Some of you probably remember that I bought 2 pieces of OA4 when it came out last autumn, and now I have noticed that they give different exposures from each other. It's something you don't notice so easily, but I've noticed it in my videos now that it's starting to get lighter outside and against the snow.

At first I thought it was because I had different firmware on them, but in this video I show the whole process and with the same firmware on both, there is still quite a big difference.

So DJI do you guys or anyone else have any advice/solutions?

Yeah, that is a bit strange as the different exposures are easily seen, especially when viewed in 4K.  Will be interesting to see what DJI says about this difference.  

DowntownRDB Posted at 2-25 07:28
Yeah, that is a bit strange as the different exposures are easily seen, especially when viewed in 4K.  Will be interesting to see what DJI says about this difference.

I was hoping they would had the same when I took off the front cover, then it would be an easy fix.
And I could only see it against the white snow.

Maybe my theory is completely ridiculous, but I'm guessing you have the 2 cameras next to each other as you film? In which case they cover different view angles. Their view does not exactly line up. And so the WB and exposure will be a little different. (see image). If you look at your video at 1:38, for example, the color swatches are shown in a different angle, which means the cameras are shooting from a different angle and so this can affect exposure and WB...?

Of course this doesn't make any sense, I guess, if the WB is set manually?



Have you tried switching cameras, so first, camera A is on the left and camera B is on the right. Then you switch them so camera B is on the left and camera A is on the right. If in both cases the exposure difference is the same (eg in both cases the left camera is darker, or casts a bluish color, etc), then it's due to placement. Otherwise, it's an issue with the individual camera - I'm guessing. But I might be completely off.

hope you iron out the issues,  to add how bout just place on a tripod film one, then the other

osmonauta Posted at 2-25 13:08
Maybe my theory is completely ridiculous, but I'm guessing you have the 2 cameras next to each other as you film? In which case they cover different view angles. Their view does not exactly line up. And so the WB and exposure will be a little different. (see image). If you look at your video at 1:38, for example, the color swatches are shown in a different angle, which means the cameras are shooting from a different angle and so this can affect exposure and WB...?

Of course this doesn't make any sense, I guess, if the WB is set manually?

They are both set to 5500K, and yes all settings are all the same. Even when I tried to check they have both the same ISO when using Auto so they should show the same exposure. Especially when I set both on Manual with ISO 100, 1/200s, and also in the dark with ISO 400 1/200s.

The reason I made this video is because I have seen difference in my ordinary videos before and was at first thinking it was because they had different firmware, so was hoping it that was the cause.

And no your theory is not ridiculous.

Fishycomics Posted at 2-25 13:28
hope you iron out the issues,  to add how bout just place on a tripod film one, then the other

The light do change fast this time of year, so I need to record at same time since the sun (that is still low on horizon) switch fast.
And I was just hoping it was the front lens that caused it.

johansenfoto Posted at 2-25 11:15
I was hoping they would had the same when I took off the front cover, then it would be an easy fix.
And I could only see it against the white snow.

So much for an easy fix.  Still waiting to see if DJI has any thoughts on this.  

DowntownRDB Posted at 2-25 14:21
So much for an easy fix.  Still waiting to see if DJI has any thoughts on this.

Yes they will probably answer in the weekdays.

Not a big issue, but it is an extra step for each video I use both cameras.

they answer troubles within a 48 hour period. and   5 sec clips  a min to set be done in 30 minutes

I'm just wondering whether these slight differences also occur with larger cameras but are not noticeable because most people don't buy two of the same cameras.
Maybe this is just an action cam problem and DJI needs to go the extra mile here.

Andreas Dueck Posted at 2-25 15:47
I'm just wondering whether these slight differences also occur with larger cameras but are not noticeable because most people don't buy two of the same cameras.
Maybe this is just an action cam problem and DJI needs to go the extra mile here.

There are sometimes differences between used DSLR cameras of the same version, I found that out after watching a video from a camera repairer in Finland. But an action camera has nothing that can be set to other than focus.

Hi. Thank you for your feedback. Let me forward this to the designated team to check. I will get back to this as soon as they provide an update. We appreciate your patience.

Hi. Thank you for patiently waiting. Based on the relevant team after checking the video, this is under the normal range. Can you confirm if both cameras have the same parameters set and have you already tried testing in different environments? Does this issue only happen in snow scenarios? If you want you can send us the original files to further check. Thank you.

DJI Tony Posted at 2-27 02:56
Hi. Thank you for patiently waiting. Based on the relevant team after checking the video, this is under the normal range. Can you confirm if both cameras have the same parameters set and have you already tried testing in different environments? Does this issue only happen in snow scenarios? If you want you can send us the original files to further check. Thank you.

