Garbled output for page images in PDFToImage

[David Newton]

Hi,

I've looked back through the message archives on this and found a few 
questions related to this, but no solutions. Through both the PDFToImage 
command line tool and the convertToImage() method, some pictures in PDFs 
come out with multicoloured 'bands' across them, or are made more 
severely illegible. From iterating through the properties of the images 
in a sample document, it looks like this happens to PNG images which use 
a DeviceRGB ColorSpace, but those that use an Indexed Colorspace are fine.

An issue that demonstrates the problem, with 'before' and 'after' 
images, is https://issues.apache.org/jira/browse/PDFBOX-958 .

Does anyone have any further ideas as to what might be going on, or what 
could be done to fix or work around this? The issue seems to be 
somewhere in or below the getRGBImage() method, as images written 
straight out from here already show the deterioration, before they're 
included in a complete PDF page image.

Thanks
David