wrong color temperature Godox ? Do I do something wrong

[rlast]

Dear photography collegues,

Today I decided to test my color gels, by shooting a greycard for each color gel attached to my flash.
I started neutral - no color gel, and the result that came out surprised me:
7200 for a flash that is specified as 5500 +/- 200K
I checked this with the White Balance color picker in Capture one

Am I doing something wrong here? Any help is appreciated!

Rudolf

Equipment:
camera Sony RX10M3, RAW
flash Godox V860ii
Capture one V11.3.2

[Manfred M]

I can only guess what is happening because we don't know exactly how you are testing. Don't forget that the 5500K ±200K is what your flash outputs, not what is hitting your gray target.

With any light source, three things can impact the light hitting your target:

1. Direct light coming from your flash;

2. Indirect light from your flash, for instance if you are in a room, the colour of the furniture, walls, floors and ceilings that reflect the light from your flash; and

3. Ambient light - for instance sunlight coming in through a window or other light sources in the place where you are testing.

All of these can impact the colour temperature of the light hitting the card.



On top of that, how do you know your card is neutral? High end cards should have been tested during manufacture, but cheaper ones, especially as they get older can discolour as the inks and base material age.

Any filters and even the optical components in your camera lens or optical stack could in theory impact the colour temperature of the light hitting your camera's sensor.

Here is a screen shot of a gray card I photographed on Monday in a small studio and illumination came from two studio lights with soft boxes. Depending on where I read the colour temperature, I got variations from 5550K to 6000K. Look at the differences in the RGB values I show when I put a Color Sampler on different parts of the card.

[rlast]

Thank you Manfred, for your instant reply.
I understand your three points, and I tried to eliminate 2 & 3 as good as possible :

I tested before in my living room, so based on your feedback I went outside in the garden, and took a place avoiding ceilings and walls - and photographed my greygard attached to the fence @ 1.8m (so some distance to the last wall : the ground)

I checked light leaking from the sun (it is an overcast day today in the Netherlands), and took a test picture in M (1/1000th = my highest sync speed, and f/8) no flash. It is quite dark and gives a luminance reading of 4 (out of 255) in Capture One.
Not perfect, but not bad either ….


Now, took a new picture with flash on, set to 200mm (to bundle the light as much as possible forward).
Again : the auto white balance tells me, it is 7400K. Nothing changes.

Though - normally my pictures don't look too bad, so there isn't a big issue right now.
Still I would like to understand what's going on, so any further suggest (If I may ask) is appreciated!

Oh - my greycard: it brand is a Mannon. Not studied what kind of quality it is, it is only 2 months old. I guess it will not be anything special than "consumer" quality.

This is the result:


Thanks for reading / your help!

[Manfred M]

1/1000th second synch speed? What camera are you using to get a synch speed that is that high?

Look at your histogram as that definitely tells us what is going on. In an ideal situation you should be seeing a single vertical line at around 128 value with identical R, G and B values (half way between 0 and 255). There is a bit of "pollution" affecting your test and that shows in the histogram.

A "perfect" histogram would look like this:

[rlast]

Manfred, thanks for the feedback. Let me take some time to produce better results.

The camera has a leafshutter, that is why it can work with a sync speed of 1/1000. (Sony RX10m3, 1"sensor)
I noticed, that there is quite some light roll off in the edges (vignetting), and I photographed a small part of the fence (left down corner) Both lead to a wider than "needle" histogram. The camera also produces more noise than the FFs most of you own, which also leads to wider distribution.

As said let me work on better results first, because the wider distribution might hint to flaws in my setup, though it doesn't explain the widebalance being that far for flash.


Thank you very much, especially for your pictures, as they are very illustrative … :-)

[george013]

Manfred, thanks for the feedback. Let me take some time to produce better results.

The camera has a leafshutter, that is why it can work with a sync speed of 1/1000. (Sony RX10m3, 1"sensor)
I noticed, that there is quite some light roll off in the edges (vignetting), and I photographed a small part of the fence (left down corner) Both lead to a wider than "needle" histogram. The camera also produces more noise than the FFs most of you own, which also leads to wider distribution.

As said let me work on better results first, because the wider distribution might hint to flaws in my setup, though it doesn't explain the widebalance being that far for flash.


Thank you very much, especially for your pictures, as they are very illustrative … :-)

Take a picture with the lens cap on and no ambient light, a dark room. That might tell you something about the behaviour of the camera.

George

[rlast]

George, may I ask what I should look at? I don't know what you expect to see.
I've done night shots and complete dark frames and these never surprised me.

Thanks, Rudolf

[george013]

George, may I ask what I should look at? I don't know what you expect to see.
I've done night shots and complete dark frames and these never surprised me.

