuracy a GPS based system can achieve. So accuracy is a function of GPS quality and nothing else. In case your requirements are not met by the above values, the only option is to use a transponder system. These will certainly deliver lap times only and completely miss data recording.

to be continued

- Harry

P.S. iPhones have GPS sensors integrated in a so called RF chip featuring all kinds of net services like UMTS, LTE, etc. The iPhone5 uses a Qualcomm RTR8606 while the iPhone4S uses a Qualcomm RTR8605. The iPhone5 adds LTE cellular network support, so this is probably the change in version number. As the internal GPSes are good enough for location services as used for 'normal' iPhone users, I doubt a lot of development efforts are going into the GPS part in the future. For better performance, specialized devices will be the best choice (see Compatibility).
Image Image Image Image
Baurice
Fewer than 10 Posts
Fewer than 10 Posts
Posts: 2
Joined: Mon Sep 24, 2012 3:06 am

Re: Numbers, numbers, numbers

Post by Baurice »

Hi Harald
A question on your comment on Timing Accuracy
iPhone3G's and 3GS's internal GPS: better than 0.5 seconds
iPhone4's and 4S's internal GPS: better than 0.1 seconds
Any iPhone directly integrated with a supported external sensors (Emprum's Ultimate GPS, Dual's XGPS150, TomTom car kit, and G-Fi): around 0.05 seconds, 5 Hz device better than 0.05 seconds
I've been using my 4S on a Tom Tom mount for quite a while now and its great. I really like your product and have been recommending it here in NZ.

I'm now thinking of upgrading and would like to improve the accuracy, From what I understand about the TomTom mount its' GPS is only 1Hz which is worse than the 4S GPS chipset can deliver. However you say above, that if I use the TT mount with my 4S, accuracy is improved (to about 0.05s - btw is that per lap? or per timing point?)
So.......my question is
How much better will the set up be if I use the TT mount (for good video stability), and add a DUAL XGPS150A? Or should I wait for the iPhone 5?

Cheers for now
Andrew
Christchurch
New Zealand
User avatar
Harry
Site Admin
Site Admin
Posts: 10518
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Numbers, numbers, numbers

Post by Harry »

Hi,

The iPhone4/4S/5 deliver 1 Hz, like the TomTom. The improvement in accuracy when using the TomTom comes from LT's ability to use the original GPS timestamp instead of depending of runtime effects from iOS. There is no reason to change to iPh5 for accuracy reasons, what you can do is switching from TomTom to the Dual. The TT mount can be replaced in a situation like this by something more stable - see www.gps-laptimer.de/Compatibility.html - Harry's Choice.

- Harry
Image Image Image Image
trackday
Fewer than 10 Posts
Fewer than 10 Posts
Posts: 2
Joined: Wed Oct 10, 2012 9:26 pm

Lap Time differences

Post by trackday »

I'm new to this forum, but did a search regarding lap time accuracy and found one topic relative to GPS accuracy of the iPhone 4, 4S, 5, etc. I apologize if this has been answered before.

What I'm looking to report or found out is why my Harry's Lap times are substantially different than ones recorded using a video feed from my goPro. Calculating the time expired per lap using standard goPro video shows lap times in the 2:17-2:20 range but Harry's reports the same laps as 2:20-2:23.

I'm wondering how the lap times are calculated, I assume based on the time elapsed between crossing the start/finish line, but if that's the case, the start/finish line timing using the goPro indicates less elapsed time.

thank you for any advice or explanation in advance.
User avatar
Harry
Site Admin
Site Admin
Posts: 10518
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Lap Time differences

Post by Harry »

Hi,

Firstly, we need to clearly differentiate between lap times and video recording times. I'm not sure which numbers you mention above, so please check that first.

Lap times are highly accurate when recorded with LapTimer, details on this can be found here: viewtopic.php?f=6&t=459 In case you record a video using goPro you actually do a start and stop anywhere. Comparing the exact *frame times* of the start / finish line when passing the first and the second time should reflect the above lap times pretty well. In case you record a video using your iPhone / LapTimer, the start of your first video will have a lag of 2 seconds or so compared to the lap times. So the lap's video recording will start around 2:00.00. Reason for this is LapTimer detects a trigger passed up to 1 seconds *behind* the start / finish line, starting video recording takes iOS another second to start. Please note that although the video is late, lap times are accurate.

- Harry

I will merge this post with viewtopic.php?f=6&t=459 later.
Image Image Image Image
trackday
Fewer than 10 Posts
Fewer than 10 Posts
Posts: 2
Joined: Wed Oct 10, 2012 9:26 pm

Re: Lap Time differences

Post by trackday »

While still in the pits, I click Start Driving, I then hit Video, which tells me the video will start after I cross the start/finish line. At the same time I start recording on the goPro.

After the session has ended, I have lap times in the lap list. When comparing the times recorded there with the elapsed times of the goPro video, the goPro is usually a couple seconds shorter for the same laps.

Hope this explanation helps. I can see there being a difference if the video delay has something to do with it.

Thank you for promptly responding. A great product.
User avatar
Harry
Site Admin
Site Admin
Posts: 10518
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Lap Time differences

Post by Harry »

Maybe there is a problem with your video viewer. Have you checked the frame times (you need to use something more sophisticated like a editing software here) of the two frames when passing start / finish. As long as the laps are recorded correctly and with good GPS conditions, you can trust in LapTimer lap times. You may upload one of your goPro vids to youtube (unlisted) and send the link plus the corresponding lap recording (Lap Details -> Export -> .hlprl format), I can check if you want.

- Harry
Image Image Image Image
User avatar
Harry
Site Admin
Site Admin
Posts: 10518
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Numbers, numbers, numbers

Post by Harry »

Just updated the stats on overlaying times (see initial post):
  • With LapTimer v15.4 or later, an iPhone 5 requires approximately 90% of the FullHD video length for overlaying that video. The sample tested has been a 9 minute video which has been overlaid with lap data in approximately 8 minutes. The iPhone5S saves approximately 30% of that time - which is a huge step and results in 5 to 6 minutes for our sample.
  • This actually means LapTimer is able to overlay videos in realtime - in theory. We will see what is possible in the future... The progress in the overlaying is pretty exciting. For the iPhone4, it took LapTimer approximately double the time of the original video - in HD only.
- Harry
Image Image Image Image
Post Reply