[Gdal-dev] Ecw resampling problem when reading region larger than output buffer?

[Asger Petersen]

Hi Frank,

I'm using version 1.3.1.0 from FWTools 1.0.0b2 on WinXP so I have just
filed a bug report with an example file attached.

Cheers
Asger

> -----Original Message-----
> From: Frank Warmerdam [mailto:fwarmerdam at gmail.com] On Behalf Of Frank
> Warmerdam
> Sent: 5. april 2006 16:12
> To: Asger Petersen
> Cc: gdal-dev at lists.maptools.org
> Subject: Re: [Gdal-dev] Ecw resampling problem when reading region
larger
> than output buffer?
> 
> Asger Petersen wrote:
> > Hi list
> >
> > I get funny results when using RasterIO to read a region from ECW
that
> is
> > larger than the outputbuffer ie doing super sampling.
> >
> > When the output buffer is same size or smaller than the read region
> > everything is fine. When output buffer is larger than the read
region I
> get
> > a result where the number of lines in the output seems to be fine
but
> the
> > number of pixels is 1/3 of what is should be (see attached files).
> >
> > I've tried different formats and this only happens with ecw.
>  >
> > Oh, and to complicate things a bit I'm using Simon Perkins' .NET
> bindings.
> > Works: band.RasterIO(Gdal.RWFlag.Read, 4000, 4000, 351, 351, data,
351,
> > 351, Gdal.DataType.Byte, 3, 1056);
> >
> > Doesn't work: band.RasterIO(Gdal.RWFlag.Read, 4000, 4000, 351, 351,
> data,
> > 352, 352, Gdal.DataType.Byte, 3, 1056);
> >
> > Isn't it possible to get GDAL to supersample from ecw files?
> 
> Asger,
> 
> What version of GDAL are you using?  Hopefully at least 1.3.1?  There
was
> a bug in the past with supersampling, as it had to be implemented as a
> specific case with GDAL doing the oversampling as the ECW SDK did not
> support returning supersampled images.   It is quite plausible that
there
> are still some problems in this in fairly specific cases.
> 
> If you still have this problem with at least GDAL 1.3.1 (or better yet
> a recent CVS snapshot), and if you file a bug report with the details
> and a pointer to an image on which it can be demonstrated I'll dig
into
> it.
> 
> The c# shouldn't be an issue so providing just the C# rasterio call
> should be sufficient.  I just want to ensure I am trying to reproduce
> it against the same image at my end as you ran into the problem with.
> 
> 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    | President OSGF,
http://osgeo.org