M2 Discussion

Ah right. Of course.... I've asked the guy on youtube to do a car test soon, he agreed, would be interesting to see what comes out....
 
Yeah a race drone is not the best platform to demo / test a camera, a cinematic quadcopter no problem cuz its flown nice and easy.

All i know is if i was flying race quads i would just get a regular TV camera for the steering, i would not care for recording the runs, actually it confuse me why race quads want to do that when most other forms of competitive racing RC things don't have the same need.

I must assume its the millennial thing with the "video or it dident happen" attitude :rolleyes:
 
if you want to do the same then that's probably a great video, if you want to use it in a vehicle it's not going to tell you anything

I think one of the challenges the M2 has had is so much of the early feedback was based on similar test footage, bouncing around on a quad copter, then they were all surprised when the results when used as a dashcam were so poor

I disagree. If it is really video from the Foxeer 3, then what it shows is that in a detailed environment, there's very little motion blur even at high speeds. The grass remains very crisp most of the time with you able to see every blade of grass, despite the movement, and so I'd expect performance in a car on eg a tree lined route to be very good. Beyond that, yes overall quality needs to be assessed in a vehicle because a quad has the advantage of a gimbal and only being used in daytime. However, it is a good indicator that this camera is shaping up nicely to be worth consideration and incar testing when available.

As for the teardown video, given it's so blurry compared to the video I posted, it's quite possible there's a difference in the re-encoding and the bit rate used. For a start the teardown video only goes up to 1080P whereas the video I posted was 4K downscaled to 1440P.
 
I forgot that I had these month old test screen shots as they were stored on an external hard drive but having just stumbled upon them I thought you guys might like to see them.
This is a comparison between the new M2 and an original M1 (A lens). Both cameras are mounted side by side, touching each other on a small wooden board.
This is firmware. v.203. There have been a couple of incremental (unreleased) beta versions since that have improved the contrast issues but not the overall IQ issues.
You'll note the narrower FOV of the original M1 A lens and the superior image quality.
Both cameras are at the default settings across the board. WDR is ON on both cameras.

Mobius 1 "A" lens
M1_a.jpg

Mobius 2 standard lens currently shipped with the camera.
M2_a.jpg
 
Last edited:
I would have said the most obvious difference is the lack of contrast on the M2, most probably because the WDR setting is stronger. This can be seen most easily from the Fords grill which is very black in the top picture but much greyer with more shadow detail retrieved in the lower picture.

I'm unsure how WDR is arrived at on Mobius cameras, but technically speaking, it should probably be the result of 2 merged images, 1 slightly under exposed and 1 slightly over exposed image to produce a balanced image with more detail. Somehow, I think that maybe a few manufacturers are simply using brightness or contrast as a means of retrieving detail rather than the 2 shot method and thus true WDR, (unsure as to what has been used here but I would guess the latter method as most of the picture appears to lack contrast to my eyes).

Also of note, is that neither lens is perfect. Look at the white line bottom left. With both lenses there appears to be some barrel distortion as the white line bends inwards. It's slightly less marked on the M1, probably down to the narrower view. In the M2 image, the lamp post bottom right in the background also curves inward.
 
Neither camera uses two merged images for WDR, nor is it merely a brightness or contrast adjustment. It is algorithmic. The M2 originally experimented with dual exposure WDR (HDR) but the technique was abandoned.

Generally speaking, wide angle lenses of this nature will always produce some level of barrel distortion, especially close up like this.
 
Neither camera uses two merged images for WDR, nor is it merely a brightness or contrast adjustment. It is algorithmic. The M2 originally experimented with dual exposure WDR (HDR) but the technique was abandoned.

Generally speaking, wide angle lenses of this nature will always produce some level of barrel distortion, especially close up like this.

a. Vico use the 2 image method. That said in my opinion the Opia 2 disappoints at night. I have footage from both the Opia 2 and M1. The M1 is as good, arguably better in my opinion.

b. Depends on the quality of the lens, but yes, I'd imagine for dash cam lenses, cost restrictions probably mean obtaining optically near perfect lenses is cost prohibitive. That said, in my opinion, number plates will always be an issue whilst lens only resolve well / have low distortion towards the centre of the lens as passing cars are usually in the periphery.
 
