On Wed, Aug 12, 2015 at 3:07 PM, Andrea Peri <[email protected]> wrote: > Hi Radim, thx for explain. > Your explanation of extimated strategy could explain why the extimated > is more precise rather than the actual strategy. > I guess are the null values to produce a wrong "actual" results. > > Can you say me in which file I can found the code for the extimate and > actual strategy > to verify my hypotesis.?
https://github.com/qgis/QGIS/blob/master/src/core/raster/qgsrasterinterface.cpp#L395 https://github.com/qgis/QGIS/blob/master/src/providers/gdal/qgsgdalprovider.cpp#L1323 https://github.com/qgis/QGIS/blob/master/src/core/raster/qgsrasterinterface.cpp#L510 Radim > Thx. > > > > 2015-08-12 11:22 GMT+02:00 Radim Blazek <[email protected]>: >> On Tue, Aug 11, 2015 at 3:09 PM, Andrea Peri <[email protected]> wrote: >>> Hi, >>> trying the rendering of a raster floating point using the cumulative count >>> cut. >>> I see it is quite good if I set the values 2.0,100.0% (not 98%) and >>> use the accuracy at value "estimate". >>> Instead always with the same values for cumulative count cut, but >>> using the accuracy "actual". >>> The result is really poor. The image became totally black with only >>> few isles of grey. >>> >>> This is a surprise for me because the accuracy "actual" take a long >>> time and the name seem to say more precision. >>> >>> I like to know what is the strategy used for actual: estimate and the >>> strategy for accuracy: actual. >> >> Actual takes all pixels, estimate takes 250000 pixels (whole data >> extent with resolution giving approximately 250000 pixels). Some >> providers may use different approach to get estimated values, e.g. >> GDAL provider may use approximated values returned by GDAL, IIRC. >> >> Radim >> >> >>> >>> Many thx. >>> >>> -- >>> ----------------- >>> Andrea Peri >>> . . . . . . . . . >>> qwerty àèìòù >>> ----------------- >>> _______________________________________________ >>> Qgis-user mailing list >>> [email protected] >>> http://lists.osgeo.org/mailman/listinfo/qgis-user > > > > -- > ----------------- > Andrea Peri > . . . . . . . . . > qwerty àèìòù > ----------------- _______________________________________________ Qgis-user mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-user
