Hi Even, I just notice that in the page http://www.gdal.org/gdalinfo.html
In the sample reported is write a bloc size of 512x16: >Band 1 Block=512x16 Type=Byte, ColorInterp=Gray Should it be a typo error ? A. 2015-08-23 11:40 GMT+02:00 Even Rouault <[email protected]>: > Le dimanche 23 août 2015 11:28:21, Andrea Peri a écrit : >> Hi Even, >> thx for reply. >> >> So actually a tiff born to be used with vrt datasource be better if >> you was with a tile=128x128 instead of tile=1024x1024 >> right ? > > I guess that should help in most cases, but almost due to the reduced tile > size than the matching between TIFF and VRT block size. For example, if the > VRT is read by an algorithm that processes line by line, a reduced tile size > will cause less data to be read. > Increasing GDAL_CACHEMAX can also be an alternative > >> >> A. >> >> 2015-08-23 10:51 GMT+02:00 Even Rouault <[email protected]>: >> > Le dimanche 23 août 2015 07:43:03, Andrea Peri a écrit : >> >> Hi, >> >> I have a tiff tiled with tile size 1024 x 1024. >> >> >> >> When I try to build a vrt-datasource on it using the gdalbuildvrt. >> >> I see the result vrt file report a >> >> block of 128x128 instead of 1024x1024. >> >> >> >> I dont know if this is correct by the gdalbuildvrt oint of view >> >> >> >> Also I like to know if is preferrably to correct manually the vrt >> >> result to report the right value 1024 >> >> or is also acceptable to maintain the 128 size reported for the right >> >> work of the gdal utilities in the vrt file. >> > >> > Andrea, >> > >> > the 128x128 block size is currently hardcoded in the VRT driver. We could >> > imagine extending the VRT syntax to have user defined block size. >> > >> > Even >> > >> > -- >> > Spatialys - Geospatial professional services >> > http://www.spatialys.com > > -- > Spatialys - Geospatial professional services > http://www.spatialys.com -- ----------------- Andrea Peri . . . . . . . . . qwerty àèìòù ----------------- _______________________________________________ gdal-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/gdal-dev
