I use theconverter DLL to generate TIFF group 4 fax encoding from PDF documents in one of our applications and I noticed that the image quality is rather poor compared to other PDF to TIFF conversion tools, but the text conversion is correct. The following code extract shows how I use it:
PDFToImageEnableErrorDiffusion( TRUE );
PDFToImageSetFileNameSuffix( "%08d" );
m_progress.Create( "Generating images from PDF...", 100, m_list.size() * 2, TRUE, 0 );
g_pageList = this;
PDFToImageSetProgressCBFunc( PageProgressCBFunc );
PDFToImageConverter2( info.m_ctrlfile.c_str(), ( info.m_imageloc + ".tif" ).c_str(), NULL, NULL, 300, 300, 1, COMPRESSION_CCITTFAX4, 100, FALSE, 0, m_list.PhyNum(), m_list[m_list.size() - 1].PhyNum() );
Below you can find a small print-screen of the image generated with the pdf2image DLL and an example generated with another application (XXXXX). You can see that the text quality is equal…
Above is generated with pdf2image...
… and this is generated with XXXXX.
But from the same run, you can see that the image quality is far less…
Above is generated with pdf2image…
… and this is from the run with XXXXX.
Is there a way I can improve the quality of the generated TIFF group 4 fax encoded images?
Can you please email to us your sample PDF file for checking?
In attachment you can find a zip file with 5 documents.
UZBE_11.pdf - The pdf file for which I’m trying to generate tiff files (1 page per file)
XXXXX_00000310.tif - The tiff file generated with XXXXX for page 310 in the document
XXXXX_00000311.tif - The tiff file generated with XXXXX for page 311 in the document
verypdf_00000310.tif - The tiff file generated with the verypdf library for page 310 in the document
verypdf_00000311.tif - The tiff file generated with the verypdf library for page 311 in the document
The verypdf_* files are generated using the code further down in this mail.
The XXXXX_* files are generated with XXXXX using the following settings:
* TIFF image, black&white, 1 bit
* High quality dithering
* CCITT group 4 encoding
As you can see, there is a big difference in quality between the verypdf generated images and the images generated with XXXXX. How can I improve the quality of the images generated with verypdf?
Thanks for your sample files, we have checked your sample files, we noticed the difference is caused by halftone technology, the XXXXX and VeryPDF are using different halftone patterns. This is not a problem, we will try to use different halftone patterns to convert color PDF pages to black and white TIFF files in the future releases.
I have still a couple of questions.
1. Which dithering will be used (halftoning, Floyd-Steinberg,...)? Personally, I would suggest using Floyd-Steinberg since this produces the best quality for photographic images and it has a very good approximation for shades of gray.
2. Is there already timeline for this implementation? I need to know this to keep my customers informed about this improvement. It is also on my list of hot topics. For one of my customers this is a showstopper.
We are using halftoning algorithm, I know Floyd-Steinberg is work better, but the speed of Floyd-Steinberg is slow, however, we will try to provide an option to choose halftoning or Floyd-Steinberg in the next version of PDF2Image SDK product.
Thanks for your message, our engineers are busy on some critical projects now, so we haven't an approximate date for when will the next version be released at the moment. However, we will let you know after new version is ready, thanks for your patience.
Continue to previous email, you can use following command line to convert your PDF file to high quality black and white TIFF file,
nResult = PDFToImageConverter("C:\\test.pdf", "C:\\out.pdf",NULL,NULL,200,200, 24, COMPRESSION_CCITTFAX4,100,TRUE,TRUE,-1,-1);
if you set "bitcount" to 24 and "compression" to COMPRESSION_CCITTFAX4, PDFToImageConverter() function will create high quality BW TIFF file with Floyd-Steinberg technology, please use above source code to try.