So what's best for a rear facing dashcam? Mobius or Mobius 2?
 
That will be hard to do right now, but since its dark and i have a flashlight, and my belly are full of too much food ( at least for my level of energy output )
Then i will take a short drive and retrieve the M2 in the car :)

Thx for the FW update shoutout.
 
interesting that they have released this, @Dashmellow only recently mentioned it was still struggling in the beta versions

The last version I've had to work with was v.2.03.2. which was sent to me mid February which I have not had a good experience with at all in two different M2 units with two similar but different lenses. Have not had much contact with Peter since then. Have not had a reply to my last several emails reporting on the M2 and M1 A2 either. (since two weeks ago) I sometimes get the impression that my negative reports on the M2 and the M1-A2 have not been so well received because the response I got back was that the camera performs better than a GoPro. Not for me so far. :(

Hopefully many of the issues have been addressed but "slight tweaks to the WDR mode" doesn't sound like it would fix the results I've been getting.

Basically, neither camera produces clear sharp video with good acuity, almost as if the lenses are slightly out of focus, the color balance and contrast has been poor and there have been other issues. I've even tried to see if I could do a manual refocusing but the lenses are glued in so tight it is essentially impossible. The more the camera sees motion the worse the visual acuity gets.

v2.03.2 - 30fps - "Higher" data rate - H.264 - CBR - WDR-OFF - all other settings at defaults

M2_landscape.jpg

I was advised to try the camera with WDR-OFF but the results I consistently get are very dark and still not very sharp looking. I thought perhaps the lens had gone out of focus but it doesn't make sense the two different cameras would do this. (also poor color balance)

M2-wdr-off_a.jpg
 
Last edited:
I wonder if perhaps constructive feedback is being misunderstood and being taken as criticism

No, I don't think that's quite it. I do send examples and detailed explanations but my results are the results I've been getting such as the above screen shots. What more can be said about results like that?

The truth is that I am very puzzled as this point. Some people seem to be reporting very positive experiences with the M2 and others not so much. Once the "green distortion flash" was resolved with the initial M2 release I was very excited about the potential for this camera but every firmware release since then seems to make things worse. For me it all started to go downhill when the dual frame WDR (HDR) was abandoned.
 
Last edited:
Just been testing a little in the living room with low light, and that do not seem to have become better, still quite noisy.

I will have to side with DM on
For me it all started to go downhill when the dual frame WDR (HDR) was abandoned
I cant remember that far back for sure, but it is something like that.
M2 daytime footage compared to the V3 GC and the XS is like it have a Russian industrial filter on it, colors are so much better on the 2 SG cameras and the M2 is like its filmed thru a really dirty UV filter.
The colors do appear true but they are not as vibrant as on the SG cameras, almost like i am filming thru a window with some degree of tint on it.

Now the M2 is no way near as good in low light was it once was, its like it have lost 2-3 ISO steps, and if i don't use WDR then it is more or less just blackness.

I remember when i first got the M2 it was actually more light sensitive than my old 9665GC, that's far from the case now even factoring in i am on a V3 9665GC now.
I have been toying with strapping the M1 to the M2 in the windscreen to demo the difference between the 2, and i really think the M1 will perform much better in low light.

The recording i madw of my flashlights tonight using the GC i wanted to upload the M2 footage of the same, but in the M2 footage you cant even see the dlg building 100 M away though it have a lit street below.
At least on the GC footage you can see the outline of the building before i crank up my flashlights, on the M2 ( footage deleted in FW update ) the building are just darkness aside the a narrow band down low where the street light is.
 
Last edited:
Some quick initial testing with V2.08 shows some very nice improvement with overall color balance and contrast over previous V2.03. Unfortunately, so far there is still a distinct lack of critical sharpness and acuity when compared to the M1 camera.

I'll try to post some examples when time permits.

One interesting thing to note is that there are now designations for A - B and C lens type options for the M2.

M2_lens_options.jpg
 
Still a mess..

WDR - On, EP - Low-light, Metering - Centre weighted.


@Dashmellow Had taken a few seconds of static video with the 3 options but couldn't really make out much on the phone. (Got overwritten due to a lot more driving than anticipated).
 
Last edited:
Anyone recording with H.265?
I seem to get a lot of corrupted frames, not sure if because of camera or card.
 
Back
Top