M2 footage dash and action

[country_hick]

It would be an exceptionally rare crash where needing a license plate number from video would be beneficial. Yes, there could be a third rig that caused the events that created a crash but if you show you were not involved in the originating action you should still be legally safe. It would be great to be able to read all license plates even when both care are going 300mph in different direction at night in a blizzard.

Having 360p video that shows one car causing a crash is enough to prove who did what in that crash. Just becayse you only get horrible picture quality does not mean you did not capture what happened.

The weak spot of all dash cams is the media they use for recording. SD cards do wear out and malfunction. Hard drives crash. Floppy discs lose their magnetic film. Some dash cams are made of quality parts and materials. Other dash cams last 34 days and die just after the 30 day warranty expires. Quality components prices are why no $15 dash cam can be expected to do well or last very long.

My biggest complaint about what I can not seem to obtain is a dash cam that will instantly turn on and work at -40f or -40c which are the same temp. I have seen -20 many times and -40 at least once. If my car uses electronic components that work at those temperatures why not a dash cam?
(I do understand tropical zone thinking that at -10f nothing could work. But being from up north I know other electronic things do work in that cold.)

 

[TonyM]

Got a few recordings of the inside of my lens cap today with the M2 in my pocket Found a few hot pixels. Checked that VBR works, as the file size was a lot smaller too! So I managed to test a couple of things even when making a mistake

 

[jokiin]

Got a few recordings of the inside of my lens cap today with the M2 in my pocket Found a few hot pixels. Checked that VBR works, as the file size was a lot smaller too! So I managed to test a couple of things even when making a mistake

have noticed the hot pixels also, not sure if anyone has reported this yet, they'd need to be aware of it first of all to be able to implement either an auto fix or a calibration method

 

[kamkar]

haha i made a few pocket recordings a while back, but i totally forgot to take the time out to look for hot pixels.

 

[TonyM]

Night time test of the exposure priority options, at 1080p30 and 1080p60, each with 'bright light' and 'low light' priority settings.
I can't see any difference between the bright and low light priority. Perhaps this is because the camera is already operating at its low light limit?
There's certainly a noticeable difference between 30fps and 60fps in this instance.

 

[c4rc4m]

I don't know quite how you reach some of your conclusions I often see posted here but many DCT members like me who have had hands on experience with a wide range of different dash cams going back many years will tell you that a great many of them are NOT reliable at all as numerous individuals have learned to their dismay only after the fact when the vital footage they thought they had captured was not. Among many of the attributes that you seem to find superfluous but that many find indispensable is that the Mobius has proven itself time and time again to be one of the most rugged and dependable on the market and therein lies much of the attraction and word of mouth popularity of the brand.

Personally I've never had a big issue but then again in fairness I haven't had a huge number of different cams. I also don't tend to buy products at 1st release like many on here. I guess if you buy a lot of cams, buy cheap or buy cams not fully developed to their full potential at 1st release, then you probably meet more reliability issues.

And some of have realized that plate capture on even the best cameras are something of a roll of the dice, i am not bothered by it at all, but if one of my cameras could capture a plate when needed i would be thankful.
But as it is now i only expect to be able to capture a plate under normal road conditions in bright sunshine from the right angle, the coming months i dont have much hope for capturing any plates due to the standard for Danish weather this time of the year.

Hence the reason why I put image quality above all else. It doesn't matter how small or reliable your cam is if it doesn't capture the required information be that the plate or drivers face.

 

[c4rc4m]

It would be an exceptionally rare crash where needing a license plate number from video would be beneficial. Yes, there could be a third rig that caused the events that created a crash but if you show you were not involved in the originating action you should still be legally safe. It would be great to be able to read all license plates even when both care are going 300mph in different direction at night in a blizzard.

What about hit and run especially if side swiped?

There's certainly a noticeable difference between 30fps and 60fps in this instance.

That's inevitable as the shutter exposure is twice as long at 30fps compared to 60fps. Hence why they recommend you to use 30fps at night.

 

[country_hick]

What about hit and run especially if side swiped?

That is one of the very rare situations.

 

[TonyM]

I'm currently beta testing a new camera which has at times delivered night time image quality better than both my M2 and GC cameras. However it has also failed to record at least once in every half hour daily commute and I certainly wouldn't rely on it as my only dashcam. The M2, like the GC, has been completely reliable.

