Re: [mapserver-users] Average/Bilinear Resampling and Alpha Bands

2010-03-15 Thread Robert Sanson
Hi Jason Well, ECW use wavelet compression, which is lossy, and data values can get changed slightly. regards, Robert >>> Jason Beverage 16/03/2010 10:27 a.m. >>> Hey again, Some more information, the file I was using was an ECW I converted from a tif. The black splotches appear if I use

Re: [mapserver-users] Average/Bilinear Resampling and Alpha Bands

2010-03-15 Thread Jason Beverage
Hey again, Some more information, the file I was using was an ECW I converted from a tif. The black splotches appear if I use the ECW, but if I use the original tif file I don't see them. However, when using AVERAGE or BILINEAR, it appears as if the alpha values are either completely transparent

Re: [mapserver-users] Average/Bilinear Resampling and Alpha Bands

2010-03-15 Thread Robert Sanson
Hi Jason That's not unexpected. RESAMPLE=NEAREST preserves the "nodata" vales used in the mask. BILINEAR or AVERAGE combine values from adjoining pixels, so the "nodata" pixels get some other value. regards, Robert >>> Jason Beverage 16/03/2010 4:53 a.m. >>> Hi all, I have some imagery th

Re: [mapserver-users] Average/Bilinear Resampling and Alpha Bands

2010-03-15 Thread Frank Warmerdam
Jason Beverage wrote: Hi all, I have some imagery that is in UTM that contains an alpha band that masks out areas with no data. When I use MapServer to reproject it to epsg:4326, the masked out area looks good if I set RESAMPLE=NEAREST on the layer. If I change the RESAMPLE option to BILINEAR

[mapserver-users] Average/Bilinear Resampling and Alpha Bands

2010-03-15 Thread Jason Beverage
Hi all, I have some imagery that is in UTM that contains an alpha band that masks out areas with no data. When I use MapServer to reproject it to epsg:4326, the masked out area looks good if I set RESAMPLE=NEAREST on the layer. If I change the RESAMPLE option to BILINEAR or AVERAGE I get black s