Task #11847 (closed)
Opened 11 years ago
Closed 7 years ago
Bug: Canon Raw pixel type always uint16
| Reported by: | omero-qa | Owned by: | |
|---|---|---|---|
| Priority: | minor | Milestone: | Unscheduled |
| Component: | Bio-Formats | Version: | 4.4.9 |
| Keywords: | codec, DNGReader, external | Cc: | nico.offeddu@… |
| Resources: | n.a. | Referenced By: | n.a. |
| References: | n.a. | Remaining Time: | n.a. |
| Sprint: | n.a. |
Description (last modified by jamoore)
See https://www.openmicroscopy.org/qa2/qa/feedback/7778/
The pixel type is set to uint16 because the white balance correction can cause the pixel values to be greater than 255. In cases where there are no stored white balance values, or all values are <= 1.0, it should be safe to set the pixel type to uint8.
See also http://www.openmicroscopy.org/community/viewtopic.php?f=13&t=7368
Change History (8)
comment:1 Changed 11 years ago by jamoore
- Description modified (diff)
comment:2 Changed 10 years ago by rleigh
comment:3 Changed 10 years ago by mlinkert
- Component changed from from QA to Bio-Formats
- Milestone changed from Unscheduled to 5.1.1
Moving to 5.1.1 for triage.
comment:4 Changed 9 years ago by mlinkert
- Milestone changed from 5.1.1 to 5.1.3
comment:5 Changed 9 years ago by mlinkert
- Milestone changed from 5.1.3 to 5.x
comment:6 Changed 9 years ago by jamoore
- Milestone changed from 5.x to Unscheduled
comment:7 Changed 8 years ago by mlinkert
- Keywords codec DNGReader external added
- Owner mlinkert deleted
comment:8 Changed 7 years ago by mlinkert
- Resolution set to wontfix
- Status changed from new to closed
Going back to the original thread, the file in QA 7778 is reported to have been acquired with 14 bits per pixel, so the expected behavior is not that the pixel type would change from uint16 to uint8. The expectation seems to be that the maximum pixel value will be greater than 255, but I don't see any evidence for this being possible in looking at the file. The image displayed by Bio-Formats in ImageJ is consistent with what I see after converting to PPM/TIFF using dcraw, or viewing in Windows Photo Viewer. Without a contradicting screenshot from Canon's software, closing as wontfix.
Is the white balance set dynamically during acquisition? Could we have datasets where the white balance varies from image to image? This might result in a mixture of images with varying bit depths, which would make things a bit inconsistent when it comes to analysis, in which case maybe having 16-bit data throughout would be useful purely from a predictability and consistency point of view.