Thinking about how to optimize image selection

[Andreas Schleth]

Hi Robert,

do you really use the mouse at all to navigate through the images in
thumbnail mode?
If tagging with the letter tags in thumbnail view, I only use the
keyboard. Same with the viewer. One hand on the arrow/pg-up/down keys,
the other on the selected letter - ready to strike anytime.
In the time it would take to find my mouse (you mention some middle
button action) I could easily flip through several images in the viewer.
However, I could imagine a keyboard shortcut to "raise" a thumbnail to
full screen size as long as the button is held down. On my keyboard (DE
layout) the "^" would be the logical key (top left and somewhat
meaningful). Whether this would be faster than the two-pass method has
to be seen.

Another idea would be a shortcut to switch the mouse to "inspect mode"
beforehand (as with the new crtl-t to enter tag mode in the viewer), so
a simple mouse-down would suffice - maybe even to blow up the image to
1:1 px per px view with mouse pan ...

Cheers, Andreas


Am 04.12.23 um 00:13 schrieb Robert Krawitz:
> I seem to have somehow gotten dropped from the list, and a family health situation over the past 18
> months has greatly interfered with time to participate in OSS work.
>
> Anyway, I'm thinking about how to more efficiently select images for export/publication.  I shoot a
> fair bit of sports for my alma mater, and at games (basketball/football/soccer) I routinely shoot
> 1500-5000 photos, from which I have to select some (usually it works out to 10-15% or thereabouts).
>
> The current process I use is:
>
> 1) Unload the photos from memory card to disk (I have a bash script for this).  This takes some
> time, but is not labor intensive.
>
> 2) Load the photos into KPA -- again, takes time, but isn't labor intensive.  I could speed it up
> with faster storage.
>
> 3) Tag the photos with the correct gender, opponent, sport, and match.  Not particularly labor
> intensive, although if I shoot multiple games in one day (as I did yesterday), it takes a bit more
> work to find the boundary between the sets.  I guess I could do it with the timeline bar.
>
> 4) Scan through the photos, tagging the ones I want to keep.  I use single letter tags, but this is
> very labor intensive, because I have to look through every photo.
>
> 5) Symlink all of the photos for later processing.  Usually not a problem, unless I have multiple
> names with the same name due to using multiple cameras.  I'm solving that problem going forward by
> using a different name prefix for each camera, so I won't have any more collisions.
>
> 6) Crop and correct rotation on everything I've selected.  This is labor intensive.  I use
> rawtherapee with a couple of patches to keep the crop tool as the default between images (that saves
> me a keystroke per image -- not insignificant when I typically have 200-500 images!) and default the
> crop aspect ratio to free.
>
> 7) Upload everything.  This is more time intensive than labor intensive.  It would be even less
> labor intensive if I wrote a script to do the upload to SmugMug, but unfortunately that's not so
> easy with Facebook.
>
> So the big step I want to optimize is (4); all of the other kpa stuff I've already done over the
> past 5 years.  Yesterday, I saw another photographer at a game looking through his photos.  He was
> using large thumbnails on his laptop.  I suspect he was using some tool on Windows, but didn't get a
> chance to ask him (this was between games, and we were both busy prepping for the second game).
>
> What I think I'd like to be able to do is quickly scan through all of the photos using large
> thumbnails to mark them for future processing.  I envision this applying (or unapplying) a specified
> tag to the images I quick select.  Even 480 pixel thumbnails (which let me get 4 rows of 6 columns
> on my UHD monitor) won't let me check for critical sharpness, but it would be enough to let me
> quickly narrow down what I want to take a closer look at.
>
> The ordinary selection mechanism (clicking on an image) is not sufficient for this purpose.  I would
> need to use control-click; if I inadvertently just clicked on an image (or shift-clicked), I would
> lose all of my work.  One idea would be a checkbox in a corner of each image (probably at the
> bottom, where there's already space for image name etc).  Checking that box would apply the chosen
> tag; unchecking it would remove it.  If I've selected multiple images, clicking on one of the
> checkboxes would toggle all of them (first click would toggle all on, second would toggle all off,
> and so on).
>
> That leaves the matter of checking for critical sharpness and other effects.  The obvious way would
> be to first pass select them as above, and then look through the selected images with the viewer,
> but this also seems inefficient.  What I'd like is a quick way to inspect each image (as a full
> screen "tool tip", in essence) that would pop down by itself.  A simple hover isn't the right thing,
> as hovering on an image would then always just pop up a quick viewer.  One option might be middle
> button click-and-hold to pop it up; releasing the button would pop it down.  Of course, even 4K is
> low resolution by the standards of modern cameras (my Canon R7 is about 2x in each dimension), but
> if I really need to worry about that, I can go back and look at them later, but I normally don't.
>
> It wouldn't hurt performance to be able to link up the memory card download with the KPA load; it
> would reduce the I/O load, which with the 8TB spinning disk I use is hardly negligible.  But that's
> a separate problem.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kphotoalbum/attachments/20231204/d17c7b2e/attachment.htm>