[gdal-dev] GDAL 1.8.0 misreads images with Type=Float32 as Type=Byte

[Greg Coats]

Even, Thanks for pointing out that this GDAL 1.8.0 bug was fixed 6 months ago.
Unfortunately, GDAL 1.8.1 has not been released, no plans have been announced for the release of GDAL 1.8.1.
The predicament is that GDAL 1.8.0 must be used to use gdallocationinfo, but GDAL 1.7.3 (and not GDAL 1.8.0) must be used to read images with Type=Float, which is the type the USA government uses to distribute all digital elevation modes.
Greg

On Jun 26, 2011, at 12:35 PM, Even Rouault wrote:

> Le dimanche 26 juin 2011 17:02:46, Greg Coats a écrit :
>> GDAL 1.7.3, released 2010/11/10, for Mac OS X 10.6, acquired from
>> http://www.kyngchaos.com/software:frameworks, properly reads geo
>> referenced raster images of Type=Float32. For the example 468 MB image,
>> gdalinfo properly reports Type=Float32, and Computed
>> Min/Max=35.024,748.615.
>> 
>> GDAL 1.8.0, released 2011/01/12, for Mac OS X 10.6, acquired from
>> http://www.kyngchaos.com/software:frameworks, improperly read geo
>> referenced images of Type=Float32. For the example 468 MB image, gdalinfo
>> improperly reports Type=Byte, and Computed Min/Max=0.000,255.000.
>> 
>> gregcoats $ source ~/programs/scripts/paths_gdal_173.sh
>> gregcoats $ gdalinfo --version
>> GDAL 1.7.3, released 2010/11/10
>> gregcoats $ gdalinfo -mm floatn40w078_13.flt | tail -3
>> Band 1 Block=10812x1 Type=Float32, ColorInterp=Undefined
>>    Computed Min/Max=35.024,748.615
>>  NoData Value=-9999
>> 
>> gregcoats $ source ~/programs/scripts/paths_gdal_180.sh
>> gregcoats $ $gdalinfo --version
>> GDAL 1.8.0, released 2011/01/12
>> gregcoats $ gdalinfo -mm floatn40w078_13.flt | tail -3
>> Band 1 Block=10812x1 Type=Byte, ColorInterp=Undefined
>>    Computed Min/Max=0.000,255.000
>>  NoData Value=-9999
> 
> You didn't mention the driver concerned by the issue, but from the file 
> extension, it looks like its ESRI HDR. So this should be fixed in 1.8.1 by 
> http://trac.osgeo.org/gdal/ticket/3933