my pdfs are too big

[drb]

Hello,
I am using the latest version of Xnview. I am running into a small issue. Kindly help me out.

I wish to convert several .tiff files into a multipage pdf. THis is what I did.
Browse the folder in xnview, select all, create>multi-page file.
In that window, in options, for write tab, for pdf parameters, I tried several things like compression none, zip etc and then quality 50%. Still my pdf file generated is too big.

For example, i have around 500 .tiff files in a folder and wish to convert them into a single multipage pdf. Each is around 60 kb in size. But the pdf produced is around 122 MB. Super big.

Please help on how to optimize the pdf to a much smaller size without reducing too much of the quality. These are mainly text and not images in the .tiff.

Thank you,

drb

[cday]

Could you upload a sample 60KB file?

If the TIFFs are small the PDF file containing them shouldn't be much larger, as the PDF format is only really a 'container' for the image(s) in the file. And if the TIFFs are black and white images with 1-bit colour depth, the 'FAX' compression option should produce a small file size.

Edit:

If your images are black and white the optimum compression option is 'CCITT G4' ; that is one of the 'FAX' options but XnView is slightly inconsistent in naming the options in different menus...

You likely need this setting:

[drb]

I made a mistake. Most of my files are around 100 kb in size.
Here is a sample .tiff

and also the screenshot of Xnview that I get when I create to get pdf looks different than yours. Why is tiff selected in yours? I thought pdf should be selected? Am I wrong?

Please help.

[cday]

Here is a sample .tiff

The image you uploaded downloads as a JPG image, a format which doesn't support black and white images; could you please upload a sample TIFF as an attachment so that it is possible to be sure about the colour depth of the original image, which will affect the optimum compression method to use.

... and also the screenshot of Xnview that I get when I create to get pdf looks different than yours. Why is tiff selected in yours? I thought pdf should be selected? Am I wrong?

I made a screenshot with an incorrect setting, sorry about the confusion...

Edit:

If your TIFFs are confirmed as 1-bit colour depth as I suspect, the following save option should result in the smallest PDF file size:

Note: The 'Quality' setting is only applicable to JPEG compression, all other available compression options are lossless.

[drb]

My apologies. Postimage.org converted the .tiff to .jpg
I have attached the original .tiff for your perusal.

Please help.
Thanks,

[cday]

Your image was a black and white 1-bit image with CCITT G4 'Fax' encoding, as you may already know.

I've created a PDF file using File > Save as... with the file type set to PDF and Options set to FAX, as in the screenshot in my last post above.

Although the original image is now in a PDF 'wrapper' with similar compression, the file size has actually decreased marginally!

To create a multi-page PDF file use the Tools > Multipage file function with similar FAX compression.

00000177.pdf

(89.32 KiB) Downloaded 158 times

[drb]

Sorry to bother you further.
As you said, that is only marginal decrease.

Is there a way to reduce it significantly without much loss in quality?
I have Nitropdf and tried to optimize with 300 dpi etc, but since the starting size of pdf that got created with xnview is around 122 MB, it crashes. Anything beyond 30 MB or more, nitropdf crashes.

thanks,

[cday]

I'm not sure that I understand what you are asking...

The image you uploaded is a 1-bit TIFF with a file size around 90KB, which is a very reasonable file size because it is optimally compressed with one of the 'Fax' options; your original request was to convert your TIFF images to PDF files without a large (massive...) increase in file size, and the PDF file I uploaded has the same (or actually marginally smaller) file size, so it should meet your needs.

However, if your original TIFF file (or files) are saved using another compression option, or saved as grayscale or colour images, you will indeed obtain a very much larger resulting file which plausibly might crash NitroPro...

But you images as 1-bit TIFFs are already optimal [within the quality limits of black and white images, which require higher DPI values to reproduce well) and are similarly optimal as PDFs when created as above...

[cday]

Is there a way to reduce it significantly without much loss in quality?
I have Nitropdf and tried to optimize with 300 dpi etc, but since the starting size of pdf that got created with xnview is around 122 MB, it crashes. Anything beyond 30 MB or more, nitropdf crashes.

Are you saying that 90KB per page, 9MB per 100 pages is still too large?

If that is the case there are options to decrease the size further, but they would mean either accepting lower image quality -- though downsampling to 300DPI or possibly 400DPI using the Image > Resize... function (or, probably better, scanning at 400DPI for future scans), or using commercial software to save the existing 600DPI TIFF images with more-efficient JBIG2 compression: the 'lossy' version may be quite effective, although with a slight risk of inserting incorrect characters in places.

Alternatively, Adobe Acrobat Clearscan could potentially shrink the file significantly while also enhancing the text quality, although a reduction in file size may only be obtained for a reasonable run of pages...

[drb]

I have 494 .tiff images approximately of that size 90-100 kb (random checking showed that range). And when I create a pdf, the size is 122 MB. Not sure how that is happening.

Now in your reduction, it shows 89 kb instead of 91 kb. Hence wondering if I convert that many images to make a single multi page pdf, then I will end up with marginal decrease in size.

After reading your responses, I guess I am pretty sure making some silly mistake or not reading something proper.

thanks,
drb

[cday]

I have 494 .tiff images approximately of that size 90-100 kb (random checking showed that range). And when I create a pdf, the size is 122 MB. Not sure how that is happening.

