Opened 16 years ago
Closed 16 years ago
#2503 closed defect (invalid)
Output of tiled tiff causes weird artifact in one tile
Reported by: | morgon | Owned by: | warmerdam |
---|---|---|---|
Priority: | normal | Milestone: | |
Component: | default | Version: | svn-trunk |
Severity: | normal | Keywords: | |
Cc: |
Description
Running the following command on the attached tif: gdal_translate -co TILED=YES -co BLOCKYSIZE=256 -co BLOCKXSIZE=256 input.tif output.tif
Causes strange permutations of a certain tile on the right-most side of the image.
Problem is not just in gdal_translate; using the programming API to accomplish the same thing causes the same problem.
Tested with GDAL 1.5.2 and svn-trunk, all with both internal TIFF library and external TIFF library (version 3.8.2).
Attachments (2)
Change History (4)
by , 16 years ago
Attachment: | 001-DRG.tif added |
---|
by , 16 years ago
Attachment: | output_gdal1.5.2_internal.tif.gz added |
---|
Example output from gdal_translate, version 1.5.2, using internal tiff. Output is the same in other versions. Gzipped so I could upload it.
comment:1 by , 16 years ago
I've downloaded your input and output images, and both have the exact checksums (gdalinfo -checksum ) and they show up the same way, and fine, in OpenEV (I could only display the gray band, as OpenEv doesn't handle natively images made of gray + alpha band.). It is displayed correctly by QGIS and gimp too.
My guess is that your viewer cannot handle those images correctly, probably because of the not-so-common combination of a gray band + alpha band.
comment:2 by , 16 years ago
Resolution: | → invalid |
---|---|
Status: | new → closed |
Blah. I hadn't expected both ImageMagick's "display" and gqview to fail in the exact same way (especially as they seem to use different back-ends). Shows up fine in the gimp, like you said...my bad.
input image that causes bad output with TILED=YES BLOCKYSIZE=256 BLOCKXSIZE=256