I can’t get my head around white balance in Darktable. As I understand it, the main point of the entire machinery is to get as close as possible to white wherever there’s white color in the picture. Am I right? (Stop here if I’m not.)
Usually, the best example of why we need to calibrate whites is what a person looks like under leafy trees: Their skin has a green tint. If I want to get rid of that tint, I need to add it to the original light source.
But what if my intention is to keep it that way?
My photos are mostly documentary. What I’m dealing here with at the moment is a couple of snowy photos from my holidays. These were taken in an open forest, in overcast and foggy weather conditions. Such as this SOOC one:
It was yesterday and I remember that the snow did not appear clear white under these conditions. It was kind of blueish, and that’s how my Lumix G9 set to AWBw took it.
Now, DT seems to understand that in some photos, and keeps the whites on the cold side automatically. But it also adds a bit of green tint to it, and I can tell that’s wrong.
So, what’s the correct approach to get the whites closer to reality instead of trying hard to make them actually white? What I do to mitigate this is that I turn off the color calibration module and go with as shot in the white balance module.
I have often had the same thoughts. E.g., people will post photos of scenes with colored lighting, asking how to make the subjects look “natural”. I’m sorry, I’m sure that in the real life scene, they actually had that bright green (or whatever) cast.
I thinks it’s an interesting question not to be solved as I would sum it as :
would you rather represent the scene as it was (not white due to color cast) or as you probably perceived it (maybe more on the white side) due to the psycho physiological phenomenon of “colour constancy”
Visual perception is by definition a subjective experience that can be affected by number of other things like by persistence of vision …
My opinion is that there is no real right answer
I there any harm in bypassing color calibration and using old white balance ? We got by with white balance for a long time so I’d do that if it showed me better results … nothing keep you from using again color calibration on more challenging situations anyway ?
Is there any fundamental benefit in using color calibration instead of white balance if you’re not using any new functionality it provide ? I’d be curious to have anybody shed an educated answer on this.
Getting whites back to white is a first order approach: let’s get as close as possible to the “real” colors. If you don’t want that, that’s a perfectly valid choice.
In such cases, I first try to figure out why there is colour in the whites.
In this case, it’s easy, the light is daylight with a high CCT. So to get that effect, lower the correction in the color calibration module, e.g. if you have daylight, lower the color temperature from what the module sets.
If the module gives me hue+chroma, lowering the chrome gives me some of the original cast back.
The way I understand it is that you pick a “white” or neutral spot and it corrects based on the color of that spot. The spot itself may not be white/neutral as whatever is lighting it may change its color. I think this only works for one source of light which is why sometimes you need multiple color calibration modules(with masks), or correct it using color balance rgb or some other module. I’m not an expert so please correct me on this.
Yes, if there is a color cast caused by the camera or an external light source that you could not influence, but whose hue you do not want.
No, if you want to keep the color tone of the scene. In a scene with the evening sun, it makes no sense to have a white shirt as white, because he is illuminated by the yellow light of the sun and should look yellow.
That’s why color calibration after white balance offers different possibilities to correct the color tone according to your own judgement.
Nevertheless, according to my experience, it is recommended to make the white balance and to restore the corresponding tone of the scene afterwards.
I see the white balance rather as a “neutral” starting position for further treatment of the colors. By neutral I don’t mean “as the scene looked”, but, in fact, first the shirt - even in the evening sun - should be white. Then you can use the hue and chroma sliders to restore the temperature/mood. Additionally, with the help of channel mixer and/or color balance module, one can then fine-tune the colors even more.
In your example, I’ll make the snow white first and then retroactively give it the color cast you saw on the scene.
This “play” with the personal color control of the scene has a decisive advantage - you pay much more attention to the color composition of the scene already when photographing and with subsequent processing, you learn a lot about the color impression.
What you describe is an often recommended workflow. The downside, for those that view it as such, is that you have to recreate from scratch what the scene looked like. Considering how easily fooled your brains/eyes are and how unreliable our memory is you’re likely to end up quite far from what the scene looked like. Now some don’t care and view photos as paintings and some want to retain some relation to the scene.
Now with all the transformations of viewing media, surround and how our brains and eyes adapt it’s pretty tricky to know how to maintain colour accuracy through all those changes.
In practical terms though it’s quite helpful to have an as shot whitebalance that looks right on the scene. Or wb values that otherwise allow you to derive some version of scene accuracy.
Well I’m not sure this will answer your question… DT can “white” balance using the legacy method or you can go the modern route using a CAT… this is now not so much white balancing as trying to match the scene in a way that on your monitor you perceive them the way it was in the scene…
But I think what you are expressing is the perceptual nature of color. Our eyes do a lot of correcting that’s not a mathematical “WB”
Snow will always be one of those things. Much like bounce light in a studio… But our brains tell us snow is white or nearly or whatever should look right and so we are happy but our camera just collects the light with no perceptual correction
Using a CAT… dt tries to account for this… A calibrated screen and correct d65 values can play into this as well…
When using the CAT I find for difficult images I would just use the global picker…this is pretty good at catching the right hue of any cast but often it is too strong and so the image is over corrected and starkly neutral…simply pulling back the chroma to reduce the correction until the image was pleasing I found to be the best strategy…
I increased the saturation in both cases with color balance module, so you can see the color differences better.
In any case, both look boring. The sky was covered with a veil of clouds, so there was no direct sun. Nevertheless, I perceived the colors of the snow in shadows very discreetly bluish.
I adjusted the hue and chroma in the Color Calibration module to mimic this mood to a degree that I felt better suited the scene as I perceived it:
The subtle difference in the color shading of the cold colors of the snow in bright area, the snow in the shadows and the color of the ice of the frozen pond. And this as a contrast to the subtle color shading of the warm colors of the branches, the tree trunks and the red leaves on the ground.
