I have been fixating with the performance of DT since I first started to use it. Things have improved, I have been asking support also here in the forum, and I quite satisfied with everything… until the other day when I opened Lightroom on my old macbook pro and I was surprised to see how fast it felt when modifying exposure and tone curves etc; everything happens in real time, very fluid etc.
Consider this: I am working on the same 24Mp file, and the macbook is a 2014 model, core i5 with 8Gb of ram, internal intel gpu – compared to the much more modern Dell XPS with i7, 16 Gb of ram and Geforce gpu.
Needless to say, but I am still grateful for DT existence, happy with the overall system, and will never go back to Mac Os/Lightroom but there is maybe space for some improvements in terms of usage. It seems to me like most of the discussions here are very technical and few people comment on a more generic usage like fast culling of photos, quick edits that get propagated, etc.
For example I frequently use
darktable-generate-cache to make previews and be able to quickly scan through a series of images taken during a day, but that involves importing images in DT, taking note of the oldest photo-id of the batch, quitting, running the binary from the cmd line, opening DT again… if I were to convince a friend to switch over from LR I would have a hard time explaining that he needs all of these steps just to do a first culling!.
Anyway, first question is this: what’s the performance like for the rest of you; is a simple exposure variation lagging somehow or fluid?
What I’ve learnt to do is this (apart from calculating previews separately like I’ve just said); I set the simplest demosaicing algorithm (e.g. VNG for Fuji files instead of Markensteijin, or PPG instead of Amaze for Nikon files), deactivate all denoising modules, do the actual editing which is usually a combination of exposure, tone curves, crop, etc. After this, at the very end, I change demosaic algos and activate denoise.
This gives me a little more speed I feel, but what I discovered the other day is that by switching off my second monitor DT feels much faster! I haven’t been able to make an objective comparison but I wonder if also other people have found this to be true.
My standard setup is this: Dell XPS 15 plugged in a thunderbolt docking station Dell TB16, driving the main monitor, a Benq 27" 2560x1440px and an older Nec 23" EA231wmi 1920x1080. I should also point out that usually DT runs full screen on the Benq monitor, and the other one is to the side with maybe a chrome window opened on the online manual or something, so it’s not actually used by DT.