Harry's GPS Suite Forum

My impression is there is some issue on system level. I understand correctly you use an iPhone4, latest iOS (7.0.3), a BT1 (no BT1A), and the internal GPS sensor? I have this configuration around and will run some tests once I have a chance to drive around a bit.

- Harry

Harry wrote:My impression is there is some issue on system level. I understand correctly you use an iPhone4, latest iOS (7.0.3), a BT1 (no BT1A), and the internal GPS sensor? I have this configuration around and will run some tests once I have a chance to drive around a bit.

Exactly.
Having the compare to another identical setup would be valuable.
Thanks!

Hi,

I had a chance to run this setup and have some bad news... (or good news in case you want some arguments for a new iPhone). Here are the scenarios I ran with exactly your configuration plus outcome:

1. LapTimer with BT1 connected and Video View active and recording: speed an other displays is not fully responsive; as an example, the time elapsed should get an update 5 times a seconds, but it sometimes lags up to 1 or 2 seconds; RPM is updated from time to time but not smooth; looking into the data recordings I get a similar picture like you had, OBD is not updated and has drop outs.
2. LapTimer with BT1 connected with Timer view active and recording w/o video recording in background: smooth update rate for both GPS and OBD, no drop out, regular refresh of display; the small load gauge in the bottom right shows system is fine (not overloaded).
3. Same as scenario 2 but with video recording in background: same picture as in scenario 1, drop outs in OBD, lagging screen update; the small load gauge shows a red hand - which means the system is overloaded / drops below 5 Hz screen update rate.

My current analysis is as follows: it seems the video recording (independent from preview shown / not shown) generates a very high load for the iPhone4 running iOS7. Remaining processing power is too limited to keep up with regular tasks like display updates or OBD processing. The problem with OBD is incoming packages do not have a timestamp allowing a relocation to a continuous timeline (i.e. values cannot be aligned to GPS time). Furthermore, there is no data sent automatically (like with GPS) - it needs to be requested instead. Due to the lags in processing, the OBD rate drops dramatically and shows time ranges without updates. It is actually no dropped connection, but simply dropped processing. Processing OBD itself (or recording data or refreshing the display) requires limited processing power (10 to 20% load), but depends on regular allocation of processing power. What I want to say is "OBD is not the root cause", the drops are a symptom of a different issue.

As this effect does not happen for the 4S (or or 5S) the reason is probably the iPhone4 has no dual core processor, making regular processing allocations more difficult if one high priority process (video recording) is going on.

I will run some more tests tomorrow and check if it is really only the video processing load or if something else can be changed. To work around the issue, you may reduce video resolution (not tested yet).

Hope this at least throws some light on the background.

- Harry

I was suspecting something in this direction.
Thanks a lot for running those tests on your side to confirm!

I've done yet another try using lower video rates. I'm not giving up easily.

Despite the fact the video is not useful at this resolution when rendered (lot's of information missing as it is cut off from the screen), I did not find a prove the video rate would have an impact to the durability of the recordings.

I first ran a test with "Wifi" resolution. That looked promising in the beginning, but I then faced 30+ seconds drop out during the middle of a 12 minute recording, and another short one later on (see IMG_7728). Still it was better than that with higher resolutions.
But when I was going to the ultimate "3G" resolution, picture looked even worse than before (seeIMG_7730). Have a look not only where it dropped out but also where the graph is going flat horizontal whereas the native GPS measurement is vibrating.

Going to test without video next week.
I'm not 100% confident this is an iPhone 4 video rate issue purely, as 1) it used to work fine before with earlier versions of LapTimer and 2) there is no ratio interpolation of video rates used compared to dropout occurrence visible to me. Still the iOS7 thing might factor in, but probably not the video rate as such.
Still looking into the option of operating a remote cam with iPhone doing data recording only. I have the GhostHD which has Radio and WiFi remote control, but I think is not supported by Laptimer DoublePlusGood Edition, right?.

But one final question, given the findings of yours and mine:
Do you intend to list iPhone 4 still as a supported device?

Remote controlling external cams is supported for iOS devices and the GoPros only currently. I will add support for Sony wifi cams in v18 (that's the plan at least).

Given the new problems that (seem to have) started with iOS7, I changed the recommendation in LapTimer's Compatibility listing some weeks ago already. Removing iPhone4 support in general would most probably not be the wish of 95% of LapTimer users running an iPhone4

- Harry