Apologies if this has been asked before. I tried searching, but couldn’t find anything that matched what I am about to ask.
I have an image that I would like to process, where there is a group of people standing in front of a background. Both the subject and the background are under exposed. I would like to increase the exposure for both the subject and background, but I’d like the subject exposed correctly and to leave the background a little under exposed.
In terms of workflow, processing speed and the quality of the final image, do I:
Increase the the exposure of the overall image to the point where I am happy with the background exposure, then mask off the subject and continue to increase the exposure of the subject until I am happy
or
Increase the exposure of the overall image to a point where I am happy with the exposure of the subject, invert mask the subject and decrease the exposure of the background until I am happy.
or
Do something else?
As you can see, I’m relatively new to this, but I have a bit of time to learn at the moment.
Thanks for the reply. Unfortunately because it is my family involved, I’d prefer not to post the original RAW. In it’s place, I have added a jpeg copy of the original RAW and one that I have applied global edits to. I have obscured the faces and badges for your safety and mine. (You might think I am hideous example of humanity and some things you can’t unsee! )
With this in mind, it would be best to treat it as a thought experiment.
The starting point is the “original” image.
The end point for the subjects, is the “edited” image, however, I’d like the background darker than it is in this image, but not as dark as it was in the original RAW. It’s so the subjects can stand out a little. This I know can be achieved by masking, and this is where I’d like to know how to organise my workflow.
This is actually the first image I have processed using anything, let alone darktable and I am reasonably happy with the results. Took me a while though, but practice and familiarity will get me there.
As this was my first processing attempt using darktable, I’m not sure if all modules I used were scene referred, but having seen many more YouTube tutorials since this effort, scene referred looks to be the way to go.
I was thinking Option 1 as well, but was wondering about other opinions or things I just don’t know about yet. I’m only at the beginning with darktable, so like they say, I don’t know what I don’t know yet.
If you are using two instances of exposure, foreground and background, you might want to try two instances of Colour Balance RGB, using the same masks, for tweaking. Or, if for instance, the background is how you want it, but you want to pay more detailed attention to your main subject, you could use just one instance, reusing the mask from the appropriate exposure module.
The ability to reuse masks in this way in different modules is one of the wonders of darktable. The only thing is the pipeline order: you have to define the original mask earlier in the pipeline for a later module to know about it.
I your example, I would probably start with the background. But it is not a one way street, as changing one will affect the perception of the other, so which ever one begins with, there will be toing and froing.
I find it is easier to think about the main subject:
I start by exposing for the subject until the mid-tones are good
I use tone equalizer if the contrast is huge
I use color balance rgb shadows/midtones/highlights for general contrast
And then I use masked instances of color balance rgb if I need to dodge and burn certain areas
So I’d choose your Option 2, darkening the background, either with tone equalizer background is already darker than the subject, or with masked color balance rgb. The advantage of color balance rgb is that you can also change saturation, contrast or color grading of the background.
Sometimes the hardest part is to manage the transitions when you do something like this…I find playing with the mask contrast and feathering…and by feathering I mean the slider not the outer border of the mask. Using that to make the mask follow edges or relax that a bit can help with blending the mask especially when there is a large difference in the foreground and background…
In your case here you could even blur the background a little and that might help…there are a few distracting elements there that don’t add anything to that photo…
I used your blur suggestion in my latest attempt and it mostly came out good. I only added a slight blur to the background because the heavier the blur the more unnatural it looked.
Starting to get off the original topic, so I’ll keep this brief…
I masked the subjects with a drawn mask, which was unnoticeable when it came to darkening the background, but when it came to blurring, there is a noticeable less blurred halo around the subjects. Good thing is though, it is only noticeable when zoomed in, not a problem when sharing the image by phone (or by social media, if I used it).
I did try a parametric mask on my first attempt a few weeks back, but there are some areas, where the difference between the subject and the background wasn’t that great. eg. hair and bark. (the small section included also shows the halo)
I’ve learnt some more about masking since my first attempt to when I next get a chance, I’ll have another go at using a parametric mask and if it continues to stump me, I’ll start a different post.
That good news…if you use the feather and enough it will usually blend in and clean up the mask…
This old blog post has a nice explanation of the edge aware blurring about 2/3 of the way through it and also a nice explanation of the retouch module…
)
