Blackmagic Forum

updated to 7.3...

I think there's a white balance bug:
Shot with Titan Tubes @ 5600k.
BMPCC4K set to 5600k in-camera.
Hasselblad X1D set to 5600k.

5600k on P4k looks way too warm and I have to set to 4400k to get a balanced white.

Attatched Screengrabs:
P4k in Resolve @ 5600k

P4K in Resolve @ 4400k

Hasselblad in Phocus @ 5600k

[/attachment]

do you have the tint corrected?
or set the whitebalance with a graycard?

Morgan,

I think the 4400K image is much too cool for the Pocket to match the Hasselblad here, casually viewing on my iPhone. I would prefer to start with the 5600K on Pocket 4K then add more contrast if I were trying to match them. The Hasselblad appears a bit yellow anyway.

The angles are different too, so there could be a slight shift.


Sent from my iPhone using Tapatalk

How do you know, that the Hasselblad is right?

P4k-5600 is closer to the Hasselblad than P4k-4400
Embedded in the jpg-files are icc-display profiles, this does not make sense

Use images including gray cards
Use Monitor and software with color management

Frank2021 wrote:How do you know, that the Hasselblad is right?

P4k-5600 is closer to the Hasselblad than P4k-4400
Embedded in the jpg-files are icc-display profiles, this does not make sense

Use images including gray cards
Use Monitor and software with color management

I wasn't really assuming that the Hasselblad was "correct" more of just another point of reference for what a 5600k setting looked like. But I'll take that advice

They were all just screen shots from the given softwares.

Ryan Earl wrote:Morgan,

I think the 4400K image is much too cool for the Pocket to match the Hasselblad here, casually viewing on my iPhone. I would prefer to start with the 5600K on Pocket 4K then add more contrast if I were trying to match them. The Hasselblad appears a bit yellow anyway.

The angles are different too, so there could be a slight shift.


Sent from my iPhone using Tapatalk

Thanks!
Wasn't trying to match. The Hasselblad was just a reference for what 5600k was looking like from a different camera.

Mark Foster wrote:do you have the tint corrected?
or set the whitebalance with a graycard?

Tint is where it's always been on the P4k.
Maybe it's an issue with the new screen calibration too. Could be why the P4K didn't match what came into Resolve.

Still looking in Resolve... there's no way that 5600k in Camera Raw tab is a correct color temp.

https://www.dropbox.com/s/jgxfaoxea0q7a ... .braw?dl=0

Weirdly if I view in Blackmagic RAW Player... it actually appears as though the file is the correct temperature.
Perhaps it's Resolve 17's interpreting of the file?

Doublecheck your workflow as well as the lighting environment. This camera doesn't match itself which means either the lighing or the settings have changed. Both are indicators of user error.

Good Luck

Howard Roll wrote:Doublecheck your workflow as well as the lighting environment. This camera doesn't match itself which means either the lighing or the settings have changed. Both are indicators of user error.

Good Luck

WB.png

That's my problem.
The lighting and settings both did not change.
Lights sat on a stand. Camera sat on a tripod and literally had no one touch it. I triggered via bluetooth.

Which version of Resolve are you using?

Morgan Gold wrote:That's my problem.
The lighting and settings both did not change.

Obviously something did, post the original Braw of the JPG that was too warm at 5600.

Good Luck

Howard Roll wrote:

Morgan Gold wrote:That's my problem.
The lighting and settings both did not change.

Obviously something did, post the original Braw of the JPG that was too warm at 5600.

Good Luck

https://www.dropbox.com/s/1ef79k63kv1p8 ... .braw?dl=0

Still think something about Resolve's interpretation is wacky.
Thank you for looking!

I've placed the first and last frame alongside one another so you can see how the lighting changes within the shot. Could be the fixture, maybe somebody opened the blinds. I'm not seeing any weirdness or inconsistency other than 4400 is a little low.

Good Luck

Hey, Howard.

Thanks for looking.

So indeed something strange is going on in Resolve rather than the file. Both of your screenshots match, and look accurate when set to 5600k.

When I have it set to 5600k in Resolve it is warmer.

Any thoughts on what that could be?

Howard Roll wrote:I've placed the first and last frame alongside one another so you can see how the lighting changes within the shot. Could be the fixture, maybe somebody opened the blinds. I'm not seeing any weirdness or inconsistency other than 4400 is a little low.

