Hasselblad X2D 100C support #463
Conversation
|
Note also that the 3FR has these tags (does unique ID matter?): but no idea what that matrix is calibrated for... I presume older Hasselblad models also use the Adobe one instead of the 3FR embedded one? (They also seem to use the Adobe unique ID instead of the 3FR one... And this doesn't seem to matter as canonical_id is not used by dt currently...) |
Codecov ReportPatch coverage:
Additional details and impacted files@@ Coverage Diff @@
## develop #463 +/- ##
===========================================
- Coverage 59.26% 59.19% -0.08%
===========================================
Files 231 231
Lines 13766 13786 +20
Branches 1917 1921 +4
===========================================
+ Hits 8159 8160 +1
- Misses 5473 5491 +18
- Partials 134 135 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
☔ View full report in Codecov by Sentry. |
|
I pulled an image from the X2d into lightroom and then exported it t a DNG file. Here is the BlackLevel info from the resulting DNG file: $ exiv2 -pt B0000050.dng| grep BlackLevel I'm not sure what to do with this, how/where to update the xml file so I just added it here in case it helps |
|
and here are the color matrix info: $ exiv2 -pt B0000050.dng| grep CalibrationIlluminant2 $ exiv2 -pt B0000050.dng| grep ColorMatrix2 from the same DNG file (see my comment above) |
So we can infer/guess that at least for that camera the camera-provided color matrix is D65, i guess. |
|
@LebedevRI Looks like we'll also need to bump up the accepted size a bit: https://discuss.pixls.us/t/looking-for-workflow-advice-with-3fr-files-from-a-hasselblad-x2d/36897/27 H6D we already had on RPU is 12000 x 8816, new X2D is 11904 x 8842 |
|
With darktable master + your PR: Opening the raw from https://nextcloud.phxicg.com/s/9eGHMdPc8Q6jTEa, I get: |
|
That is pretty weird as the RPU sample is uncompressed. @kkempter-getswift Does the camera support both uncompressed and compressed raw output? Please upload all possible outputs to RPU, as requested there. |
I haven't checked, but that sounds like #144. |
|
Hi everyone I could use some help. I am a DBA and less of a developer, I use git & friends but minimally and even then I am quite rusty. I was the one who uploaded the .3FR files I downloaded 4.2.1 source code from https://www.darktable.org/install/ and compiled on my local Fedora laptop. Now I want to grab the source code that would include this PR and re-test. Can someone help me with how to grab the right code? Thanks in advance |
But it should be uncompressed! And the size seems to corroborate that - 2*100MB... |
I'm not sure what you mean? rawspeed/src/librawspeed/decoders/ThreefrDecoder.cpp Lines 52 to 69 in a02d62e |
|
per this request: That is pretty weird as the RPU sample is uncompressed. @kkempter-getswift Does the camera support both uncompressed and compressed raw output? Please upload all possible outputs to RPU, as requested there. The camera allows me to select either RAW or "RAW + JPG" and then I can choose 14bit or 16bit there are no other options. The files I uploaded are 16bit I just uploaded a 14bit example file (B0000077.3FR) |
Ok, that would explain it then 😉 At least this 16b X2D sample is uncompressed so needs to be hooked up differently. |
There is no difference whatsoever between the 14-bit and 16-bit modes it would seem - still uncompressed, and data is 16-bit, with no gaps in the values. |
|
Also, the X1D II 50C samples on RPU are uncompressed, and I think supported? @LebedevRI Edit: Looks like there is also a compressed X1DM2-50c (.fff) sample, and the X1D II 50C (.3FR) uncompressed one. Only compressed mode was checked perhaps? So this report then makes more sense? Yep, out of e.g. these |
|
Also added (copied w/ crop tweak for Mk II as well) support for the X1D while we're at it since the sample is there. |
|
Please split this into 3+ sequential PR's - uncompressed, and then per camera. |
Done. This now depends on #466 |
| <Color x="0" y="1">GREEN</Color> | ||
| <Color x="1" y="1">BLUE</Color> | ||
| </CFA> | ||
| <Crop x="126" y="94" width="-118" height="0"/> |
There was a problem hiding this comment.
Same, crop still leaves a few garbage pixels in.
How do you come up with it? Have you tried the methodology i outlined in some previous PR?
There was a problem hiding this comment.
Sorry, I indeed had no time to go through the zoom & scroll procedure, this was from raw data PPM dump...
|
Thank you! |
|
I suspect you all have fixed this already, and I suspect I'm doing something wrong to get the latest code with the x2d fixes, however just in case it's helpful: I checked out the master branch and compiled I also tried via checking out the master branch then: In both cased above I get this when I open darktable, import my 3FR images and then double click on one of the images: |
|
@kkempter-getswift All the necessary changes have been merged now, so you can just pull the develop branch again, no need for individual PR checkout... Also, as mentioned, if possible, please upload the equivalent .FFF file to RPU please. |
|
Thanks, I dont see any .FFF files, when I connect the camera as a storage device I have a DCIM directory and in that directory a sub directory (this one is named 114HASBL) and within the sub directory I have only the .3FR files, here is a directory listing: file:///run/media/kevink/X2D-100C-CF/DCIM/114HASBL/B0000004.3FR If there is something else I can provide let me know, or if I am doing it wrong Thanks in advance |
I think you get .FFF only if you transfer from camera using Phocus. Please test transferring both a 16-bit and a 14-bit file - i.e. for the exact two 3FR files you uploaded to RPU (B0000001 and B0000077), create the matching FFF ones by transferring them in Phocus please. |
|
I just uploaded to RPU a file named Job_0002.fff it was created by the Hasselblad Phocus software when I imported B0000050.3FR I think I already uploaded B0000050.3FR but I will do so again to be sure |
|
No
|
|
I have uploaded the below set of files, each of these are more in line with the guidelines: B0000009.3FR and Job_0009.fff |
|
I have also uploaded the below set of files, each of these are more in line with the guidelines: B0000010.3FR and Job_0010.fff |
|
Just so we are at the ideal conditions, could you please upload 3 samples //with same content//:
|
|
I have uploaded the following 4 files: B0000079.3FR -> Job_0079.fff (16bit) B0000080.3FR -> Job_0080.fff (14bit) |
|
@kkempter-getswift thank you, verified the uploads! |
|
All I compiled the latest development branch, everything is working, Thank You!!! I am however seeing these output errors on the command line: |
|
Nice! |
Addresses darktable-org/darktable#14345
Stub only (using ADC 15.3), levels and crop need to be verified.
Depends on #466 for OOC 3FR uncompressed support.