I'm not sure either, I would expect the size of the resulting PDF file to be little larger than the total size of the images if similar compression is used, or around 50MB. Just possibly Tools > Multipage file > Create uses a different 'Fax' compression option??

If you care to place as many TIFFs as you can in a ZIP archive and upload that file, I would be interested in having a quick look at the resulting PDF file produced, and possibly look at some alternative options. The maximum attachment size allowed at the moment is 1MiB.

[drb]

I am attaching around 10 files zipped.

I tried some of your methods, still it is enormously bigger than the .tiff files all put together. Not sure what's going on.

Thanks so much.

[cday]

I have no problem producing a multi-page PDF of about the expected size from the TIFF files you uploaded using this method:

Tools > Multipage file > Create :

Image_1.png (16.23 KiB) Viewed 3561 times

Xn_Multi-page_FAX.pdf

(983.03 KiB) Downloaded 141 times

How important is file size to you, given that your images are black and white and now encoded about as efficiently as is possible using readily available freeware software? Image files are large, and if you had to scan to grayscale (necessary for any pages that have half-tone graphics or photographs) or to colour, the reulting image files even efficiently encoded would be much larger?

Nonetheless, I have had a quick look at resizing the original images to reduce the final PDF file size, and the result may possibly be of interest. I'll include it in a later post.

Although your images are scanned at 600DPI, the small size of the text and the second non-English language used means that the images have quite a high level of detail that tends to increase the file size and somewhat complicate effective use of more efficient compression methods.

Could you please identify the second language used in the text, and indicate whether it is included in this list of languages:

http://finereader.abbyy.com/professiona ... pecs/#lang

[drb]

The second language is not in the abbyy.

However, I am not sure what went wrong, I updated to latest version of xnview and did follow exactly your settings as above and I got all the 494 .tiff files into a multipage pdf file at 39 MB size.

That should work great for me. I must have done something wrong in my previous steps hence turned out to 122 MB in size.

I am not sure how to thank you. Words fail me.

Each second you spent on reading my post, analyzing, working in your computer and posting it back etc.......I very very deeply appreciate it. Not many are there like you. Whether this process had failed or not, I would appreciate the same.
Thankfully, it worked for me.

Again, thanks so much for spending your time. My humblest salutations to you.


PS: I tried to optimize file size in nitropdf - even downsized resolution to 150 dpi, but still it stayed at 39 MB. Maybe as you said, i shall covert the images to lower resolution first and then convert into pdf.

[cday]

... I got all the 494 .tiff files into a multipage pdf file at 39 MB size.

That should work great for me. I must have done something wrong in my previous steps hence turned out to 122 MB in size.

If the resulting file size is acceptable using the above method with 'FAX' / 'CCITT G4' compression that is the easy way forward.

If you are scanning yourself and wish to experiment a bit to see if it is possible to obtain a smaller file size while maintaining acceptable quality, you could try scanning at 400DPI or 300DPI. You might also explore the settings on your scanner to see if it's possible to obtain a slightly 'cleaner' image without the breaks within characters present in the existing scans. Some scanner software also provides a selectable enhancement mode for text which can be very effective, but I'm not sure if it is available for black and white images.

You mentioned the possibility of downsampling the images to reduce the file size. I did a quick test yesterday and was surprised that the image quality did not seem to be too badly affected: many characters were slightly degraded, but some were actually enhanced slightly! The file size was reduced from 983KB to 514KB and you can assess the image quality yourself if you download the file below.

I used Image > Resize... to reduce the height and width by 50% while maintaining the page size: note that the Resize operation produces a grayscale image, so it is necessary to follow it with an operation to convert the resulting image back to 1-bit colour depth. Both operations offer a choice of resampling methods and I used the default settings, although they may or may not be optimal.

The two files below are the original file again and the downsampled version:

Xn_Multi-page_FAX.pdf

(983.03 KiB) Downloaded 154 times

Xn_Multi-page_FAX_50%.pdf

(513.68 KiB) Downloaded 118 times

Out of interest I also looked at JBIG2 compression using ABBYY FineReader but didn't obtain good results with either the lossless or lossy versions, possible due to the complexity of your images including the breaks in many characters, which makes it hard to minimise the number of instances of a character bitmap that need to be stored, the way in which file size can potentially be mimimised.

ABBYY FineReader also offers an option of producing page images using scalable fonts as used in a word processor to reproduce the output from an OCR process, which also results in a file that is searchable. Partly at least due to the second font used in you images, that didn't work well either as you can see from the file produced:

FR12_Text_over_image_Windows_fonts_Embedded.pdf

(143.98 KiB) Downloaded 52 times

The 'text over image' ( a rather opaque name) generally requires careful proof-reading to correct any character or layout errors, but was defeated in your case by the second language which it couldn't process. You can see the notable reduction in file size, and high quality of the font characters produced, although it is a synthetic image as if the recognised text had been retyped.

I also had a quick look at Adobe ClearType, which uses OCR results when possible to generate matching scalable font characters, and arbitrary scalable shapes for unrecognised characters. That produced good quality searchable output but the file size increased substantially, although on a long document with good text quality very small files can be produced. This is just the first page:

00000016_ClearScan.pdf

(383.92 KiB) Downloaded 53 times

With your NitroPro PDF software with its tabbed viewer you can easily compare the above versions if you download them. You can then decide whether it is worth considering anything except the straight forward option you have now used successfully...