The M2 is however only a 1920x1080 device and there's only so much detail you can capture at 2 megapixels, no matter how sharp the lens or how capable the image processing.

 

[c4rc4m]

Not sure you can blame 2mp alone. However, it doesn't help.

 

[kamkar]

And the old V1 SG9665GC missed out, so good it will be updated soon so it dont miss out.

 

[TonyM]

I've run a series of tests to compare the various M2 metering options in the latest firmware, to see what works well for dashcam use, and what doesn't.

I ran three tests in the same day:
The 8am test has some clouds with side lighting.
The 1pm test has some clouds with back lighting.
The 10pm test is at night!

I took short video clips, allowing the exposure to adjust to each setting before grabbing a full-size frame. I took both 1080p30 and 1080p60 clips. There was absolutely no difference between 30fps and 60fps during the daytime, so I have just used the 60fps clips. At night there was a difference, so both sets are presented. WDR was turned off in all tests.

The sequence for the images is:
Top row = Bright light priority
Bottom row = Low light priority
CW / CS / MS / UW / LW = Centre-weighted / Centre-spot / Multi-spot / Upper-weighted / Lower-weighted

8am (final clip ruined by white van arriving )


1pm


10pm - 30fps


10pm - 60fps


My thoughts on the test results:
- Metering options seem to have almost no effect at night
- Upper-weighted and Lower-weighted are a bit too severe for variable lighting whilst driving
- Centre-spot looks good in this static test, but I'm not sure I would trust it whilst driving - further tests needed
- Low light priority offers good exposure for the road and cars ahead, but sky tends to wash out
- Bright light priority exposes the sky well, but can be too dark for the road and cars

My personal preference is centre-weighted, low light priority. Other M2 owners with different car setups may get different results or prefer a different priority. The good thing is that the image adjustment options are available to let you tailor the output to your liking.

 

[kamkar]

Low light priority offers good exposure for the road and cars ahead, but sky tends to wash out

I agree, was hoping WDR would be able to catch this when it get better.

 

[jokiin]

30fps looks better at night, none of the other settings seem to matter much, that's good in one respect I guess as it's just a matter of deciding which is preferred during the day and night will be pretty consistent by the looks of things

 

[TonyM]

@TonyM said: Low light priority offers good exposure for the road and cars ahead, but sky tends to wash out

I agree, was hoping WDR would be able to catch this when it get better.

At some stage I'll test some of these metering options with and without WDR, as the WDR is pretty good with highlights but not so good with shadow areas.

 

[TonyM]

30fps looks better at night, none of the other settings seem to matter much, that's good in one respect I guess as it's just a matter of deciding which is preferred during the day and night will be pretty consistent by the looks of things

Yes, it's useful in that regard, as the metering options are global - i.e. they apply to both mode 1 and mode 2. So you could choose 1080p60-WDR for daytime and 1080p30-noWDR for night recordings, and choose the metering type that suits your daytime preference

 

[TonyM]

Here's the full-size 8am test images with centre-weighted metering:

Bright light priority

Low light priority

 

[TonyM]

Here's the full-size 1pm test images with centre-weighted metering:

Bright light priority

Low light priority

 

[TonyM]

And here's the full-time night test results, showing 30fps and 60fps this time as the metering options made almost no difference. And for good measure, there's a comparison frame from the SG9665GC.

1080p30

1080p60

SG9665GC 1080p30WDR

 

[c4rc4m]

I think the most important thing at night is the sensor. Some sensors are just more sensitive to light.

Adjusting the metering will only adjust the exposure. If the sensor has poor night time sensitivity then it will always be noisy because it has to use a lot of electronic gain to amplify the signal received which results in noise. If Mobius have a noisy sensor, then the only option as I see it is going to be abandon the sensor and start again.

I know we're talking in a different league but DSLR's show the difference between sensor sensitivity quite markedly and how it's progressed over the years. eg. Not many years ago ISO 400 was grainy. Now many DSLR's will take pics / video at ISO 5,000 without a lot of grain. The best camera (admittedly at a huge price and full size sensor) now manages ISO 400,000 (!) and is quite good in quality at ISO 250,000 (Sony A7S):

From 1 min 13 secs onwards if you want to skip the scene setting shots - here I believe the camera is cycled between ISO 6,400 and 80,000. Notice the lack of grain even at 80,000!