[Gdal-dev] MrSID driver with 16bit output

[Frank Warmerdam]

Chapman, Martin wrote:
> I have a 16bit (GDT_UInt16) tiff file that I am trying to copy to a 
> UInt16 jp2 file using the MrSID esdk 6.x driver.  When I try to remap it 
> to 8-bit uchar it works fine.  When I try to just copy it to jp2 it 
> fails.  It looks like the only diff between the two operations is the 
> datatype used to initialize the JP2WriterManager object in the 
> JP2CreateCopy() method of the mrsiddataset.cpp file.  In 
> MrSIDDummyImageReader::initialize() the datatype is set to 
> LTI_DATATYPE_UINT16 instead of LTI_DATATYPE_UINT8.  Can anyone confirm 
> this to be a problem in the GDAL mrsid driver?

Martin / Kirk,

Could you verify your file is UInt16?  I tried with an Int16 file, and
it crashed.  It turns out I wasn't checking return values from stuff
like the initialize() method, and it was returning an error code indicating
"unsupported datatype [50903]".  If I force the output type to UInt16 things
seem to work fine.   This was all with Geo_ESDK-5.0.6.962.  Perhaps behavior
has changed in ESDK 6?

I am going to tighten up the error checking a bunch, but I don't see much
other action I can take for now.  Kirk can you comment on LTI_DATATYPE_
values that should work with JPEG2000?

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent