CX350 White Balance Discussion

[Bassman2003]

Hello,

I am still getting adjusted to my CX350 and am trying to work it into my filming vs the PX270. Overall, it is doing well. The rear zoom controls have been greatly improved. Thank you Panasonic for sticking with it.

I am writing tonight regarding white balance. I have to say, I am not having the best experience. I will break this into two parts and I welcome your experiences & input:

1) Just getting a solid white balance is difficult. Tonight I was filming a football game that started in daylight and ended up under the sodium vapor lights. When under the lights I tried to get a white balance and it took me more than 10 tries to settle on something that worked. I have learned over the years to gauge off of the viewfinder as the numbers will show as close almost every time but the color will be drastically different. No way could you just click it and trust the numbers as in the old days. I know these lights are difficult but something just seems too difficult.

2) I have been trying out the variable WB setting so I can easily walk the WB down as the sun goes down. Which is a decent idea except for the variable WB seems to have a different scene file compared to a normal WB "A" or "B". In V WB the saturation is a lot higher. In direct sunlight it looks quite nice. When I get to under the lights it looks crazy over saturated. I switched to WB "A" and the saturation went back to a true representation.

Any idea what is going on here between the VWB and the set WB? Shouldn't the image be the same but with different a color temp?

Thanks for your input.

[NorBro]

Is it possible the variable WB also affects any color matrix values?

In other words - for the lack of a better description and also understanding of how your camera works - if your color matrix is 0,0 and you're in the regular manual WB mode and you change from 5600K to 3200K, your color temperature should only be affected (not color values).

But if you have a variable color temperature (again, not sure how it works in the CX350) the Kelvin change may also change the matrix from 0,0 to 10,-20 (random example) affecting your colors as well.

[Peter C.]

I used variable wb with the cx350 today. I didn't have the same problem because it was indoors under a consistent light source. I used it to warm up the scene a little more than what was metered wb off a gray card.

Barry Green, a Panasonic expert, doesn't recommend using variable wb because you're only changing one axis between cool and warm. The problem is color can change on another axis green to magenta and variable won't correct for that.

The situation you describe is difficult because it's mixed lighting that is changing during a live event that you need to continuously record. You might have better luck with auto white balance but if it's multi camera then that isn't ideal either.

[Bassman2003]

Thanks for your replies. I will do some tests in controlled lighting and see how it behaves. Norbro, I am not sure how the matrix is implemented. It could be. Maybe Barry can shed some light on this.

I wonder how the WB mechanism works on this camera anyway. Such wild variances in color under the same 300-400 kelvin range makes me think there is some non-traditional way to set color going on.

Here are three screenshots: (these are straight out of the camera. I fix the color in post btw. Yes, the middle one is green but VWB was falling down a bit and the game was going on!)

Daylight only - VWB 5400k
Daylight 5400k.jpg

Sodium Vapor lights only. Right before halftime (green...)
Lights 4500k VWB.jpg

Right after halftime with a set WB. This is pretty close to how it looked in real life. (this field had some weird color areas!)
Lights 4200k Set WB.jpg

[Bassman2003]

Follow up - What bothers me is the highlight handling. Look at the offensive line in shot #2 & #3. The highlight handling appears to be different between the VWB and the Set WB. This is the Achilles heel of this camera imho. The set WB highlights are not as controlled. Or is this a relative saturation thing?

[Peter C.]

Seeing the images, the color of the grass is most noticeable since it takes up a large portion of shot. No doubt the color of lights effect it a great deal. There is clipping in the highlights but sometimes there isn’t much you can do under hard lighting. Comparing results with broadcast sporting events that use far more expensive cameras you’ll always be left unsatisfied.

In my experience the profiles that are punchy and intended to be used straight out of the camera have more contrast and saturation. If you’ve color graded you’ll notice everything is interrelated. For example as contrast is increased saturation needs to be reduced. For a live sporting event color grading isn’t possible.

[NorBro]

Bassman, I've shot hundreds of football/lacrosse/field hockey/soccer games under these lights.

