Hi Even, The gdal_merge.py behaviour I meant was exactly that the resulting directly created compressed tiff can waste space.
-Jukka- ________________________________________ Even Rouault wrote >Le vendredi 22 novembre 2013 19:18:56, Rahkonen Jukka a écrit : >> Hi, >> >> I think that the reason is that is is hard/impossible to make an optimal >> deflate compressed tiff when gdal_merge goes through the circle "open file >> - add data - close file". There are other alternatives to test: - You can >> use a non-optimal deflate compressed tiff as a temporary file - is is >> lossless and the final result will be identical. - Use gdalbuildvrt for >> making an interim .VRT file and convert that to final product instead of >> using gdal_merge.py. Perhaps GDAL can make an optimal deflate compressed >> image from the VRT file with one run. > >> A hint: If you play with big datasets and deflate compressed images GDAL >> sometimes does not guess when it should jump into BigTIFF. Give the -co >> bigtiff=yes parameter manually if you feel that it could be needed. > >> If Even is still sneaking here he can say if the behaviour you see with >> gdal_merge.py is expected or a bug. Otherwise you can ask it from gdal-dev >> mailing list. > I'm not sure which gdal_merge behaviour Jukka is refering too, but indeed it > might not be the appropriate tool to mosaic stuff into a compressed geotiff > due > to the way it operates (it can have to do uncompression / recompression cycles > that will waste space in the geotiff). gdalbuildvrt to do the mosaicing > (without a disk footprint of only a few bytes) followed by gdal_translate to > compress will lead to the optimal file size. > > -Jukka- > > ________________________________ > Jonathan Moules wrote: > > On 21 November 2013 21:23, Rahkonen Jukka > <jukka.rahko...@mmmtike.fi<mailto:jukka.rahko...@mmmtike.fi>> wrote: I > would try if it makes difference to save result from gdal_merge first into > uncompressed tiff and compress it with deflate in a separate run. > > > Yep, that worked. Although that's unfeasible for the full dataset; I don't > have enough diskspace for an uncompressed version. > > Also, doing that (deflate, level 9), the tiled version was about 15% > smaller than the untiled version. > > Thanks! > Jonathan > > > Jukka Rahkonen > > ________________________________ > Jonathan Moules<mailto:jonathanmou...@warwickshire.gov.uk> wrote: > > It is not completely surprising. If there are no repeateable patterns in > the uncompressed stream, deflate can result in (a bit) larger files. You > might want to try with -co INTERLEAVE=BAND added to see if it improves > things. But generally DEFLATE is not appropriate for aerial imagery. > > > The "bit" larger files are typically about double the size of the > uncompressed one! > > I know deflate isn't as well suited, but I'm experimenting with a whole > bunch of different settings to see what sticks. > > ====== > > I'm now trying to get the four band into GeoServer (regular GeoTIFF), but > it's not having it. I've created it the same way as the 3 band: gdal_merge > -q -o rgbi.tif -of GTiff -co TILED=YES -co BIGTIFF=YES -co COMPRESS=JPEG > -co JPEG_QUALITY=25 -co BLOCKXSIZE=512 -co BLOCKYSIZE=512 --optfile > tiff_list.txt (low compression for testing). No pyramids yet. > > The first thing I notice is that the 4 band is *smaller* (118MB) than the 3 > band (164MB). > > The gdalinfo for the four band: > > Driver: GTiff/GeoTIFF > Files: rgbi.tif > Size is 8000, 24000 > Coordinate System is: > PROJCS["OSGB 1936 / British National Grid", > [brevity removed] > Origin = (419000.000000000000000,240000.000000000000000) > Pixel Size = (0.125000000000000,-0.125000000000000) > Metadata: > AREA_OR_POINT=Area > Image Structure Metadata: > COMPRESSION=JPEG > INTERLEAVE=PIXEL > Corner Coordinates: > Upper Left ( 419000.000, 240000.000) ( 1d43'22.27"W, 52d 3'27.49"N) > Lower Left ( 419000.000, 237000.000) ( 1d43'22.87"W, 52d 1'50.38"N) > Upper Right ( 420000.000, 240000.000) ( 1d42'29.76"W, 52d 3'27.37"N) > Lower Right ( 420000.000, 237000.000) ( 1d42'30.39"W, 52d 1'50.26"N) > Center ( 419500.000, 238500.000) ( 1d42'56.32"W, 52d 2'38.88"N) > Band 1 Block=512x512 Type=Byte, ColorInterp=Red > Mask Flags: PER_DATASET ALPHA > Band 2 