Updated EOS 2000D BlackAreas cameras.xml
This change solves https://github.com/darktable-org/rawspeed/issues/717 and is cropping out the first 4 columns from left optical black area when the first 4 columns are not black.
for RPU sample, exiftool says:
[MakerNotes] SensorWidth : 6096
[MakerNotes] SensorHeight : 4051
[MakerNotes] SensorLeftBorder : 88
[MakerNotes] SensorTopBorder : 47
[MakerNotes] SensorRightBorder : 6087
[MakerNotes] SensorBottomBorder : 4046
[MakerNotes] BlackMaskLeftBorder : 0
[MakerNotes] BlackMaskTopBorder : 0
[MakerNotes] BlackMaskRightBorder : 0
[MakerNotes] BlackMaskBottomBorder : 0
Is the output identical for that raw too?
https://github.com/darktable-org/rawspeed/issues/389 is the main culprit here, a few bad rows/cols shouldn't normally cause so dramatic effect.
Same. IMG_2637_CR2.txt
The sample from RPU says firmware 1.0.0. The one I got has 1.1.0
To me it seems like Canon didn't mask those 4 columns at all with this camera. Never seen that before with Canon.
Huh, that's bizarre. And the horizontal black area in cameras.xml also (already) does not match the makernotes...
Huh, that's bizarre. And the horizontal black area in
cameras.xmlalso (already) does not match the makernotes...
Or do you mean the part that is 0?
BlackMaskLeftBorder
BlackMaskTopBorder
With permission to share this raw file here for everyone, https://drive.google.com/file/d/1fAzOCo2dGLM6yJfyJB-yG34KyFTRmfgg/view?usp=drive_link
I asked the photographer about high ISO shot because the different parts of the optical black area are more visible.
The red part is the missing masked pixels.
I don't know what the blue part is but many raw files from Canon look like that. Is it how things work in sensor when the image first is captured? It begins to read the sensor from that part?
Purple part is the masked pixels we use to set the black level.
What about the green part? I see that part at higher ISO from different Canon models.