Hi,
And thanks for your tests ! For those who want to make some tests with
the same file than me (tiff RGB - 293 Mo),I published it on a FTP server :
Adress : ftp://interne.crige-paca.org
Login : gdaldevecw
Password : ftc65*gt
I also published in this FTP the comparative table of my tests and
results. I agree with Jean-Claude. It could be great to make tests with
the same original file and the same gdal_translate command
(gdal_translate -of "ECW" -co LARGE_OK="YES" -co TARGET="90" -co
DATUM="RGF93" -co PROJ="LMFRAN93" file_in.tif file_out.ecw).
I made some complementary tests on GDAL 1.9 (Windows-Linux-MacOSX /
32-64 bits) and always get the same differencies...
Best regards,
Sylvain
--
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
<http://www.crige-paca.org/>
Le 28/02/2012 08:28, Jean-Claude Repetto a écrit :
On 02/28/12 02:15, andreaerdna wrote:
- the 32-bit GDAL versions/builds 1.7.3 from www.gisinternals.com/sdk
and
1.7.0b2 from FWTools 2.4.7 on 64-bit/32-bit Windows tested using
gdal_translate without a "TARGET" creation option behave the same way
producing an output ECW file of 164.417.131 bytes (with the same
checksums)
that is about 3 times bigger than that produced by the other 32-bit GDAL
versions tested.
Hi,
The default TARGET parameter value has been changed from 75% to 95%
between GDAL 1.7.3 and GDAL 1.8.
So please allways give the TARGET parameter when you make your tests,
and use the same TARGET for all your tests.
It would be useful if you can make your test image available to
everybody, so that people can compare their results with yours. If it
is not possible, use a freely available image, such as NASA BlueMarble.
Could you also test on Linux 32 bits ?
Best regards,
Jean-Claude
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
