[gdal-dev] gdalbuildvrt and nulls again - now from Photoshop

[William Kyngesburye]

On Jul 14, 2009, at 1:28 PM, Even Rouault wrote:

> Le Friday 10 July 2009 21:14:18 William Kyngesburye, vous avez écrit :
>> This time I'm taking some Photoshop-created tiffs that I georeference
>> with a TFW.  These tiffs I save with transparency mask (not layers)  
>> to
>> handle nulls.
>>
>> Reported by gdalinfo, they appear as:
>>
>> Band 1 Block=12775x5 Type=Byte, ColorInterp=Red
>>   Mask Flags: PER_DATASET ALPHA
>> Band 2 Block=12775x5 Type=Byte, ColorInterp=Green
>>   Mask Flags: PER_DATASET ALPHA
>> Band 3 Block=12775x5 Type=Byte, ColorInterp=Blue
>>   Mask Flags: PER_DATASET ALPHA
>> Band 4 Block=12775x5 Type=Byte, ColorInterp=Alpha
>>
>> But when I run gdalbuildvrt on them, they are skipped:
>>
>> gdalbuildvrt does not support heterogenous band numbers. Skipping
>> [sometiff].
>
> This warning means that you are giving a list of datasets that don't  
> have the
> same number of bands. For example, if the first dataset examined by
> gdalbuildvrt has only 3 bands, it will ignore any further dataset  
> whose band
> number is not 3. (if the error message isn't clear, you're welcome  
> to suggest
> a better one)


Ah, yes, some of the images don't have nulls, and so I didn't save the  
transparency.

Is it to generate the vrt with mixed rasters?  Maybe merge the vrts of  
one set of no-null rasters and one set of rasters with nulls in a text  
editor?

Is there a technical reason (in the vrt format) that gdalbuildvrt  
can't mix 3-band (no nulls) and 4-band (nulls) rasters?  Or is it just  
rejecting the mismatches for simplicity?

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"Time is an illusion - lunchtime doubly so."

- Ford Prefect