It might be good to read also about the open options in https://gdal.org/drivers/raster/gpkg.html
"BAND_COUNT=1/2/3/4: Number of bands of the dataset exposed after opening. Some conversions will be done when possible and implemented, but this might fail in some cases, depending on the BAND_COUNT value and the number of bands of the tile. Defaults to 4 (which is the always safe value)." -Jukka Rahkonen- Even Rouault-2 wrote > On mercredi 2 décembre 2020 15:16:44 CET Nicolas Cadieux wrote: >> Hi, >> >> This follows a discussion in the qgis-user mailing list. We are using >> gdal_translate (and QGIS "export as") on a 8bit unsigned single band >> tiff to a .gpkg. When using no -ot switch or when using -ot Byte, this >> results in a 4 band .gpkg raster with the wrong values. (Values in band >> 1 to 3 are similar but off the mark). Band 4 is set at 255 as expected. >> Looks like some type of compression is going on in the band 1 to 3). >> >> When saving using the -ot Int16 or Float32 switch, all is well and the >> result has one band with the correct values. >> >> ex: gdal_translate -ot Int16 -of GPKG "E:/Users/Nicolas/Google Drive >> Nicolas/Partage_temporaire/qgis/int16.gpkg" "E:/Users/Nicolas/Google >> Drive Nicolas/Partage_temporaire/qgis/int16.gpkg" >> >> Is this a possible problem with the driver or is the problem sitting 18 >> inches in front of the screen? Problem similar on a Mac running QGIS >> 3.10.1-A with GDAL 3.0.2 and QGIS3.16 on Windows running GDAL 3.1.4. >> >> Test files are found here: >> https://drive.google.com/drive/folders/1fMl9odowUFFRQFK61rTxgHeom_D8ymN6?usp >> =sharing >> > > The answer is there: > https://gdal.org/drivers/raster/gpkg.html#tile-formats > > For Byte input (that isn't recognized as to be used by the tiled gridded > coverage data), the default tile format will use JPEG lossy compression > for > tiles with a fully opaque channel. > > -co TILE_FORMAT=PNG will use lossless compression > > GeoPackage raster aim was initially for visualisation purposes, hence this > default. The more analytic side was a later addition to the format, and > actually an extension > > Even > > -- > Spatialys - Geospatial professional services > http://www.spatialys.com > _______________________________________________ > gdal-dev mailing list > [email protected] > https://lists.osgeo.org/mailman/listinfo/gdal-dev -- Sent from: http://osgeo-org.1560.x6.nabble.com/GDAL-Dev-f3742093.html _______________________________________________ gdal-dev mailing list [email protected] https://lists.osgeo.org/mailman/listinfo/gdal-dev