Good Luck

drift.jpg

How is your color management set up?

Morgan Gold wrote:The lighting and settings both did not change.

I'm showing how and where they did. Do you inderstand what is presented here? The lighing shift happens at 16:53:13:10.

Morgan Gold wrote:Both of your screenshots match, and look accurate when set to 5600k.

There are three they don't match and they're set to 4400.

Howard Roll wrote:other than 4400 is a little low.

The camera seems to white balance a little low but there's no reference. If you can establish some consistency in the testing environment you may be able to draw some conclusions. The Resolve angle seems way off base.

Good Luck

Morgan Gold wrote:Weirdly if I view in Blackmagic RAW Player... it actually appears as though the file is the correct temperature.
Perhaps it's Resolve 17's interpreting of the file?

Once the sidecar is updated in Resolve then the Braw player will use those setting, not weird or unexpected. Here's an extreme example.



Good Luck

Morgan Gold wrote:Hey, Howard.

Thanks for looking.

So indeed something strange is going on in Resolve rather than the file. Both of your screenshots match, and look accurate when set to 5600k.

When I have it set to 5600k in Resolve it is warmer.

Any thoughts on what that could be?

I'm guessing the lighting in your setup is closer to 4400K than 5600K, since the Hasselblad also looked warmer. If you can recreate the setup and white balance off of a gray card with the Pocket 4K and another color meter, even the cine meter ii app on the iPhone, then post those measurements.

The Pentax 645Z matched surprisingly close to my Pocket 4K for example when setting the same white balance in the same lighting.

Ryan Earl wrote:
I'm guessing the lighting in your setup is closer to 4400K than 5600K, since the Hasselblad also looked warmer. If you can recreate the setup and white balance off of a gray card with the Pocket 4K and another color meter, even the cine meter ii app on the iPhone, then post those measurements.

The Pentax 645Z matched surprisingly close to my Pocket 4K for example when setting the same white balance in the same lighting.

Unless the Titan tubes are waaaaaay off.
I set each one by hand on the tube to 5600k.

Gonna do a more controlled test tonight!

According to Andrew Lock from the Youtube channel Gaffer and Gear the Titan tubes are very precise in CCTs.

But don’t forget that your environment might reflect and color the light. A wooden door (if hit by the lights) for example can lower the perceived CCT by several hundred Kelvin.

What does firmware 7.3.1 suppose to have improved?

New test (no color checker, just paper and me):
Titans set to 5600k
BMPCC4k set to 5600k

Camera monitor looks balanced, raw file (though log and probably isn't the best reference) looks balanced but in Resolve @ 5600k looks very warm. Just ball parking the paper to white in Resolve is somewhere around 4900k.







BRaw clip:
https://www.dropbox.com/s/vbx5o1zc6ft3t ... .braw?dl=0

Robert Niessner wrote:According to Andrew Lock from the Youtube channel Gaffer and Gear the Titan tubes are very precise in CCTs.

But don’t forget that your environment might reflect and color the light. A wooden door (if hit by the lights) for example can lower the perceived CCT by several hundred Kelvin.

Thanks for the thoughts.

In both shots the subject is in a neutral room. The original shots were in a 80% white room with a slightly off-white background that lent to a very desaturated sand color if I was to describe it.

New shots are 100% white room with the exception of a dark grey carpet.

Why do you suppose there is so much magenta in the shadows? What does that look like when normalized?

Good Luck

Morgan Gold wrote:Camera monitor looks balanced, raw file (though log and probably isn't the best reference) looks balanced but in Resolve @ 5600k looks very warm.

Were you able to do an auto white balance off of a gray card or the white piece of paper in the shot? With the P4K or another camera / meter?

Ryan Earl wrote:

Morgan Gold wrote:Camera monitor looks balanced, raw file (though log and probably isn't the best reference) looks balanced but in Resolve @ 5600k looks very warm.

Were you able to do an auto white balance off of a gray card or the white piece of paper in the shot? With the P4K or another camera / meter?

Apologies. I did not.
I can try another with auto-white balance.

I will say that I've never up to this point ever had this issue before with any of this combination of hardware/software. I'll definitely try though!