Thanks, Rudolf

Forget it. Something else happens. I thought to show something like the black and white cats.

George

[rlast]

Still experimenting ….
Found out that I needed to change the color response to linear --> that make the luminosy more even...

did some shoots in the dark, to avoid light leakage
now finding out to damp the reflections playing a role … :-o
still not there yet ...

will be continued ...

[rlast]

Manfred, thanks for the help / suggestions so far...
I can get a much smaller distribution now, showing the exposure is around 127



though no better results obtained so far …

What did I try

* get closer (distance) to the greycard, to reduce the relative impact of light bouncing from the walls
* created a hood (tube) which I put over the flash, to bundle the light so reduce reflections
* shot in the dark to reduce the impact of light other than the flash

Still when I click on auto white balance in C1, the reading is Always around 7200K …

Most important for me to know now, is to understand : is it my greycard, or flash that is so far off in color temperature? :-(
So still like to understand …
If it is the flash - strange, but I have my greycard to get the right white balance
If it is my greycard - then I can simply replace …

Any hint to go forward is more than appreciated! Sorry that I have not more equipment (studio) for more reliable testing)
Rudolf

[george013]

Manfred, thanks for the help / suggestions so far...
I can get a much smaller distribution now, showing the exposure is around 127



though no better results obtained so far …

What did I try

* get closer (distance) to the greycard, to reduce the relative impact of light bouncing from the walls
* created a hood (tube) which I put over the flash, to bundle the light so reduce reflections
* shot in the dark to reduce the impact of light other than the flash

Still when I click on auto white balance in C1, the reading is Always around 7200K …

Most important for me to know now, is to understand : is it my greycard, or flash that is so far off in color temperature? :-(
So still like to understand …
If it is the flash - strange, but I have my greycard to get the right white balance
If it is my greycard - then I can simply replace …

Any hint to go forward is more than appreciated! Sorry that I have not more equipment (studio) for more reliable testing)
Rudolf

I'm not sure what you try to investigate now: white balance or color temperature.

I opened a nef in 3 Nikon converters, in DxO and in RT. Nikon tells the color temp is 5601, DxO says 5400 and RT says 5367. I don't think it's hard science in photography.

I'm still puzzled with color temp. High color temp is white, low color temp is red. In photography it's the other way. More like a correction on the 'measured' color temp. That means there's something as the 'measured' temp.

George

[xpatUSA]

<>
Still when I click on auto white balance in C1, the reading is Always around 7200K …

Most important for me to know now, is to understand : is it my greycard, or flash that is so far off in color temperature? :-(
So still like to understand …
If it is the flash - strange, but I have my greycard to get the right white balance
If it is my greycard - then I can simply replace …

Any hint to go forward is more than appreciated! Sorry that I have not more equipment (studio) for more reliable testing)
Rudolf

If it's any help, my later Sigma camera manuals list flash WB as 7000K. In the absence of published information on your camera's WB or your flashes' CCT, it doesn't seem that your test is that far off.

On the other hand, the ever-consistent Sigma says this of their older EF-500 flash model:

"Flash

approx. 5850 K

Sigma EF-500 DG flash

pardon the messed up text formatting ...

You pays your money and takes your choice.

As a cross-check, Kodak R27 10x8" white cards are pretty cheap on eBay last time I looked.

A pity that your camera WB or your flash color temps are not specified anywhere ... ?

[rlast]

Thanks xpatUSA :
this is the material I'm using:

grey Card Manning - 17.8%, flat spectrum, though rather cheap one (12US$?)
Flash - Godox V860ii - 5500K +/- 200K specified
Camera Sony RX10m3 - WB set to 5500K

so my expection was that a picture taken of the greycard with this setting should be rather neutral.
But if I compare it with real grey (background of C1) it looks very blue-ish, so a correction to 7200-7500K is enough.

hope this clarifies the pecifications?
cheers, Rudolf

[Manfred M]

Manfred, thanks for the help / suggestions so far...
I can get a much smaller distribution now, showing the exposure is around 127



though no better results obtained so far …

What did I try

* get closer (distance) to the greycard, to reduce the relative impact of light bouncing from the walls
* created a hood (tube) which I put over the flash, to bundle the light so reduce reflections
* shot in the dark to reduce the impact of light other than the flash

Still when I click on auto white balance in C1, the reading is Always around 7200K …

Most important for me to know now, is to understand : is it my greycard, or flash that is so far off in color temperature? :-(
So still like to understand …
If it is the flash - strange, but I have my greycard to get the right white balance
If it is my greycard - then I can simply replace …

Any hint to go forward is more than appreciated! Sorry that I have not more equipment (studio) for more reliable testing)
Rudolf

I personally feel that you are overthinking this. Colour vision is a mix of physics, physiology and psychology. The lower the colour temperature, the higher the human vision system's sensitivity to changes in colour temperature. A difference between 3000K and 3300K is going to be far more noticeable than moving from 5000K to 5500K.

