[Xastir] Problem with geotiff files

[Tom Russo]

On Mon, Oct 11, 2004 at 08:34:27PM +0100, we recorded a bogon-computron collision of the <j_cox at swiftdsl.com.au> flavor, containing:
> On Mon, 2004-10-11 at 09:58, Tom Russo wrote:
> > On Mon, Oct 11, 2004 at 05:32:16PM +0100, we recorded a bogon-computron collision of the <j_cox at swiftdsl.com.au> flavor, containing:
> > > Hi all,
> > > 	I am trying to setup geotiff files. I keep having problems with xastir
> > > reorting back:
> > > 
> > > *** geoTIFF file /usr/local/share/xastir/maps/south/newy.tif is not in
> > > the proper format.
> > > *** Please reformat it and try again.
> > 
> > This rather generic sounding error message actually means the file has either
> > more than one sample per pixel, something other than 8 bits per sample, 
> > or more than one plane.  Those aren't actually geotiff parameters, but tif 
> > parameters, so you need to use "tiffinfo" to see that.  
> > 
> > The only way to get those to display is to do some kind of conversion
> > so that they fit the parameters xastir needs, or to hack xastir
> > so it supports a wider range of tiff images.
> -- 
> 
> Thanks Tom,
> 	That solved the problem - I re-exported from global mapper and it 
> pulled up the pics OK. It seamed to have a positional error in it the image 
> positions as they don not line up with the mif/tab files. It is also slow so,
> I might have had debug turned on in one of the lib's.
>  Only question left is dose Xastir use GoeTiff from GDAL or direct to the libgeotiff ?

Xastir doesn't use gdal for very much at all, it gets its geotiff capability
from libgeotiff.  That is supposed to change eventually.

If your image is very much larger than a USGS 7.5" quad, you might have
a significant positional error due to the approximate way that xastir
attempts to convert UTM rasters to lat/lon.  It doesn't use gdal (yet),
it reads direct from the geotiff, and does something like a linear 
rubbersheeting.  For USGS quads this isn't so bad, but for larger scale
the positional error can be significant.

One thing I've had luck with is to use gdalwarp to convert rasters to 
an unprojected lat/lon, and then use those with xastir.  If the boundary
warps to something that isn't rectangular, though, tiling doesn't work
so nicely.  Also, the method xastir uses to skip over scan lines is optimized
for 7.5" USGS quads and that didn't work properly when I slammed in the 
ability to read unprojected rasters.  Rather than spend time to come up with a 
new optimized method, I just disabled that optimization when unprojected
rasters are used. 

When xastir's gdal support is done all this should change.  I don't know
the status of that effort.

-- 
Tom Russo    KM5VY     SAR502  DM64ux         http://www.swcp.com/~russo/
Tijeras, NM  QRPL#1592 K2#398  SOC#236 AHTB#1 http://www.qsl.net/~km5vy/
(1) Ignorance of your profession is best concealed by solemnity and silence,
which pass for profound knowledge upon the generality of mankind.
                -------"Advice to Officers of the British Army", 1783