Although the scene is actually low in color, it was the subtle color variations between complementary color ranges that, in my opinion, made the scene visually interesting.
Of course, I only vaguely perceived this color mood on site, but it was enough to draw my attention.
It’s an interesting point. I prefer to have the color calibration module switched off by default, I start with WB set to “as shot” and I tweak it from there in the WB module until it ‘looks’ how I want it to.
However.
It’s my understanding that at least some of the tools in darktable work best and how they are designed, when when fed a technically neutral image. A neutral image also means that things like LUTs, and styles using color balance RGB for color grading for example, work more consistently from image to image.
I think that’s the main points that I know of - I choose to ignore those, but they are valid.
Just noticing your color harmony on the vectorscope. In your last video I found it really interesting around 54-55 min. When you shared the tip and hinted for your next video. At that point in the edit on my screen the green foliage in the foreground was a bit neon looking but that simple use of the triad harmony and the adjustment with the channel mixer left that foliage a much nicer and realistic looking hue to me anyway… definitely looking forward to the next video to see how you leverage that tool …
You have been given some good answers here and @s7habo makes very informative and excellent videos on using DT.
For my 2 cents worth. I find DT is the best program I have used to correct the WB of images taken under extreme lighting such as red lights in the nocturnal house of a zoo. I can recover color in a scene like this using the white balance module alone with the color calibration module disabled. I have not found any program that can match DT under these extreme conditions.
For more normal photographs I often find the camera has done a nice job with WB and using as shot works fine for me. This can be achieved with WB module alone or WB module set to camera reference and color calibration set to as shot in camera. For the life of me I don’t understand why using two modules to achieve as shot in camera is benefical. I would love some one to explain to me the benefit if there is.
Now if I am not happy with as shot in the camera that is where the choice of two different modules can help. Firstly the white balance module alone provides a drop down selection of lighting conditions such as daylight, cloudy, incandesent etc for most but not all cameras. This is often sufficient for my needs. The WB module also has kelvin temperature and tint sliders which are very intuitive to use if you are experienced with that concept. So WB module alone for many images in my experience can be fine on its own to set white balance.
However, color calibration (CC) module in my opinion is not so intuitive to use. Especially for an inexperienced user. However, I concede for some images it can be very advantageous. For instance, you can use masks and multiple instances of CC module to correct mixed lighting conditions. But another great feature is the spot color mapping feature. I often use this to white balance a picture based on either a neutral color region or skin tones within an image. Lets be fair, many images we only have skintone to judge white balance on. So having some known values for skintones I can bring the colors in close.
The chroma sliders are really interesting. If I set the chroma slider to zero in the spot color mapping optons it will correct the image for a neural ‘grey’ target value. But then if I am unhappy with the correction because it has been too strong I can lower the chroma slider in the top section of the module. I call this the output chroma slider. This same technique works well when using skintone as a target. Because not everyone has the same skintone adjusting the output chroma slider can let me increase or decrease the strength of the correction applied.
Because I grew up in a world of color film color balanced for daylight I often set an image’s white balance to daylight in the white balance and/or calibration module and see how I feel about that color before doing adjustments. This is especially the case when processing another photographer’s image such as in the PlayRaw category of this forum.
I’d like to know that as well. It was probably discussed even CC was introduced, but I forget. What was the motivation for splitting white balance into two modules? What can CC do that the single module can’t?
If the result using just whitebalance module is ok, then go with it. There’s no benefit in using additional tools just because they’re available if you’re already satisfied.
If you’re not satisfied with a simpler whitebalance approach, then it makes sense to use the additional capabilities of cc.
The benefit of cc is to have further options if the simple whitebalance isn‘t sufficient e.g. if you need to match a color target or want to handle different illumination situations etc.pp.
There’s no one size fits it all approach…
I have read this, but since photography is about the look if the look is right from the white balance module alone then I wonder what the advantage is. My current default is white balance camera reference and CC module set to as shot by camera. But am I really gaining anything from this? The deveopler AP is very good at telling us what is wrong with certain modules and I respect his incredible knowledge, but I always remember even him saying photography is about the look and if WB module alone gives the look I want then I wonder why I want to complicate the process by introducing the CC module. After all, other respected programs only offer the white balance module and are they so bad because of it? I say these comments with respect and not trying to be antagonistic.
The journey from capturing the scene to the displayed image or exported jpg is all about controlling the loss of information.
It starts with selecting an exposure where you need to decide if you want to ditch shadows or highlights and continues through the digital process.
It’s all about control - if you use limited tools then you might not be able to keep information you want to keep. So it’s worth to have the algorithms that aren’t limited if you need to squeeze out something the limited functions can’t provide.
That’s the difference between darktable and several other tool: dt gives you the option to use it; it doesn’t expect you to be satisfied with tool that’s just ok for a 80% case
But it doesn’t make sense to worry about the third decimal place when rounding to an integer in the end.
My Fuji camera tends to give strangely greenish renditions in daylight if I use WB+CC. With WB alone, everything looks normal, in the sense of “similar to the JPG”, and “similar to other raw developers”.
This is easy enough to counteract with a little bit of magenta color balance, or indeed a tweak to the CC white point, but it is annoying.
Then in general it doesn’t make sense to use cc for wb in your case.
Maybe it’s useful if you need to match a colorchecker.
(But then you‘re losing the Fuji color rendition )
But even then: Quote from APs introduction post:
I guess the whole question is : is 12 to 20% extra color precision worth a whole new complex module ?
The answer is yours
But remember the highest precision bonus (20%) was given in the worst lighting conditions, and that’s usually when you need it most. Make from that what you want.
)