Q: We use PDFNet SDK 5.02 on Linux 64bit in its Java flavour for PDF/A
validation and conversion from generic PDF.
We have to create PDF/A documents from souces that are out of our
control. Using new PDFACompliance(true, inputFile, null,
PDFACompliance.e_Level1B, null, 10).saveAs(outputFile, true) in the
Java SDK, the resulting file, however, still exhibits a number of
errors:
• e_PDFA41: Transparency used (ExtGState with soft mask) on 10
objects
• e_PDFA44: Transparency used (Blend mode is not ‘Normal’) on 1
objects
• e_PDFA45: Transparency used (‘CA’ value is not 1.0) on 1
objects
• e_PDFA46: Transparency used (‘ca’ value is not 1.0) on 1
objects
Why does PDFNet not fix these errors? Is there any procedure that can
be performed to also fix these issues?
A: Regarding other errors they can’t be resolved because they could
imply significant loss of information (e.g. transparency flattening
and converting pages to raster images) which could affect accurate
document reproduction and is usually not desirable.
Is there any procedure that can be performed to also fix these issues?
The affected pages could be selectively converted to image (using
‘pdftron.PDF.PDFDraw’ in PDFNet SDK) and the image can be used to
replace the original page. The procedure is:
- Perform PDF/A validation using PDFACompliance.
- Rasterize the affected PDF pages to image (using PDFDraw as shown
in PDFDraw sample) - Add the rasterized page to the document (see AddImage sample)
- Edit the existing PDF page (see the ElementEdit sample) by
replacing all elements with a single image from step 2. - Run PDFACompliance in conversion mode to fix any other PDF/A
errors.