> What about an approach where the out-of-bound values are assigned the last<br>
> value only if reprojection is used (and if the layer does not have a null<br>
> value)?<br><br>My two cents, from a user perspective.<br>We're facing this dilemma with Geotools, where reprojected color mapped rasters (thorugh SLD) result with borders having the lower color from the SLD rastersymbolizer. This is a big problem for many users, and we have to reproject the raster before serving it.<br>
It would be very important to be able to manage nodata values for out-of-border pixels, to be able to manage transparency and avoid those ugly borders...<br><br>giovanni<br><br><div class="gmail_quote">2011/1/18 Marco Hugentobler <span dir="ltr"><<a href="mailto:marco.hugentobler@sourcepole.ch">marco.hugentobler@sourcepole.ch</a>></span><br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div class="im">> Agreed, int to double conversion could be optimal.<br>
<br>
</div>oops, that should be 'Agreed, int to double conversion could be problematic'<br>
<br>
Am Dienstag, 18. Januar 2011, um 09.19:23 schrieb Marco Hugentobler:<br>
<div><div></div><div class="h5">> > I thought that if data type of the source is integer the provider<br>
> > could represent them as floating point. Byte can be represented as<br>
> > integer. Bad solution however.<br>
><br>
> Agreed, int to double conversion could be optimal.<br>
><br>
<br>
><br>
> This would be quite similar to your current approach, except that people<br>
> will have the appropriate raster appearance if they don't use reproj. And<br>
> if they do, they have an undesired border, but still the right colors in<br>
> the raster area.<br>
><br>
> Regards,<br>
> Marco<br>
><br>
> Am Montag, 17. Januar 2011, um 17.34:22 schrieb Radim Blazek:<br>
> > On Mon, Jan 17, 2011 at 10:06 AM, Marco Hugentobler<br>
> > <<a href="mailto:marco.hugentobler@sourcepole.ch">marco.hugentobler@sourcepole.ch</a>> > Using NaN sounds like a good idea<br>
> > and Qt has platform independent support for<br>
> ><br>
> > > it (qIsNan & co.).<br>
> > > All other solutions I can think of seem to be more complicated (e.g.<br>
> > > force a transparency value only if raster is reprojected).<br>
> > ><br>
> > >>float + NaN for byte/int?<br>
> > >><br>
> > > This is not clear to me. Could you explain your approach for byte/int?<br>
> ><br>
> > I thought that if data type of the source is integer the provider<br>
> > could represent them as floating point. Byte can be represented as<br>
> > integer. Bad solution however.<br>
> ><br>
> > Radim<br>
<br>
<br>
--<br>
Dr. Marco Hugentobler<br>
Sourcepole - Linux & Open Source Solutions<br>
Churerstrasse 22, CH-8808 Pfäffikon SZ, Switzerland<br>
<a href="mailto:marco.hugentobler@sourcepole.ch">marco.hugentobler@sourcepole.ch</a> <a href="http://www.sourcepole.ch" target="_blank">http://www.sourcepole.ch</a><br>
Technical Advisor QGIS Project Steering Committee<br>
_______________________________________________<br>
Qgis-developer mailing list<br>
<a href="mailto:Qgis-developer@lists.osgeo.org">Qgis-developer@lists.osgeo.org</a><br>
<a href="http://lists.osgeo.org/mailman/listinfo/qgis-developer" target="_blank">http://lists.osgeo.org/mailman/listinfo/qgis-developer</a><br>
</div></div></blockquote></div><br>