White Balance variations with different RAW converters

[george013]

I toke a NEF-file and converted it to JPG in different converters. Within them I set the colortemp to 6000. This is the result.

Photoshop Paint.



Capture NX


RawTherapee.


UfRaw or Gimp.


I did set the WB in those converters all on 6000, the "subcorrection", tint, on 0.

I wouldn't be surprised when I got differences when setting the WB by example on daylight, but setting on a specific temprature I wouldn't expect this.

George

[xpatUSA]

I toke a NEF-file and converted it to JPG in different converters. Within them I set the colortemp to 6000. This is the result.

~

I did set the WB in those converters all on 6000, the "subcorrection", tint, on 0.

I wouldn't be surprised when I got differences when setting the WB by example on daylight, but setting on a specific temprature I wouldn't expect this.

George

Excellent, George, thanks for taking the time.

Both UFraw and GIMP use Dave Coffins' DCraw as I'm sure you know. Which means that FastStone Viewer would give a similar result to them, probably.

[Shadowman]

I usually set my white balance to cloudy and just go about my shoot (RAW), the only time I might change it is if I'm shooting indoors or at night and don't want to look at the oddly colored LCD review. I'll change WB during processing if I think it's necessary, on a recent outdoor shoot my "As Is" white balances measured between 4600-6200, I changed most to either daylight or Cloudy within the Lightroom panel depending on how I wanted the final image to look, these choices gave me either a 5500 (daylight) or 6500 (Cloudy) temperature.

I have a friend who shoots with the same camera as you and she wanted to know which white balance to use when shooting with flash indoors and images are recorded in jpeg. I told her to either use AUTO WB or Flash, she stated that she didn't like the outcome as the images always came out with a yellow cast. My next suggestion was to use Daylight WB, waiting to see how it turns out. Before it's suggested, she is required to shoot jpeg only, setting is usually in school gymnasium. She uses one off camera flash and an umbrella.

[george013]

Ted,
Imageviewers use the embedded JPG. I know IView had a tool to convert the RAW but I can't find it back anymore. One photo took a very long time.

John,
Those presettings as daylight are global settings. Daylight on the equator is different from daylight on the north pole and daylight in the morning is different from the afternoon or evening.
What surprised me is the big difference when giving in a number, in this case 6000.

George

[NorthernFocus]

... What surprised me is the big difference when giving in a number, in this case 6000.

I ran into this a couple of years ago when I switched to LR as my RAW converter. Prior to that I had been using Nikon software as my converter. Most definitely the color at a given K value is different. In theory it shouldn't be but so it is. Recognizing the fact is important as is what to do with that knowledge. In my case I decided to pick one program as my RAW converter and stick with it. At least I can achieve consistent, predictable results that way. I do use multiple software packages for other aspects of PP but use LR exclusively for RAW conversion.

[Shadowman]

I ran into this a couple of years ago when I switched to LR as my RAW converter. Prior to that I had been using Nikon software as my converter. Most definitely the color at a given K value is different. In theory it shouldn't be but so it is. Recognizing the fact is important as is what to do with that knowledge. In my case I decided to pick one program as my RAW converter and stick with it. At least I can achieve consistent, predictable results that way. I do use multiple software packages for other aspects of PP but use LR exclusively for RAW conversion.

Dan,

Have you tried Lightrooms WB selector tool?

[NorthernFocus]

Dan,

Have you tried Lightrooms WB selector tool?

Yes I use it when appropriate. Though it seems to always add a lot of red tone.

[Manfred M]

No surprises here at all George. Different people / groups code different algorithms to do the raw conversion process.

You can see these types of differences in commercial products as well. Adobe's "Engine" was downright terrible until the re-write it for the 2012 update. DxO Optics Pro and Phase One Capture One are two other commerical raw converters that get excellent results and have a solid following.

The biggest concern I've heard about the freeware tools that you have used is that while they may be interesting from a technical standpoint, their output is far less nuanced than the better commercial ones. I haven't used any of the ones you did, George, other than View NX2, which I understand uses the same conversion engine as Capture NX.

[Geoff F]

Also bear in mind that some converters have their own default settings for various adjustment options. So setting just the colour temperature to the same value may look different depending on the default settings for brightness and black point etc in each converter. Sharpness can also be auto applied differently.

