SG9665GC High Contrast/Dynamic Range Flaw and other Discouragements

joss002 said:
Ive found a happy medium with the angle of the lens but there is still quite a bit of sky bleaching and buildings fade into the sky,if I take it up 1 click the sky is fine but the road is pretty dark,down 1 click and the sky is a white out and most buildings fade into it .
SG9665GC v2
Firmware is SG20160125.V2 and Ive fitted the CPL filter.


And a couple of raw files if they are of any help !.

https://www.dropbox.com/s/b7gx7kgf8gdoene/2016_0316_140037_059.MOV?dl=0
https://www.dropbox.com/s/efg9380z8janrnr/2016_0316_140338_060.MOV?dl=0
Click to expand...

thanks, that's nowhere near what @Dashmellow has been experiencing, interesting about the effect the angle has and also that you have a largish black hood in the bottom of the screen, if anything it does seem this is more pronounced on darker vehicles, we may need to look at how that bottom section of the frame is tabled for AE, maybe a selectable option with a different AE table to suit dark bottom 20% of screen or not will sort that out, will be something we'll need to experiment with I think to get the right balance
 
By bonnet is dark blue and the dash is black so would make sense. It is mostly out of view with the camera aimed slightly higher but it would be good to be able to bring it back down again.
 
I just did a longer drive today ( 60 km or so in bright sunshine with a fjew puffs of white cloud here and there ) other than the overly bright sky i cant really find anything wrong.
Its not like mobile masts - wind turbines and trees with the overly bright sky as background get blurred/faded out, allso no overly bright road in front of the car.
Same spot the X cam handle the sky better but even it suffered a little under the bright lights of the sun today, i will update with a fjew screen grabs where the SG camera was at its worst, and then trow the same spot with the x cam in for comparing.
Will update this post with those pics.

Got sun head on driving south, and a little later sun 90 degrees to my left as i was driving west.
 
jokiin said:
thanks, that's nowhere near what @Dashmellow has been experiencing, interesting about the effect the angle has and also that you have a largish black hood in the bottom of the screen, if anything it does seem this is more pronounced on darker vehicles, we may need to look at how that bottom section of the frame is tabled for AE, maybe a selectable option with a different AE table to suit dark bottom 20% of screen or not will sort that out, will be something we'll need to experiment with I think to get the right balance
Click to expand...

The question of having a black or dark blue hood (bonnet to our UK friends:)) is an interesting one. It does seem to be playing a role here and this is exactly the sort of feedback I was hoping for in posting this thread. As I said in my initial post I was basically hoping for a constructive discussion rather than about merely complaining, even if that aspect is an inherent and unavoidable part of the mix in a thread like this.

So, I've been thinking about the question of aiming the camera so that there is dramatically less hood in the frame. I'm going to try again and see what happens but it raises a different question for me.

I run four cameras facing front, rear and on both sides. (SG and three Mobius cams) One of the great things about my four cam set-up is that I capture and overlap some coverage all around my vehicle. Each camera picks up a section of the extreme side coverage of the one next to it. I realized today after fiddling with the aim on my SG is that if I go too high to avoid my hood I lose a section of where the coverage overlaps with the side cameras and the SG's individual coverage of both sides of each front fender as well. There is a certain point where I would have to give up more than I'm willing to unless there is a firmware remedy for this problem. So, the multi-camera system I've built here was based on the traditional 60/40 or 65/35 sky road aiming recommendation we are all familiar with, but rethinking this concept creates a problem. At this point I'm not sure there is a suitable compromise that would get enough of the hood out of the FOV to fix the problem, assuming that could indeed fix the problem.

Edit: See THIS post for a rough demonstration of the overlapping camera coverage concept I'm describing here. (only three sides and not all the same cams mentioned above)
 
Last edited:
I would think possibly aim the camera up to see if losing all that dark section at the bottom means your exposure reacts like it normally should when transitioning is a good test, obviously not a permanent fix but if it shows improvement using this method then the bottom section of the AE table is probably what we should look at first

if that is the issue my initial thoughts are that we take the built in AE options, centre, spot and average and rewrite the values of each to suit what we need and then have user selectable menu options for which you select, renamed to suit their purpose obviously but it does look at face value like if you have a darker view in the bottom 3rd of the frame that the exposure results will differ compared to if this section of frame is lighter
 
arcticfire said:
:rolleyes:
Click to expand...

As I responded to you last time, previous experiments along those lines didn't have much of a positive effect even if I am willing to try it again. Certainly when someone like you suggests eliminating a large truck hood from the FOV entirely as you did it comes across as an illogical, impractical, idiotic solution. Suggesting doing this as a temporary experiment as @jokiin does where he qualifies his remarks comes across quite differently. We could do without the gratuitous eye-rolling emoticons here, friend.

Edit: Where I live, remounting or mounting an adhesive plate is essentially impossible during frigid weather but we are almost out of winter's grip.
 
Last edited:
jokiin said:
I would think possibly aim the camera up to see if losing all that dark section at the bottom means your exposure reacts like it normally should when transitioning is a good test, obviously not a permanent fix but if it shows improvement using this method then the bottom section of the AE table is probably what we should look at first

if that is the issue my initial thoughts are that we take the built in AE options, centre, spot and average and rewrite the values of each to suit what we need and then have user selectable menu options for which you select, renamed to suit their purpose obviously but it does look at face value like if you have a darker view in the bottom 3rd of the frame that the exposure results will differ compared to if this section of frame is lighter
Click to expand...