The other factor in play is that in most cases that while a good flash will have a consistent light output, but that may depend on power level or flash duration. Very few flash manufacturers publish these curves. Some of the ones I have seen do show variations in colour temperature of 800K.

Another issue is that there are at least two inputs (sometimes three) that affect the colour temperature and those are the result of the direct light hitting the subject and also the indirect light (flash bouncing from the walls, floor or ceiling in an indoor environment and grass, walls and other structures in an external environment. The third component is ambient light, which may or may not be a factor.

If you have a decent quality gray card, I suspect what you are seeing are a combination of what I have written about; the variability in flash performance plus the environment you are shooting in.

[xpatUSA]

Thanks xpatUSA :
this is the material I'm using:

grey Card Manning - 17.8%, flat spectrum, though rather cheap one (12US$?)
Flash - Godox V860ii - 5500K +/- 200K specified
Camera Sony RX10m3 - WB set to 5500K

Thanks for the info, Rudolph.

so my expectation was that a picture taken of the greycard with this setting should be rather neutral.

That is a reasonable expectation, I agree.

But if I compare it with real grey (background of C1) it looks very blue-ish, so a correction to 7200-7500K is enough

My understanding is that if you click on the gray card image with a color-balancing tool the displayed color temperature changes to 7200-7500K.

On your camera WB setting, is there also a tint control? Not that it should make that large of a difference, I would have thought.

[Manfred M]

so my expection was that a picture taken of the greycard with this setting should be rather neutral.
But if I compare it with real grey (background of C1) it looks very blue-ish, so a correction to 7200-7500K is enough.

If the colour does not look neutral, then perhaps your computer screen and viewing environment are the issue:

1. Is your computer screen sRGB or AdobeRGB compliant?

2. Have you calibrated and profiled your computer screen? Your computer screen should be set to operate somewhere between 70 and 120 cd / sq m and your working environment should be less than 70 lux.

3. Is your work environment set up in a way that your colour vision is not influenced by external factors like the colour of your work space?

I'm assuming your colour vision is "normal".

[rlast]

Thanks for the info, Rudolph.



That is a reasonable expectation, I agree.



My understanding is that if you click on the gray card image with a color-balancing tool the displayed color temperature changes to 7200-7500K.

On your camera WB setting, is there also a tint control? Not that it should make that large of a difference, I would have thought.

XpatUSA thanks, see my remarks:

answer 1 : yes, C1 = capture one, and the 7200-7500K is the color temperature I get, when I click on the grey card with the WB-colorpicker tool.
answer 2 : There is a tint control on my camera - didn't use it so far, left it in the middle.

[rlast]

If the colour does not look neutral, then perhaps your computer screen and viewing environment are the issue:

1. Is your computer screen sRGB or AdobeRGB compliant?

2. Have you calibrated and profiled your computer screen? Your computer screen should be set to operate somewhere between 70 and 120 cd / sq m and your working environment should be less than 70 lux.

3. Is your work environment set up in a way that your colour vision is not influenced by external factors like the colour of your work space?

I'm assuming your colour vision is "normal".

Manfred, thanks! my anwers to your questions:

1) my computer screen is sRGB, IPS screen, factory calibrated, 1/2 yr old. brightness set to 70%
I don't have equipment to redo color calibration myself, so I try not to draw conclusions ased upon what I see.
my vision is "normal"' ;-)
2) my working environment is between 0 and 35 lux. illuminated with 3x 7.5W LED, 6500K
3) No daylight in my computer room. I turn off the illumination of the room (LED, see above)
Normally I Judge my pictures with a 18% grey - proof margin.

My remark blue-ish was judged relative to the grey proofmargin, and before the BW correction to 7400K was made.

[Manfred M]

Your brightness is likely much to high. As a guess, I suspect a value closer to 30% would be where it should be. 7400K seems to be unreasonably high. I suspect that 5500K is a more reasonable value. If C1 is showing 7100K, I suspect that the colours in the image are fooling it.

[rlast]

Manfred, I might be overthinking this, understand your remark - but a big part of this exercise for me is learning!

your remark about sensitivity is valuable and relief, because I was thinking that the observed delta (5500K vs. 7400K) would apply to temperatures around 3000K as well what would make it a day and night difference. Thanks!


I will see if I can replace the greycard with another one, to exclude that from being the problem!
The other factors (indirect and ambient light) I tried to exclude as good as possible, but might also need updated equipment.

Let me re-iterate my Original goal :
Characterize the effect of colorfilters attached to my flash unit.
This to let the flash match various artifical lightsources during different occasions.


great help so far, thanks,
Rudolf