[gdal-dev] Doubt about memory raster

[Jorge Arévalo]

2010/2/9 Frank Warmerdam <warmerdam at pobox.com>:
> Jorge Arévalo wrote:
>>
>> Mmm... But I suppose this is in case of rasters with pixel type =
>> Byte. In that cases, should I always follow this scheme? pArray for
>> first band, pArray + 1 for second band, and so on (interlaced)? I
>> mean, why don't pArray for first band, pArray + raster_width for
>> second band...(non interlaced)?
>>
>> Anyway, I'm working in a "WKT Raster format to GDAL format" parser. I
>> get all the bands in different arrays: void * pBandData, for example.
>> And the pixel type of these bands can be Byte, Int16, Float32, etc.
>> So, I don't have one array for all the raster data, I have a different
>> one per band.
>>
>> In such situation, has sense this scheme?
>>
>> DATAPOINTER = pBandData
>> PIXELOFFSET = (GDALDataTypeSize(raster_pixel_size) / 8) (this is, 1 =
>> pixel size byte, 2 = pixel size int16, 4 = pixel size int32, etc)
>> LINEOFFSET = PIXELOFFSET * raster_width_in_px
>
> Jorge,
>
> Yes, this is correct, but you can just omit the PIXELOFFSET and
> LINEOFFSET values in this situation since this is the default
> configuration.  You only need to specify them if you are doing
> something complicated, like interleaving.
>
Ok. So, interleaving is not the default behaviour when creating a
raster from scratch, from mem.

Many thanks!

Best regards,
Jorge

> 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
>
>