in short, warp will give you exactly the target extent of a (possibly) new grid specification in that -te subdivided by the -ts.
translate will give you approximately the extent provided to projwin, but snapped to the source grid alignment at a particular zoom implied by the projwin subdivided by the outsize. Further, translate can do scaling but warp cannot, which I guess is why you have a vrt in the workflow - you can scale a source virtually with translate to vrt, then warp (or translate) that. (Finally it's a bit sloppy to have such arbitrary extent values that warp will preserve and I routinely snap them to the source alignment, and I'm interested to provide such a feature in the warper options too.) Hope that helps! On Fri, 6 Jan 2023, 03:16 Daniel Mannarino, <daniel.mannar...@gmail.com> wrote: > Hi folks! > > > > We have a process that generates web mercator zoom levels by resampling > from the next higher up, from the original resolution down to zoom level 0. > Currently we do so with gdalwarp, but a team member realized that > gdal_translate is capable of the same, and at first glance is much faster. > Unfortunately, using the two commands (below) which I intended to do > exactly the same thing winds up with slightly different results (can/should > I post a small image here to illustrate?). > > I won't ask "Which is correct?" because the answer is almost certainly > "That depends on what you want". But can someone help me understand the > difference between what is done with the gdalwarp command versus the > gdal_translate one so that I can decide if using gdal_translate is an > acceptable solution? Here are the two commands: > > > > gdalwarp -te 12523442.714243278 0.0 15028131.257091936 2504688.5428798534 > -ts 65536 65536 -r mode -co TILED=YES ../source_tiles/everything.vrt > 007R_013C_warp_mode.tif > > > > gdal_translate -r mode -projwin 12523442.714243278 2504688.5428798534 > 15028131.257091936 0.0 -outsize 65536 65536 -co TILED=YES > ../source_tiles/everything.vrt 007R_013C_translate_mode.tif > > > > Thank you! > > Daniel Mannarino > _______________________________________________ > gdal-dev mailing list > gdal-dev@lists.osgeo.org > https://lists.osgeo.org/mailman/listinfo/gdal-dev >
_______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/gdal-dev
