Technical Details:
- MacOS 15.4.0
- Macbook Pro (M4 Pro, 24GB RAM)
- Siril version 1.2.6 (build 12871)
- Canon R7
Problem Description:
I have been using Siril version 1.2.0-beta2 (build 7216) for over a year due to a previous Canon RAW decode issue detailed in this earlier post. Recently I got a newer laptop (specs above) and decided to give the current version of Siril a try.
The old issue (mirroring across the y axis) is gone, however Siril now has the wrong bayer pattern and/or offset for my camera (R7). I can tell that this is the issue because there is still color noise in the debayered images, however the objects in the image appear almost completely grayscale and what little color remains is completely wrong (i.e. reds showing up as greens).
Converting to .DNG format first does nothing to fix the issue.
I’m leaning towards this being an offset issue, because if you let Siril do its thing it claims to be using the GBRG pattern – but if you force Siril to use the GBRG pattern with offsets of x=y=0 the files are decoded correctly.
As with the prior issue I posted about, 1.2.0-beta2 (build 7216) handles the files with no errors whatsoever.
The debayer is done by LibRaw. If your file is new, this is probably why LibRaw doesn’t handle correctly your file.
The EOS R7 was released almost 3 years ago.
Canon R7 was added to libraw version 0.21
https://github.com/LibRaw/LibRaw/blob/master/Changelog.txt
LibRaw 0.21
== Camera format support ==
* Phase One/Leaf IIQ-S v2 support
* Canon CR3 filmrolls
* Canon CRM (movie) files
* Tiled bit-packed (and 16-bit unpacked) DNGs
* (non-standard) Deflate-compressed integer DNG files are allowed
== Camera support ==
* Canon EOS R3, R7 and R10
* Fujifilm X-H2S, X-T30 II
* OM System OM-1
* Leica M11
* Sony A7-IV (ILCE-7M4)
* DJI Mavic 3
* Nikon Z9: standard compression formats only
Without files, I can’t do anything.
Here is a 15 second exposure of M27:
2U9A9858.CR3 (46.7 MB)
Tried with default parameters, looks ok to me
That’s very strange, I’m not getting that result. when I send the exposures from that sequence through debayering in the “calibrate” tab they come out like this:
Could you show me a screenshot of your Siril preferences.
Hold on, I just tried having it debayer in the “Conversion” tab instead and that worked fine. The issue appears to be specific to the “Calibrate” tab.
Here’s that screenshot:
Well, I cannot test the calibration with no calibration files.
Here’s the bias image I am using. The individual bias frames were decoded with the preferences exactly as shown in the screenshot from above.
I have to send it compressed as this website will not let me upload anything with a .fit file extension.
R7ISO12800Bias_stacked.fit.zip (80.4 MB)
Don’t you use flats or darks ?
I usually use flats, but I didn’t get to collect any on the night I took those images.
I almost never use darks. Modern CMOS sensors have dark current suppression systems that are surprisingly effective.
Again, calibration gives me what I expected:
I even tried with 1.2 series.
Here’s an image I took with 3 hours of exposure and no dark frames:
2 Likes
That’s very weird. I guess it must be a problem with my machine.
