[Update 2] WHITE BALANCE!

Tanckom

Active Member
Joined
May 1, 2016
Messages
179
Reaction score
61
Country
Germany
>>> PLEASE READ PREVIOUS POST <<<
>>> CLICK HERE <<<
So, the previous post has now 20 pictures, which takes hell long to load, and i get closer to it, so for those who read the previous post, keep continuing reading here.
white-balance-chart.png

+++++++++++++++++++++++++++++++++++++++++++++++++++++++
UPDATE 26.05.2016

--> after new trys, yes, my sleep was too short to be registered
Conclusion:
After another test, i found out, that only the three first values of

t ia2 -adj awb 0 0 0 0 0 0

are doing some changes.

FIRST VALUE:
Currently, i think the first value is some color value. However it changes between green/yellow/orange.
SECOND VALUE AND THIRD VALUE:
Maybe they are some levels correction? Because directly after using 5000 values, they change from yellow to clear images.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
UPDATE 27.05.2016

All test have been done in dark room with artifical lights.

Test 8:
Scaling the first value everytime +2000 up, and doing at the same number for the first value +2000 for the second value.
test1_by_bend96-da429v0.jpg


Since 4000 3000 looked best, i tried another round with lower differences:
test3_by_bend96-da42cng.jpg

And looks like the first value for this environment is good between 4000 and 4500.
Now looking at the best second value:
test4_by_bend96-da42d7a.jpg

I think that: t ia2 -adj awb 4250 3500 0 0 0 0 looks best. Info, my wall is not white, but bright-beige.


Test 8 Conclusion:
First value between 4000-5000 looks the clearest. All other value are either blueish, or green/orange yellow.
Second Value between 2500-5000 looks the clearest. All other value are either blueish, or green/orange yellow.



TEST 9:
Did several previous tests, where second value was 0, however didnt see any changes.
Setting second value to 4000 doesnt bring any changements too.
Here the results.
test2_by_bend96-da42bdy.jpg

Test 9 Conclusion:
Only the first two values seem to have an impact on white balance

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
USED SCRIPT, DOWNLOAD BELOW. (TAKES PICTURES AT BOOT)
 

Attachments

  • autoexec.zip
    806 bytes · Views: 58
Last edited:
OP
Tanckom

Tanckom

Active Member
Joined
May 1, 2016
Messages
179
Reaction score
61
Country
Germany
If anybody could help me with this, would appreciate it, because I'm slowly comming to a knowledge end ^^
p.s. Have only done some tests in dark room with artificial lights.
 

Tehnicni

New Member
Joined
Apr 15, 2019
Messages
20
Reaction score
2
Country
Slovenia
I have new YI HW version 23L FW 1.5.6.
"t ia2 -adj awb" work tottaly different
t ia2 -adj awb 0 0 0 0 0 0 = neutral
t ia2 -adj awb 1 0 0 0 0 0 = halve greenhalve green.jpg
t ia2 -adj awb 2 (or greater) 0 0 0 0 0 = totally greenfull green.jpg, 2, 4, 2000 or 8000, result is the same.
Al other number on any place have no effect
On pictures there is Notepad window on my monitor.

With 1.2.13 situation is the same.

I'm desperately looking for super wide angle camera for sunset Time Lapse and YI is OK solution except WB. With auto WB on startup, picture is desperately short of red.
@Tanckom which HW & FW you have and did you found solution.

2 samples of my sunset TL
 
Last edited:

meanie2

Member
Joined
Nov 27, 2017
Messages
212
Reaction score
20
Country
New Zealand
This Z23L flickers with a bluish tint, camera status lights are off:
DependableEveryAcornwoodpecker-size_restricted.gif

BadPoorIberianmidwifetoad-size_restricted.gif


On custom firmware from

And during playback of 2k resolution video file on i5 4th gen computer(using USB 3.0 card reader),the audio lags behind the video.
Using kmplayer or media player classic.
Wondering if others have such problem?
 
Last edited:

Timzed

New Member
Joined
Jun 19, 2020
Messages
3
Reaction score
0
Location
Ulm
Country
Germany
I tried these in video mode, and so far I noticed that the white balance varies when you move around.
So it seems that the first value is kind of a min and the second a max value, telling the camera the range to be used for auto white balance. This would also explain why there was no change when you kept on using the same number for the first and the second value.
So when it comes to video, I assume by using 2 identical values one can define a static white balance shift, e.g. daylight only, which would be great.
I need to play more with it.

I have the 25L with the latest 2020 firmware, and color-wise my camera shows the same results than in your screenshots.

EDIT: Just tested. Unfortunately the color still shifts around even when both values are equal, which means so far this is unuseable for video.
I found this filter, however I don't how how to use it:
t is2 -filter_sw 0x00004000
If someone could help here would be great.
 
Last edited:
Top