Block=512x512 Type=Byte, ColorInterp=Green > Mask Flags: PER_DATASET ALPHA > Band 3 Block=512x512 Type=Byte, ColorInterp=Blue > Mask Flags: PER_DATASET ALPHA > Band 4 Block=512x512 Type=Byte, ColorInterp=Alpha > > ---- > > The GeoServer logs contain an error (WMS GetMap request): > 2013-11-20 17:56:07,498 ERROR [org.geoserver.ows] - > java.lang.RuntimeException: javax.imageio.IIOException: Unsupported Image > Type at > com.sun.media.jai.imageioimpl.ImageReadOpImage.computeTile(ImageReadOpImag > e.java:706) at > com.sun.media.jai.util.SunTileScheduler.scheduleTile(SunTileScheduler.java > :904) at javax.media.jai.OpImage.getTile(OpImage.java:1129) > .... > Caused by: javax.imageio.IIOException: Unsupported Image Type > at com.sun.imageio.plugins.jpeg.JPEGImageReader.readInternal(Unknown > Source) at com.sun.imageio.plugins.jpeg.JPEGImageReader.read(Unknown > Source) at > it.geosolutions.imageioimpl.plugins.tiff.TIFFJPEGDecompressor.decodeRaw(TI > FFJPEGDecompressor.java:278) > > > Ideally I'm looking to have one source file which can be symbolised as not > just RGB, but also as any number of false-colour possibilities. > > Any suggestions anyone? I'm stumbling into new (for me) territory here. > > Thanks, > Jonathan > > This transmission is intended for the named addressee(s) only and may > contain sensitive or protectively marked material up to RESTRICTED and > should be handled accordingly. Unless you are the named addressee (or > authorised to receive it for the addressee) you may not copy or use it, or > disclose it to anyone else. If you have received this transmission in > error please notify the sender immediately. All email traffic sent to or > from us, including without limitation all GCSX traffic, may be subject to > recording and/or monitoring in accordance with relevant legislation. > > --------------------------------------------------------------------------- > --- Shape the Mobile Experience: Free Subscription > Software experts and developers: Be at the forefront of tech innovation. > Intel(R) Software Adrenaline delivers strategic insight and game-changing > conversations that shape the rapidly evolving mobile landscape. Sign up > now. > http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktr > k _______________________________________________ > Geoserver-users mailing list > Geoserver-users@lists.sourceforge.net<mailto:Geoserver-users@lists.sourcefo > rge.net> https://lists.sourceforge.net/lists/listinfo/geoserver-users > > > > This transmission is intended for the named addressee(s) only and may > contain sensitive or protectively marked material up to RESTRICTED and > should be handled accordingly. Unless you are the named addressee (or > authorised to receive it for the addressee) you may not copy or use it, or > disclose it to anyone else. If you have received this transmission in > error please notify the sender immediately. All email traffic sent to or > from us, including without limitation all GCSX traffic, may be subject to > recording and/or monitoring in accordance with relevant legislation. > > --------------------------------------------------------------------------- > --- Shape the Mobile Experience: Free Subscription > Software experts and developers: Be at the forefront of tech innovation. > Intel(R) Software Adrenaline delivers strategic insight and game-changing > conversations that shape the rapidly evolving mobile landscape. Sign up > now. > http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktr > k _______________________________________________ > Geoserver-users mailing list > Geoserver-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/geoserver-users -- Geospatial professional services http://even.rouault.free.fr/services.html ------------------------------------------------------------------------------ Shape the Mobile Experience: Free Subscription Software experts and developers: Be at the forefront of tech innovation. Intel(R) Software Adrenaline delivers strategic insight and game-changing conversations that shape the rapidly evolving mobile landscape. Sign up now. http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktrk _______________________________________________ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
