known bugs from beta tester

Okay got some more to share, and on video this time ( though OOF as action camera was too near to LCD screen )
So i have just modified the Sj6 air to near field focus so i can retake video tomorrow.

But what i found out and replicated several times was:

If you let the camera go into parking mode with the engine still running, and then after the camera have gone into parking mode turn off the engine a little later.
Then after letting the camera sit and do parking mode recording ( i assume don't have memory card as i left that in the car overnight ) well then when you start your engine again the camera remain dark and make no sound.
BUT ! it is recording something, cuz if you poke the camera the G sensor wake it up and the LCD screen come on, and it change to regular recording ( so i think it are still in parking mode though i have started the car again )
But the camera don't respond to changes in power but only to G- sensor, but i would assume if the camera are in parking mode and detect whatever on the power it should change to normal mode.

Which it also do fine if you turn off car before camera go into parking mode, in that case the camera record regular for a little while then change to parking mode and turn off LCD screen, and then upon turning on car again the LCD screen come on and the camera change to regular recording as one would expect.

This are probably not that big of a problem as people would not sit still in a car long enough for the camera to go into parking mode before you turn off the car.
Tomorrow i will also test what happen if you let camera go in parking mode with engine running, but then rummage around to get G-sensor to wake camera up again, but i assume it transition just fine into normal mode again, but better check that to be on the safe side too.


I call this issue No wake up screen, the camer are awake, just not 100% sure what it is doing in this case until G sensor wake it up, which you getting into car and drive off would do of course.
 
Next problem, and this is more of a problem, but i only managed to make the camera do it one time out if 10 - 12 cycles.
But it is all normal, camera are recording - i shut off car - camera record for a little while and then change to parking mode for 2 minutes ( speed up in edit ) - then i turn car back on- and camera come on too,,, just to shut down seconds later.
I call this issue power off see you.

 
I think I got your point now.
1, you pluged both OBD and microUSB cigar charger for power, when the engine off, K2S will be off. ---- it is because only one power source will provide current to K2S; when cigar charger working, the OBD power unit will be standby then go to sleep, then the turn off occur. It is not necessary to power the K2S with 2 power source unless you can make sure both power source have exactly same voltage.
2, you need to use the supplied data cable to transmit OBD information to K2S. Other USB-C cable from market will not work.
 
Sound reasonable about the OBD cable, it was too short so i got a longer one, will remember this when i relocate devise from current position, where i plan to move the main unit to the original OBD wire will be plenty long.
The cigar charger have just been plugged into the camera i do not use power that way when using OBD as i was told that is bad, only on day 1 did i use the 2 ways of power at the same time, since then i just use 1 route of power at the time.

I am a little unsure of my tests today was with only the OBD cable connected ( 5 wire plugged into camera ) or if the cigar charger was also plugged into camera but no powere in wire as USB A was not in cigar charger.
I have only used the OBD for power the past days, in the other plug have been in the camera it have not been powered.
I will be more careful later today when i redo my experiment, and make sure to note down what exact way of plugs plugged into the camera.
 
Okay i have become a little wiser today, and want to share that with you guys.
First of all the camera shutting down at strange times, that are probably due to the OBD having a low voltage cut off at 12.4 volts, i changed to 11.8 volts today and the problem have not been replicated today.
So in that regard i think it is safe to rule user error on my part :oops: and a almost 8 year old battery in my car.

The issue with the camera going into parking mode before i turn off the car, and then upon starting car again later the camera do not wake up and change to normal mode, at least not before a G-sensor event.
And i think it should wake up and change to normal with key on, cuz with the G-sensor set to low you can drive for a while before a event from the G-sensor will make the camera change to normal recording.
So i do think it is a major issue if some one arrive at a place, and talk a little in the car before turning it off, but what do you other guys feel on this ??

But i have good video with better focus of that today.


I do hope that the 720p in parking mode will get changed to 1080p, at least for time lapse modes, but even regular 30 FPS would be nice just at a lower bitrate.
When parked and the camera get e G-sensor event it seem to snap out of parking just fine and into regular video, giving you audio too if you have been in time lapse parking mode.
I have not yet studied if the transition back and forth between parking mode and G-sensor activated normal recording have some loss of time in between, but this i will take a look at a little later today.
 
Okay welcome back to the K2S party, tonight i took a closer look at the loss time as the camera transition in and out of parking mode while parked.

