has been explained by gwbarn. I move it to the left to give myself more real estate for the modules used for editing. I also like the RGB parade to just see the range of brightness for each channel. I don’t look at the histogram much for editing.
5 Likes
thanks @kofa @gwbarn @Terry this is why I love following this forums 
I’m someone who always read new features, and documentation but I completely miss this ones.
4 Likes
To get back to the topic, my culling strategy for personal photos is:
- Reject the obviously bad ones. I may use darkroom for this , to do quick adjustments, for example tone equalizer or agx (or even slow adjustments like color calibration), which I will then copy to all similar shots.
- Rate the obviously good ones 2-4.
- rate the questionable ones 1. These images don’t impress me, but they may still be the best of a given subject, and I may want to keep them.
- Undecided images stay unrated.
- Pay special attention to the focus between images in bursts. Only one image per burst should be kept after the first pass.
- Repeat, this time probably noting and culling very similar photos that aren’t next to each other.
- At each pass, a one star photo loses when compared to an unrated photo, because one star means I already decided it’s nothing special.
I don’t use 5 yet. But you could use it and consider 1 to mean unrated, and 3-5 to mean good quality. But I think it’s more intuitive that one star be a bad rating.
1 Like
I use a somewhat simplified version of that using flags (red/yellow/green) in digikam.
- red = rejected
- yellow = not yet decided
- green = keep
- Set filter to show only unflagged images
- In a first round images with obvious flaws (overexposed, out of focus, etc.) are marked red, everything else yellow.
- Change filter to show yellow images
- For each yellow filter decide whether to keep (green) or reject (red). This can involve going a bit back and forth to compare, potentially using digikams lighttable to decide which shot of a pair is more in focus or the like. In the end of this, all images should either be red or green.
- Delete all red images.
With the filters enabled, it will automatically jump to the next image once the flag is assigned, so this is a pretty fast process, where I just old the ALT key and press 1/2/3 to assign the flags.
I also struggle sometimes with burst shots, though, when there are several images that are more or less identical. I think it’s best to be brutal and simply randomly chose one and delete the rest in that case. Realistically, you won’t miss any of the deleted ones, once they’re gone. However, I don’t always manage to do that in practice 
.
Already for quite some time, I’m playing with the idea of implementing a culling tool for this, which shoes me all pairs of photos and for each pair I simply have to decide which one I like more. With this, it could then do some smart ranking (e.g. using OpenSkill) which then helps in deciding which one(s) to keep. Unfortunately, I likely won’t have time to actually working on something like that anytime soon :(.
1 Like
My culling strategy is first in camera: I assign 3 stars to the pictures with potential (3 stars for no reason, I just picked that once as an accidental solution).
When I have to share things on location, that’s great because I can just look at those files and use them for direct sharing.
When I get home, in Capture One (sadly not FOSS, but mostly because I haven’t found the FOSS workflow pleasant enough for this), I go over all images using its grouping feature in the cull view, then I assign all good enough pictures from each group a rating. Often I don’t do much more than changing the rating of the pictures I sorted in camera.
All pictures without a rating can then be deleted, but I usually keep them for quite a while.
Wow. I envy you.
My culling strategy is roughly this:
- When I am on a longer trip (eg a 2 week holiday), I try to cull obvious rejects (eg a stranger walked into a landscape shot) from within the camera. I don’t do this for simple event shots.
- I try to copy & import all files to my computer immediately after. The point is that it should be in a film roll so that if I have 10 spare minutes to cull, I do it. I assign two stars to everything.
- I try to review with a 3x3 grid to get a feel for the story I want the pictures to tell. The story can be simple (“I was on holiday at … and I liked …”), but it helps me decide what to keep and what to throw away. I go through everything and try to keep around 30% (3 stars) and reject the rest (1 star).
- Only at this point do I start preliminary development, crop (maybe rotate) and exposure. I cull again after this.
- I try to narrow down to a 100 shots from a single theme, often less. Not every good photo needs to be kept if it does not fit. When I have a gazillion shots of a single theme, I eyeball them all, looking for minor differences in focus, composition, and bloopers, but if they are the same, I keep the last one.
I get fatigued mentally much more by culling than by developing, I find it boring.
3 Likes
indeed. I was thinking
…(big elephant in room ignoring,no I won’t call it AI to keep certain people happy 
)…
we could use a classifier to pre-select at least those photos that are our of focus or have motion. Maybe there is already something like that
I have been looking at:
if I only had unlimited time
When I do bursts, these are not the challenge. It is trivial to identify both, even at a small size.
What’s difficult is getting the facial expressions right. Not only do people blink (which happens quite fast, and again is quite easy to cull), but they have subtle changes in their faces that are quite fast. Catch the wrong moment, and you are telling the wrong story about an event. OTOH if you catch the right facial expressions, you are 90% there, never mind about motion blur or focus.
I usually go with 6–10/s burst rates. The camera & lens can do more but it would be redundant for people.
3 Likes
Culling group shots is the worst. That’s the one Machine Learning tool I’d wish for: auto detection of the photo with the most smiles and auto-face-replacement from the other photos to maximise the smile-to-face ratio 
.
3 Likes