I know exactly what you're talking about and I always wrote it off as cameras' automatic or manual white card WB being confused by the combination of the lights and turf and where you are standing on the bleachers, press box or field.

So years ago I just started using manual Kelvin (no color matrix adjustments) and increased/decreased (it's preference) the temperature every 30 minutes if I started shooting a game around 4-5pm around this time of the year.

Once it's almost dark, you don't need to adjust the WB anymore. Pick one and stick with it so it's easier to correct an entire quarter or two at the end rather than multiple chunks (if you needed to).

[NorBro]

And if you'd like your saturation to match what you're seeing later at night then your daylight has to start off at around 6000K-7000K.

Sometimes 5400K is too cool for the sun setting and then matching nighttime visuals with artificial lighting.

[Barry_Green]

Sorry I can't do any testing to get to the bottom of things, I've got a broken spine so I can't do much of anything right now for the next couple of months.

Quick thing: what you're saying about the manual white balance acting like a scene file sounds very wrong to me, that shouldn't be happening. If you dial in a manual white balance to 5600k, and then switch to the 5600k preset, you shouldn't see any difference in color at all.* If you dial in to a manual white balance at (say) 5200k, and then you do an automatic white balance and it reports 5200k, then you can still see a notable color difference depending on how the automatic white balance chose to set the magenta/green correction.

(*assuming that the presets are using a neutral magenta/green setting, and the VAR white balance also uses a neutral setting, although that's one of the things that should be tested rather than assumed)

[Bassman2003]

Thanks for your replies. First off, Barry, sorry to hear about your back! Hope you get well soon!!!

What is frustrating to me is that I use a custom profile I made that is setup to maximize the highlight rolloff. What I am going after looks a little more on target using the variable WB. When I did a set WB it seems my highlights were more contrasty and clipping more. What has bothered me from the start of using this camera is the highlights and how they draw your attention. More like a consumer camera imho. But I have maxed out the menu settings to correct it.

I will do some testing under more controlled lighting as these field lights alternate between red and green and are a difficult environment for cameras. So from what you are saying Barry, the presets and the VWB should be identical but it is normal for the VWB and set WB to have differing color choices. What should not be happening (the reason for my post) is the scene file should remain constant through any WB setting right?

I will watch for this in my testing. Thanks.

[Barry_Green]

Well, yes, but with a potential misunderstanding.

The scene file should absolutely remain constant and untouched regardless of the WB setting. But, the wb setting will definitely affect the color, even as much as a scene file would, if it's set drastically different. But the choice of variable or preset should be identical IINM, when both are set to the same kelvin temperature.

The manual white balance can look quite a bit different from the preset or the variable white balance, depending on how much magenta/green correction the manual white balance does.

No white balance setting should have much visible effect on the highlights unless you're talking about a white balance setting that's so substantially boosting one of the chroma channels that it pushes that chroma channel into clipping.

[Bassman2003]

ok.

[Peter C.]

Sorry I can't do any testing to get to the bottom of things, I've got a broken spine so I can't do much of anything right now for the next couple of months.

I'm sorry to hear that. I have benefited from your knowledge in the past. I wish you a speedy and full recovery. We all take our health for granted until we lose it.

[Barry_Green]

Thanks for the kind wishes!

[Bassman2003]

So, during tonight's game I did some back and forth to see what is happening between the VWM and set WB. I will do controlled testing soon but what I noticed is the VWB seems to have more saturation as I stated earlier. When I had a set WB, I went into the scene file I bumped up the chroma saturation. After about 25 points the image looked very close to the VWB. Keeping that menu screen up, I changed to VWB and the image looked very saturated. So the scene file changes were affecting the VWB. I then lowered the chroma saturation back down 25 points and it looked like before.

This made me realize that I do a lot of my work on custom profiles with a Lowell Tota light under WB 3200k. So the look I settle on is related to the preset & VWB. That explains why I am always a little under enthused about the camera's image when I have used it. The scene file is not represented the same way when a manual WB is used. I would like to ask if anybody else with this camera could test this and see if their scene files look the same or different in switching between the different WB choices. Thanks.