[george013]

I already asked ones more. Why gives a higher colortemperature a warmer, more red, color. Blue is hotter than red.
I still think that the colortemperature doesnt indicate a temperature but a correction. That would explain to me why a higer colortemperature gives more red, to much correction. Leaves the question: a correction from what.

I must think about it more.

George

[Mike Buckley]

No surprises here at all George. Different people / groups code different algorithms to do the raw conversion process.

No surprise to me either for the same reason.

[dje]

Also bear in mind that some converters have their own default settings for various adjustment options. So setting just the colour temperature to the same value may look different depending on the default settings for brightness and black point etc in each converter. Sharpness can also be auto applied differently.

Yes I think this has a fair bit to do with it. The colours in the image will be determined not just by WB temperature but also the colour calibration used for the camera by the software. It is probably not surprising that Nikon's Capture NX gives the best results in this comparison, given that it is a Nikon camera you are using. You can also see the effect of a tone curve being applied by Capture NX.

Dave

[xpatUSA]

I already asked ones more. Why gives a higher colortemperature a warmer, more red, color. Blue is hotter than red.
I still think that the colortemperature doesnt indicate a temperature but a correction. That would explain to me why a higer colortemperature gives more red, to much correction. Leaves the question: a correction from what.

I must think about it more.

George

Basically, the so-called color temperature is equivalent to the temperature of something like steel (actually a 'black body'). When you heat steel it goes opposite to what you would expect starting at red and going towards blue the hotter it gets.

You are quite right, it is indeed a correction, being used to describe a point on the well-known CIE horse-shoe diagram that is supposed look white - in spite of what the camera saw. The "tint" part allows a further correction toward green (+) or magenta (-). In fact, the standard CIE illuminants such as D55 lie slightly on the green side of the black-body curve.

You probably knew all that, I think.

[Manfred M]

I already asked ones more. Why gives a higher colortemperature a warmer, more red, color. Blue is hotter than red.
I still think that the colortemperature doesnt indicate a temperature but a correction. That would explain to me why a higer colortemperature gives more red, to much correction. Leaves the question: a correction from what.

I must think about it more.

George

George - from a physics standpoint you are 100% correct; when we look at a black body emitter, blue reflects a higher colour temperature (if I remember correctly)10000K to 20000K than the reds or yellows (1500K - 3000K).

Physiologically, humans have it reversed. Cooler black body radiator colours are viewed as having warmer tones than hotter black body radiators that emit blues and violet colours; which are viewed as colder tones.

Bottom line is human vision and physics don't always align.

[NorthernFocus]

I already asked ones more. Why gives a higher colortemperature a warmer, more red, color. Blue is hotter than red...

You have to keep in mind what is happening when you set a color temperature. You are essentially "correcting" white balance. In other words you are correcting the image to make it appear as it would have had it been lit by "neutral white light". So when you set the color temperature you are telling the software what color light to assume was shining on your subject. So when you set a high color temperature you are telling the software that blue light was shining on your subject. The software corrects my adding warmer tones (yellow/red) to the image. Therefore when you set a high (blue) number the image turns red. You aren't setting the color that you want the image to be, you're telling the software what the color of the light was when you took the shot so it corrects opposite. Make sense? It is counter intuitive.

[pnodrog]

George - from a physics standpoint you are 100% correct; when we look at a black body emitter, blue reflects a higher colour temperature (if I remember correctly)10000K to 20000K than the reds or yellows (1500K - 3000K).

Physiologically, humans have it reversed. Cooler black body radiator colours are viewed as having warmer tones than hotter black body radiators that emit blues and violet colours; which are viewed as colder tones.

Bottom line is human vision and physics don't always align.

Manfred I know where you are coming from but I don't think you can blame human vision. It is more a result of convention and terminology. I assume you have done a bit of welding etc and your vision enabled you to easily appreciate that a white hot plasma cutter was a little warmer than the dull red used when brazing.

Not unusual for terms and words to have different meanings in every day conversation compared to it's specific meaning when used in a specialist area such as law, medicine, science etc.

[george013]

Ted,
The horse shoe is describing colors on a light emitting screen. I don't think one can explain the color temperature and a correction with it.

Manfred,
I think it was that answer that didn't satisfy me. And still doesn't.

