Hey folks, I am experiencing a problem where loading files from my X-T5 and then applying Darktable’s default processing causes them appear something like half of a stop underexposed. I was wondering if anyone else has seen this and has suggestions before I simply set up a default style that bumps up my exposure.
Before you jump to your keyboard, yes, I know about the issues with Fuji dynamic range setting! My camera is set to DR100, but I am still experiencing the problem. This has always been the case with my processing workflow and I have noticed that the exposure module starts out at +0.7EV and I have to raise it to more like +1.2 EV to look like what I remember in camera.
Hi and welcome. the default setting of +0.7 is a rough approximation for all cameras. If you’re consistently needing +1 or 1.2 then just set the exposure to that and auto apply it. And it’ll be fixed. darktable is mean to be modified and bent to what you want, thus the out of the box isn’t always great.
I think that’s typical, my X-T2 also tends to underexpose a bit, but that’s good for RAW as highlights are not clipped, and shadow can be pushed easily (AI denoise is pretty good, pushing a stop or two won’t add more noise).
Watch out for the “black level correction”, in some versions it’s default to -0.0002, which causes weird noise/black issue in other modules. I have to create a new default auto-preset to set it to 0.
Thanks, glad it isn’t just me. I’m looking more into exposure defaults. I guess raw files are even more low level than I thought. I didn’t realize that even the brightness from them is sort of up to the camera’s built in tone curves.
I have a Canon G16 and I need to set the exposure slider to 0.0EV to match the cameras JPG. On the overhand I have a Canon R7 and need to set the exposure slider to +0.9 or +1.1 EV to match the cameras JPG. So as already said the 0.7 default is just a educated guess for the ‘average camera’. I have created styles for my cameras which include customised exposure module settings to match the camera. Good luck with your edits.
I wasn’t trying to say the JPG was the perfect aim point, but when I start my edits I find having the edit start at the same point as what I saw in the camera is a good starting point. The point I was making was that two different Canons had two different settings required for the default.
Also does setting the fuji or any camera to neutral have any effect on the RAW? I thought it wouldn’t. I thought that was just the JPG.
On a Fuji you use the flattest possible film simulation, Eternal or Classic Neg, set shadows and highlights to -2. That’ll get you a lot closer to linear (but not all the way there).
It doesn’t change the raw, but it does make the jpg more accurately reflect the colour and contrast captured in the RAW file. Which is good for giving you an idea of where your editing will start from.
This is a practice that I’ve been using, I set my X-T2 to Pro Neg. Std, it’s easier for culling since details in shadow are more visible (Geeqie groups JPG+RAF together). Another benefit is that videos are also flat, easier to adjust in post (contrast, saturation, …) I don’t use the SOOC JPG for anything other than culling (the X-T2’s embedded preview in RAF has low resolution).
However, with the X-T2 sensor being ISO-invariant, together with dual-gain, I’ve capped AutoISO at 800 to retain highlights, thus my SOOC JPG are often way too dark, defeating that practice Because of that, for indoor events, I usually have to import the RAFs into darktable, copy/paste exposure with “auto” mode to all RAFs, and export quickly with no additional setting just for culling. Then re-import the culled RAFs another round for actual edit.
I don’t think the X-T5 has dual-gain, but capping AutoISO is still a good idea to retain highlights in dark scenes since exposure (A+SS) is already maxed out.
As @paperdigits implied, you needing a different correction isn’t surprising. I use Sony, and I also need to add a bit more exposure (typically +1EV instead of 0.7). The default is a compromise, probably biased towards Nikon/Canon (simply because those two brands are/were the most common).
In-camera ETTR might not be as accurate as we think
Not to open a can of worms, but I remember the topic of ETTR got debated several times in the Fuji-X forum, basically using a flat film-sim to make the in-camera histogram (obviously based on JPG) to be as close to RAW histogram as possible. Yet, another reason I set my X-T2 to the flat Prog Neg. Std.
It isn’t, I just try and get as close as possible. Fuji’s seems quite bad in this regard, much harder than on my nikon. Metering was the reason i sold mu X-T3 and converted the X-T20 to IR. Anyway, this is why I said:
Hey Cedric, it is meant to set the camera to a recipe / setting that doesn’t put too much of an emphasis on color, contrast, etc. - i.e. a as “neutral” setting as achievable with the camera to have a histogram, i.e. OOC JPG view somehow near to the actually saved RAW. Cheers, Lars.
On a lot of camera systems, the jpeg rendition effects the metering, so setting a lower contrast jpeg file gets you closer to linear (though not quite there on any system I’ve seen) and thus closer to a “raw histogram”. On many systems leaving the default jpeg will leave a lot of unused headroom in your raw file.
Because of that, for indoor events, I usually have to import the RAFs into darktable, copy/paste exposure with “auto” mode to all RAFs, and export quickly with no additional setting just for culling. Then re-import the culled RAFs another round for actual edit.
