pdvpjepj
(bio pppj)
October 11, 2022, 3:11am
1
I just got a Sony a7iv and it seems darktable 4.0 doesn’t have noise profiles for it. Is this correct or am I missing something? It seems it’s a very popular camera so I assume I am not alone.
If there are aren’t any, how difficult is it to create them?
1 Like
Rajkhand
(Raj)
October 11, 2022, 6:28am
2
Somehow this model is treated like an orphan. Even lens profiles are not like A7R iv. I am planning to buy this camera and was playing with some downloaded Raw files and have noticed this lack of support.
You can try here to find out what the reason is that there is no profile yet. Probably you want to join one of these.
opened 03:07PM - 18 Jun 22 UTC
scope: camera support
I uploaded a bunch of RAW-Images to raw.pixls.us to improve support for the came… ra.
Uploaded Fotos:
- Full Frame and APS-C
- in RAW, Lossless Compressed RAW and Compressed RAW
- in 3:2, 4:3, 16:9 and 1:1
opened 10:21PM - 12 Mar 22 UTC
scope: noise profile
no-issue-activity
[darktable-noiseprofile-20220312.tar.gz](https://github.com/darktable-org/darkta… ble/files/8238565/darktable-noiseprofile-20220312.tar.gz)
I hope this is usable :)
I ran into issues for ISO 102400 128000 160000 and 204800, I believe at such crazy high ISO levels, a full pinky veil appears on the darktable-interpreted raw files (jpegs from the camera don't have this), which means a total loss of any pure black on the picture or anything close.
The noise profile script thus complains that those pictures are not under-exposed enough.
Maybe the data produced for those profiles should be discarded, or maybe something is wrong in my methodology and I should re-do those. Don't hesitate to give me pointers :)
It was produced with darktable 3.9.0~git716.e00836f030-1, maybe I should have used a stable version? I can re-generate if you feel it's safer :)
Next I'll probably do wb extraction for this camera
darktable-org:develop
← koolfy:Fix-incorrect-white-levels-and-black-levels-on-high-iso
opened 02:14PM - 14 Mar 22 UTC
While trying to debug this: https://github.com/darktable-org/darktable/issues/11… 323
I realized black level values change starting from ISO-102400 until max ISO, applying 1024 prevents blowing out all channels and restores a proper image interpretation as far as I can tell.
While checking the "Black level" and "White level" fields extracted from exiftool on all my raw files, I also noticed metadata always mentions white level values of 15360 instead of the current 16283 value, so I added another commit to address this.
Feel free to discard that second commit if there was a good reason to use 16283 instead ;)
Peter
(Peter)
October 11, 2022, 8:40am
4
You can add it yourself in Lensfun. 20 seconds of work when you have localised where the file is.
mil-sony.xml is what you will need to open and change.
Edit: It is already in the database
1 Like
Peter
(Peter)
October 11, 2022, 8:46am
5
Possible to use the already created noise profile file from Github. Open, copy and paste into your own noise profile. I think the name is noiseprofiles.json
1 Like