Dan,
That's what I wrote. But when I can choice for a correction of 6000K then I think that that 6000 means a mathematical approach and being equal for all converters. This is different as saying to correct for something as daylight. That's undefined to me.


In CaptureNx I can correct the WB in different 2 ways. With the use of color temperature and then selecting either a predefined temperature, sun, tungsten etc, or a specific temperature, 6000K. The other way is the gray-point. Select a point or area that should be gray and CaptureNx calculates the correction for the R and B channels. The G channel stays the same.

An example.
A studio with a light source with a color temperature of 3000K, nothing else. I set the WB in the camera on 3000K being the value of the light. Will there be a difference using a Nikon, Canon, Pentax or whatever other mark camera?

George

[dje]

Ted,
The horse shoe is describing colors on a light emitting screen. I don't think one can explain the color temperature and a correction with it.

Manfred,
I think it was that answer that didn't satisfy me. And still doesn't.

Dan,
That's what I wrote. But when I can choice for a correction of 6000K then I think that that 6000 means a mathematical approach and being equal for all converters. This is different as saying to correct for something as daylight. That's undefined to me.


In CaptureNx I can correct the WB in different 2 ways. With the use of color temperature and then selecting either a predefined temperature, sun, tungsten etc, or a specific temperature, 6000K. The other way is the gray-point. Select a point or area that should be gray and CaptureNx calculates the correction for the R and B channels. The G channel stays the same.

An example.
A studio with a light source with a color temperature of 3000K, nothing else. I set the WB in the camera on 3000K being the value of the light. Will there be a difference using a Nikon, Canon, Pentax or whatever other mark camera?

George

George the colour characteristics of a camera are fundamentally determined by the characteristics of the red blue and green filters in the Bayer mosaic in front of the sensor. Sensors will not all be the same in this regard. Hence each camera needs some sort of calibration information to be used. I'm not sure exactly how this is done in camera but for an external raw converter like dcRaw, multipliers are used for the red and blue channels. Software like ACR and LR use a more sophisticated approach to colour management using Hue and Saturation adjustments.

White balance correction is used to allow for the fact that for a reflective light scene, the spectral distribution of the light reaching the sensor will be different for different light sources (with different colour temperatures and correspondingly different spectral distributions). This will change the colour correction used as described in my first paragraph. The two go hand in hand.

All cameras and software will be different, there are no absolutes here.

Dave

[george013]

George the colour characteristics of a camera are fundamentally determined by the characteristics of the red blue and green filters in the Bayer mosaic in front of the sensor. Sensors will not all be the same in this regard. Hence each camera needs some sort of calibration information to be used. I'm not sure exactly how this is done in camera but for an external raw converter like dcRaw, multipliers are used for the red and blue channels. Software like ACR and LR use a more sophisticated approach to colour management using Hue and Saturation adjustments.

Somewhere any software needs a refence point. If you know that, than you know how AWB works.
ACR uses hue and saturation when you let the software determine what's good. I didn't do that. I did set the temp on a specific value and the others at 0 or 1.

White balance correction is used to allow for the fact that for a reflective light scene, the spectral distribution of the light reaching the sensor will be different for different light sources (with different colour temperatures and correspondingly different spectral distributions). This will change the colour correction used as described in my first paragraph. The two go hand in hand.

All cameras and software will be different, there are no absolutes here.

Dave

And should give equal results. I just Googled for color temp meters, Just an example http://www.bhphotovideo.com/c/buy/Co...ture+Meter/N/0.
I can't believe those meters are sold starting from 800 dollars and won't give a reproducable result on cameras.

George

[Mike Buckley]

In CaptureNx I can correct the WB in different 2 ways.

There are actually four ways. In addition to the two that you explained, you can change the values of the Red and Blue channels of the white balance manually until the image is appealing to you. Similarly, you can use the Color Balance tool to make the change (though I understand that technically there is a difference between white balance and color balance even though changing either one of them can produce the same results).

An example.
A studio with a light source with a color temperature of 3000K, nothing else. I set the WB in the camera on 3000K being the value of the light. Will there be a difference using a Nikon, Canon, Pentax or whatever other mark camera?

There will be a difference in color if you use two different lenses, so I would expect a difference using two different cameras.