Interesting idea I guess. My truck has a 4 inch wide dark tinted strip across the top of the windshield, just out of view of the camera. If I tilt the camera up too far the strip ends up in the field of view. Perhaps I could remount the camera but that would make this a more involved experiment, of course. I'll consider it.
 
if you need some spare mounts or whatever for temp setup and experiments just let me know and I'll get some to you
 
Dashmellow said:
As I responded to you last time, previous experiments along those lines didn't have much of a positive effect even if I am willing to try it again. Certainly when someone like you suggests eliminating a large truck hood from the FOV entirely as you did it comes across as an illogical, impractical, idiotic solution. Suggesting doing this as a temporary experiment as @jokiin does where he qualifies his remarks comes across quite differently. We could do without the gratuitous eye-rolling emoticons here, friend.
Click to expand...

I would have thought it obvious that pointing it upwards was a test? It was a friendly icon, it's not rolling eyes on my phone? Chill out and stop being so insulting to people "like me" who are only trying to help.
 
Just to infuse some really posetive stuff to this thread, yesterday i did finally manage to remount my 9665GC and fit it with the noise dampning pad.

The pad on the mount do now make it a little more problematic to take the camera off the windscreen due to the snug fit, but its still more of a pleasure than before where i had to pull back in my headliner and at the same time push back and pull on the camera to wiggle it off the mount.

PS. That iconology/smily stuff i am not good at using them either, tho they figure a lot in my posts on the internet.
I see them as a result of my frustrations over my spelling / having to rephrase myself all the time to be able to write something, and me trying to be happy on the internet cuz IRL right now i am far from the happy chap i been the first 44 years of my life.

Just adding this one for good mesure :)
 
Black hood ! :eek: Its Kelp green :p lol but it is a dark colour.
Would a heated front screen affect anything on the cam?. At night you can see a slight star affect on street lights from the heating elements,the cam picks it up as well,as did my 0805 and 06.Just curious if it would affect anything during day light !
 
Okay back to the matter at hand, as i said yesterday was no big deal for the IQ in general.
But looking closer today i actually found 1 spot there the SG camera trew a fit of brightness rage, but to be fair so did the x cam, just not as much and the X semed to get over it faster.

So here is 14 frames grabbed fromm the footage.

first up the sun dident manage to completely kill off the cell mast on the right of the picture, and allso the 440Kv masts on the left is there too.
Its not perfect but they are till there at least.
sgsun1.jpg

xsun1.jpg


Next grab is a little down the road, there is some trees lining the road and footage, and the wind turbine on the left is allso there, and i am in the shade of a cloud, again not perfect but at least its still there.
sgsun2.jpg

xsun2.jpg


Next up is the forrest run, and this is where the SG cam trew a brightness fit, going into the shade of the trees it seem lke the SG camera ajust too much for that making the road go white too.
The X cam allso did the same but not to the same degree, it allso seem to ajust faster to the new conditions where i have left the shade of the trees.
It is as if the X cam can go strait from 1 setting to another where the SG camera seem to sample every setting in between before landing at the new setting for optimal use, and so the SG seem slower to implemt those changes.
sgsun3.jpg

xsun3.jpg


The following pics is the same place but about 40/20 frames later on in the footage ( x cam on 60 FPS / SG cam on 30 FPS )
sgsun4.jpg

xsun4.jpg


Next up is a fjew pics going in a west by northwest direction having the sun on my left side, this both cams handle much better, and they both trow a fit when they get in the shade of a truck.
Here the SH cam react less violent to beeing in the shade, i assume this is due to the overall more bright situation.
sgsun5.jpg

xsun5.jpg

sgsun6.jpg

xsun6.jpg

sgsun7.jpg

xsun7.jpg


I know the sky/ground ratio is not the same for both cameras, belive me i have been banging my head against the computer table mumbeling the word dumbass many times.:oops:
I will clean this up for the next run, but i dont think it will change much, but still worth a try.
 
Last edited:
Thank you @kamkar1

I think those screen shots are extremely helpful because they clearly demonstrate the issue at hand in a situation where there is no dark vehicle hood in the FOV, just a small section of dashboard. Regardless of how you've aimed the cameras, about 80% of what is in the image is sky and fairly bright road surface.
 
Last edited:
I dont have much hood to display :( i wish i had somthing hiding a V8, but thats still only in my dreams that happen.

Far more dash i have on display in my videos, but i do have a CPL for the x cam in the mail to me :cool: but i will leave that out for this comparing stuff.
Dont take eagle-eyed expert users to see we have a blocking issue on the X camera, it can get really bad in the most difficult situations, i do think its a matter of code/settings and not bitrate ( this is on a 22 mbit trial FW for the x cam )
 
While I can't say I've encountered any of the issues brought up by the OP, it is of concern to me because I live in Phoenix and we have strong sun.
If there are issues with blowout in reflections, that is an area I will encounter eventually, if not often, but I don't check my video that often.

So I have an interest in seeing this corrected as well.
 
Dashmellow said:
Thank you @kamkar1

I think those screen shots are extremely helpful because they clearly demonstrate the issue at hand in a situation where there is no dark vehicle hood in the FOV, just a small section of dashboard. Regardless of how you've aimed the cameras, about 80% of what is in the image is sky and fairly bright road surface.
Click to expand...

it's not to the extent you were getting though which was quite severe, was yours worse though because of the extra dark section in the lower part of the screen, at this stage I'm not sure but always looking to establish any patterns
 
You must log in or register to reply here.
Back
Top