Yes I have seen this happend before in two other ordinary videos.
And yes I have tripple checked both cameras when to be sure they were both set with same settings.

But I find it strange that it is only happen when recording in D-Log M and not with color mode.

johansenfoto Posted at 2-27 07:20
Yes I have seen this happend before in two other ordinary videos.
And yes I have tripple checked both cameras when to be sure they were both set with same settings.

I see. Will it be fine if you send us the original files? Thank you very much.

DJI Tony Posted at 2-28 02:06
I see. Will it be fine if you send us the original files? Thank you very much.

https://transfer.pcloud.com/down ... f7R1oUabuUGH8spBGcy

There you have it.
Both are shot with the protective front cover and with same settings.

If only it was the front cover giving the issue, then it would be easy to match, like I could get two new covers that match for free hehe

johansenfoto Posted at 2-28 02:54
https://transfer.pcloud.com/download.html?code=5ZhTxJ0Zof51OFHl64BZOUP3Z0AIFThWK7f7R1oUabuUGH8spBGcy

There you have it.

Thank you very much. I will get this forwarded and update you as soon as they respond. Thank you. for your patience.

DJI Tony Posted at 3-2 00:26
Thank you very much. I will get this forwarded and update you as soon as they respond. Thank you. for your patience.

No problem.

I had a writing error in my post, both are shot WITHOUT the protective front glass.

johansenfoto Posted at 3-3 12:39
No problem.

I had a writing error in my post, both are shot WITHOUT the protective front glass.

Thank you very much for the added information. I'll keep you posted.

johansenfoto Posted at 3-3 12:39
No problem.

I had a writing error in my post, both are shot WITHOUT the protective front glass.

Hi, johansenfoto. Thank you for patiently waiting. I sent you a DM, please kindly check your messages. Have a great day ahead.

DJI Tony Posted at 3-5 01:16
Hi, johansenfoto. Thank you for patiently waiting. I sent you a DM, please kindly check your messages. Have a great day ahead.

I will do, thank you

Here is a follow up video if others also want to see. Standing still and also use the official LUT

You may notice some color variations or one being a tad darker than the other. But this can only be noticable when you film the exact same thing with both cameras. How often is that going to happen? You usually point one camera in one direction and the other one in a different direction and now all of a sudden no one will know the difference.

Andreas' comment makes sense. After all, no 2 processors have exactly the same speed rating (as in microprocessors in computers). They all rated, for example 3.2GHz, but one might be 3.17 and the other might be 3.23 but both fall within the same threshold so they market them both as 3.2GHz (...whatever the threshold may be).

Btw I didn't mention, but this second video I also gave as a reply to the PM I got from DJI earlier today, but posted here so you guys could check if interested.

You are right Osmo, for most users this actually doesn't matter. But I found out this difference since I use both my cameras like when plowing snow where they are each recording in different directions most of times but I twice could see there were a big difference in the exposure/color.

Here is the first video I discovered it in, and it can easily be seen when I switch to the next scene (camera) how much darker it is (from 3 min mark)

It's hard to tell which camera is which. If you look at for example 8:55 vs 9:01 those are two different angles yet they look very similar in brightness and color. So no significant darkening can be seen. The question is whether the two different angles were provided by two different cameras or was it the same camera and you just repositioned it?

osmonauta Posted at 3-5 15:39
It's hard to tell which camera is which. If you look at for example 8:55 vs 9:01 those are two different angles yet they look very similar in brightness and color. So no significant darkening can be seen. The question is whether the two different angles were provided by two different cameras or was it the same camera and you just repositioned it?

They were shot at different cameras, and yes they do look like the same I did see now regarding the brightness.

That's interesting to see and read.
Did you get any concluding answer?

Supposing all cameras are calibrated and color is 99.8% matching (as it should), you should not see any difference between both or even 10 cameras...
But this means to have a very strong team to calibrate and match all.

Montfrooij Posted at 3-13 23:47
That's interesting to see and read.
Did you get any concluding answer?

Yup got yesterday that this should be normal/close enough.

johansenfoto Posted at 3-14 03:05
Yup got yesterday that this should be normal/close enough.

I'm not sure how these auto exposure algorithms work, but apparently it has margins which will make it have slightly different outcomes in a similar situation.

Montfrooij Posted at 3-14 03:10
I'm not sure how these auto exposure algorithms work, but apparently it has margins which will make it have slightly different outcomes in a similar situation.