So i have come to this conclusions by looking at the footage:

1: There are a 1 second overlap between segments, this is a tad old fashioned IMO, hopefully it can be changed to 0 overlap in a future firmware.
2: There seem to be 2 seconds lost as the camera transition to parking mode from normal mode, and back out of parking mode to normal recording triggered by a G-sensor event ( me kicking my Left front wheel )
This i think are compounded by the 1 second overlap in general, so a total of 3 seconds lost, and i don't think thats too bad comparing to other cameras i have not yet had the pleasure of testing in the same way :sneaky:
3: The night time footage as you can see, well it could use some TLC for sure, but its all of the footage that need some tweaking, contrast - sharpness - 3D NR algorithm ( noise reduction ) and a dash of color tweaks too.

 
Time going into parking mode isn't concerning- after all you're still recording so nothing can be lost. Time going out of parking mode is important, and should be acheived rapidly since an incident can happen just a few seconds after a car is started then moved- losing anything there could be a problem.

Right now I think basic functionality should be the development aim. As we've seen with the DCTeam and Viofo's cams image quality can be sorted out quickly with the right approach and enough people. But without all the basic functionality working, image quality alone will not sell the cam or make for happy owners.

New cards are here, will test and get one in tonight. Going to look back at what everyone has been doing and see if I can find something different to focus my efforts on.

Phil
 
yes function first.
 
Okay a little weird seem to have happened while my car have been parked overnight with parking guard on, somehow the camera get triggered to record normal ( G-sensor / time lapse mode used )
But i must say using 12.2 V cut off the camera recorded from like 20:30 last night and until 8:30 this morning.
The few ( 4 ) normal recordings the camera have made during parking mode, i can not see any reason for the camera to have been triggered, but i will try to change G-sensor sensitivity to lowest to see if that make any change.
It was a little windy over night but nothing too bad compared to how it can at times be stormy in Denmark.

Regarding the old problem i have mentioned where the camera seem to loose track of time and so after a overnight parking the first segment was with the camera fallback time / date, well that seem to have been resolved, no such issue in the past 2 days. (y)
 
I just noticed the same for motion detect. The camera just continued recording in normal mode.
 
I will change to motion detect in a few days, while mu car are parked in a back yard, there are some traffic and so also things for motion to detect.

Aside for what i have mentioned above G-sensor time lapse seem to work just fine.
 
lacibaci said:
I just noticed the same for motion detect. The camera just continued recording in normal mode.
Click to expand...
More info: The camera seems to think that it is in Parking mode (the icon is yellow) but it just continues recording normally.
 
I have just looked over footage from last night, and 2 hours into the parking mode session the camera appear to have stopped recording for no reason.
Which go hand in hand with my surprise today as i got into the car and saw the battery voltage was still over 13 V, so it cant have been voltage cut off.
The memory card are full now, so i assume the camera have not been able to delete old tiles to make room for new ones.

Also the camera seem to have been in and out of parking mode again for no apparent reason.
As you can see in the picture i did 2 X 40 minutes of driving earlier in the day, so my battery was charged after yesterdays overnight parking mode test.

45666

Funny thing the creation of LOG files have been going on after recording stopped.

45667

The GPS log on the other hand seem to have stoipped at the same time as recording ( + 2 hour summer time )
45668


I will try to clear a little room on the memory card by deleting some files, and then repeat the test again today.
 
Last edited:
I have made a second confirmation that the camera some how stop recording if the card fill up while in parking mode, i can stop the camera and start it again in manual mode, and then it will keep recording like normal.
I have now formatted the memory card and have proceeded with G sensor / motion detect parking mode, but it will be a little while before the card fill up that way.
 
Okay.
So i have been using the G-sensor motion detection parking mode for a few days, and the memory card have filled up, but in this mode there seem to be no problems with recycling older files to make room for new ones.
Most of my parking files are from the back yard where my car are parked, and most of them i cant see what the hell triggered the recording as there are no buffer.
But as my car are facing the tunnel thru the buildings i think it is cars passing by out there on the street that trigger a motion detect event recording

45752
 
A compare screen grab between the K2S and my SG9663DC ( same sensor )

45753
45754
 
This is quite good for an IMX323 sensor. It's a tiny bit oversharpened for normal videos, but this helps you see the lettering/number plates more clearly so IMHO this part is near-perfect. I've seen a lot worse from this sensor and none much better- even the much acclaimed A119S with modified firmware is only very slightly ahead of this IMHO.

Phil
 
SawMaster
agree, oversharpeness sometimes help. I remember finevu 300/500 models with too sharpe but its better for plate. and here is good optimization firmware for imx323. But here iv talk about sensitivity in night shooting. Look at right side or lighting under the lantern. there is a clear advantage in sensitivity for imx291.
 
Yes i am not sure it is using IMX291, a few sensor models was talked about early in the game.
 
You must log in or register to reply here.
Back
Top