Canon GX10 AWB issue.

[vstrglv]

There is a very strange Automatic White Balance behavior of GX10. If White balance (WB) is set to AWB and I turn on a cam, sometimes it sets Day light WB instead of AWB. There is no sun icon on display of course. I have tried this in LED illuminated room. If cam sets Day light instead of AWB an image has a yellow casts and it does not change for a long time. If cam off-on WB sets to AWB, but sometimes Day light again. The same issue takes place if I change WB in FUNC. For example, if WB is set to Day light (or another WB) and then is set to AWB sometimes WB goes to correct AWB, but sometimes it goes to Day light WB. It’s a random procedure. I have tested P and M mode and full manual mode. What is the problem?

[vstrglv]

Hi, Inapickle.

Have you seen my video?    https://youtu.be/Pc-ep5q3fNE

[Inapickle]

I have now.

Ok, so that demonstrates the 'issue', but what is the color temp of the LED light ?

---

@Inapickle wrote:

What is LED lamp you are testing with - is it an LED panel designed and calibrated for photograpy/videography specifically?

What is the color temperature of the LED lamp as specified by the manufacturer and as determined by the camcorder when setting a manual white balance ? The color temp will appear on the screen after setting the WB.

---

In your video the 'Tungsten' preset is closest to neutral so it must be around 3000K.

Actually I had a similar experience with an HF-G30 I owned before. The overhead and wall lighting in my home at the time was exclusively halogen bulbs (mostly PAR20; MR16 in the basement) with a color temperature around 2800K, some brands even lower IIRC. Not my choice, they came with the house. AWB on the HF-G30 could never achieve neutral balance under those lights. Didn't matter if I turned the camcorder off and on again and gave it time to adjust, there was always a warm cast. So if I was shooting (casual) video at home in the evening I would set the color temp to 2800K and leave it at that. The HF-G40 I have now copes with those lighting conditions way better. Granted we have progressively replaced the halogen PAR20's with dimmable LED's that are a tad cooler (around 3000K, I think) but even in those areas where we are still using halogens, AWB on the HF-G40 manages to achieve neutral balance without any problem.

[vstrglv]

Thank you again, Inapickle,

LED lamps are general type warm lights for home.  The color temp is about 3000K I guess. 

The problem is not because AWB sets WB not correctly, but sometimes AWB freezes at one color temp. I means it does not work at all at this case. I do not prefer to use AWB because WB changes during recording. But my wife prefer.

And it is very unpleasant to have this issue for new, not cheap cam.

And I hope to get replies from GX10 owners about this issue.

[Inapickle]

Technically AWB is not 'getting it right' in any of the shots in your video. In those instances where you said 'AWB works' there is still an orange cast, albeit not as strong. Maybe your wife prefers that because it comes closer to reproducing the actual ambience in the room under those 'warm' ('soft white') household lights, but in those shots it's only the tungsten preset that comes close to acheiving 'neutral balance'.

That said, it does state in the User Manual (p 66) that:

Using a custom (i.e. manual) white balance may provide better results (than AWB) in the following cases:-

- Changing lighting conditions

-Close-ups

-Subjects in a single color (sky, sea or forest)

- Under mercury lamps, certain types of fluorescent lights and LED lights

[vstrglv]

I know this, thank you. My wife prefer don't think about WB, simple turn on cam and start recording.

I don't know is this issue only in my cam or in others GX10. Is it an occasion to change camera?

[Inapickle]

---

@vstrglv wrote:

My wife prefer don't think about WB, simple turn on cam and start recording.

---

LOL. My wife is just the same. She does have a point though.

Actually, thinking about this afterwards, that test system (whilst appropriate for determining the color temp of the light source) is probably not the best way to evaluate the behavior of AWB. AWB systems, in essence, work by evaluating the scene and determining the most appropriate white (and/or nearest to neutral) reference(s) for correcting the color balance. Generally speaking they work best when the color temperature of the ambient light is in the 3000 - 7000K range and I think that's probably true of all digital cameras. If there is an over abundance of one color (or similar hues) and/or no reliable white/neutral reference in the scene, it can be easily 'fooled' resulting in an image that has a color cast.

When you manually white balance with a grey card (or in this case a sheet of white), you are telling the camcorder that what it 'sees' is white/neutral. And when you select a WB preset you are telling the camcorder what the color temperature of the scene is, so it can make the apprpropriate corrections for shifting to 'neutral' color temperature (which is essentially 'daylight balance'). Testing AWB in this way you are, in effect, presenting the WB system with a uniform 'color' (i.e. white tinted by light reflected off the paper), so it has to interpret that as best. The further away that 'tint' is from neutral the harder that gets and the AWB alogorithm could well incorprorate limits. That might well explain why AWB is getting 'stuck' with a warm cast in your tests. 

Really a better way to evaluate AWB under those conditions (short of judging whether it looks 'about right') would be to place a digital grey card (ideally) in the scene and bring the test clip into an editing/grading software that has an 'RGB picker' tool so you can assess how close to neutral the RGB values are on the card.      

[vstrglv]

If AWB is getting 'stuck' with a warm cast it means that AWB does not work at all. I understand that AWB is a not very precision tool for WB, but it should not be "stucking." 

I have tested  another scheme today. I have directed the cam outdoor through the window and set AWB. After that I returned indoor with LED light. It was yellow-red cast of course. After one minite nothing changed. It means that AWB is getting 'stuck'. I have tried this scheme several times - the same result. I guess my example of GX10 is defected. Unfortunately  there is no any reply from GX10 owners. By the way a month I tested Canon HF G50. AWB worked very good.

[Inapickle]

It's not getting 'stuck', as such, it just can't reconcile the low color temperature under those indoor LED lights. If you move from indoor to outdoor again, or switch off the LED lights and have just window light coming in, the AWB readjusts doesn't it ?

---

@vstrglv wrote:

By the way a month I tested Canon HF G50. AWB worked very good.

---

Was that under the exact same conditions ?  I have no hands-on experience with the HF-G50, but there was this strange AWB quirk that someone posted a while back:

https://community.usa.canon.com/t5/Camcorder-Discussion/HF-G50-strange-white-balance-issue/td-p/3097...

Not sure what was happening there, but manual WB was the remedy.

Anyhow, you are, understandably, keen to get feedback from other GX10 users to know if this is the normal behaviour. If you don't get any response here, maybe try posting on other forums  - DPReview, which has a camcorder 'Digital Video' section, and DVXUser and DVInfo, both of which have a Canon section.

[vstrglv]

Inapickle whote:

Was that under the exact same conditions ?  I have no hands-on experience with the HF-G50, but there was this strange AWB quirk that someone posted a while back:

---------------------------------------------------------------------------------------

About HF G50,  I have not done the same experiments because AWB worked good in different light conditions.

I have to test AWB outdoor again.

Thank you for your advice about cam forums.