Different brands or models have different exposure like ISO etc. but would think the same model should be the same. For me it is a big difference that I can see in my tests, but in ordinary shooting (without snow) they will look the same.

johansenfoto Posted at 3-14 03:38
Different brands or models have different exposure like ISO etc. but would think the same model should be the same. For me it is a big difference that I can see in my tests, but in ordinary shooting (without snow) they will look the same.

I'm quite sure there are situations where the AE will often be 'wrong' (snow) since it thinks 'this can't be right' and tries to make for a 'good exposure' based on averages it was told to aim for.
So in these cases minute differences in the information it gets, can probably cause bigger differences in the end result.
Long story short, yeah, I can see how that happens with a snowy situation.
Most of my cameras get that wrong anyway.

And for the scientific part of this story we should probably do some indoor testing with multiple camera systems of the same type (like 2x Pocket 2, 2x Nikon D750, 2x Action... etc. etc.) and see if they give different results with the exact same environments.
But I don't have 2 'the same' camera's, so that won't happen here.

Montfrooij Posted at 3-14 04:07
I'm quite sure there are situations where the AE will often be 'wrong' (snow) since it thinks 'this can't be right' and tries to make for a 'good exposure' based on averages it was told to aim for.
So in these cases minute differences in the information it gets, can probably cause bigger differences in the end result.
Long story short, yeah, I can see how that happens with a snowy situation.

Would think when they are shot side by side at same time they would produce the same image.
I only have 2 of OA4, not 2 of other cameras.
Will think they are probably not produced at same day/week at the factory.

If only those who have like 5 of the same camera try to check theirs against white.

johansenfoto Posted at 3-14 05:14
Would think when they are shot side by side at same time they would produce the same image.
I only have 2 of OA4, not 2 of other cameras.
Will think they are probably not produced at same day/week at the factory.

If the shot was 100% the same yes, but minute differences can cause changes. And with the white snow, these could be exaggerated.
And with the wide angle of view, there is a big chance there is something different in the view because they don't align exactly.

I'm curious what happens if you take 2 shots with 1 OA4 in a similar environment from a tripod, moving the OA4 slightly sideways as it was in your original test.
Maybe they are different  too (that would mean it's the algorithm, not the OA4 )

johansenfoto Posted at 3-14 05:14
Would think when they are shot side by side at same time they would produce the same image.
I only have 2 of OA4, not 2 of other cameras.
Will think they are probably not produced at same day/week at the factory.

As you can see in the video, there are differences in where the black part (trees or the ISO white balance thing) is. One camera has them slightly less in the center.
I'm quite sure that is why because the algorithm will try to do its best to base the exposure on certain parts of the frame.
And with the super wide angle, there is simply too much difference between the 2 bodies which will make the algorithm react differently.
This is exaggerated even more because of the white snow, so the average brightness (which is what the algorithm will aim for) will change a lot for certain areas of the frame. (and some areas are prioritized over others probably because it's likely the important part of the image is there)
If you really want to test this properly, you have to find something controllable (like a big curtain or something) that is one color and to a new test there.
So both OA4's get exactly the same brightness on their sensors.

Montfrooij Posted at 3-14 23:30
As you can see in the video, there are differences in where the black part (trees or the ISO white balance thing) is. One camera has them slightly less in the center.
I'm quite sure that is why because the algorithm will try to do its best to base the exposure on certain parts of the frame.
And with the super wide angle, there is simply too much difference between the 2 bodies which will make the algorithm react differently.

I had this in mind while taking theese tests, and that's why I also did them in manual on some of theese tests with same results where one is brighter than the other one.
The only thing I haven't tested is switched sides while recording them.

johansenfoto Posted at 3-15 04:46
I had this in mind while taking theese tests, and that's why I also did them in manual on some of theese tests with same results where one is brighter than the other one.
The only thing I haven't tested is switched sides while recording them.

If you really want to eliminate variables, I think a grey blanket would be (something that has only one brightness value and that fills the entire frame)
99,5% sure they will give equal results then.

Montfrooij Posted at 3-15 05:47
If you really want to eliminate variables, I think a grey blanket would be (something that has only one brightness value and that fills the entire frame)
99,5% sure they will give equal results then.

I doesn't have a grey blanket (that I can remember) but yeah if I find one I will try.

johansenfoto Posted at 3-15 15:18
I doesn't have a grey blanket (that I can remember) but yeah if I find one I will try.

Anything big that is not reflective and one color will do (just not very bright or very black, as the auto exposure will try to make a 50% exposed image, so that will be hard).
Maybe a brick wall on a cloudy day.