PDF MRC and JPEG compression problems

[xchangeit]

Hello,

when I try to use the MRC compression in PDF-Tools to get small files, I run into several problems:

1. the compression with predefined settings, even best, makes some coloured text parts look foggy or cloudy, in my case a red text area of 5-6 lines at the end of a scanned page.

2. when trying to change the PDF tool by administrating MRC settings, I cannot find a setting that would avoid this error.

3. the only way to avoid the blurred red text is using JPEG compression instead, but this compression seems to completely ignore its own compression settings. It makes very big PDF files like uncompressed (and shows no JPG artifacts), regardless of the compression strength.

My version of PDF XChange Editor is 10.2.1.

[Daniel - PDF-XChange]

Hello, xchangeit

First and foremost, the latest release is 10.3.1.387, so please do update the software and double check that these issues are not already resolved.

After that, if any of these issues persist, could I ask for a copy of a document you are testing with here, as well as screenshots, or an export of your settings for each case?

Kind regards,

[xchangeit]

Hello Daniel,

it happens with the newest version, too.

here you get a scan file and an enlarged screenshot of the problem with MRC compression.

2024-08-24_PDFToolsScan_3_anon.pdf

(164.74 KiB) Downloaded 60 times

The preview of the scan is perfect, and compressing with JPEG does not show these misty areas.
It should be reproducible with every red text compressed with MRC in PDF-Tools.

Thank you for your attention.

[Daniel - PDF-XChange]

Hello, xchangeit

Thank you for the sample, could I also ask for an export of the tool in question? You can export a tool by right clicking on the tool itself through the PDF-Tools UI:

image.png (10.55 KiB) Viewed 6211 times

You may need to place it in a zip folder in order to upload it to the forums. Or you can send it to us by email <Support@PDF-XChange.com>, with a link to this forum post so we can connect the dots.

It would also help to have a screenshot of the preview image, before the issue appears in the file. Though I expect that may contain some sensitive information, perhaps you could blank it out with a cutout on the scanner? Otherwise, please do send this by email, for security sake, we can ensure the file will be handled with care, and deleted when the case is resolved.

Kind regards,

[xchangeit]

Hello, xchangeit

Thank you for the sample, could I also ask for an export of the tool in question? You can export a tool by right clicking on the tool itself through the PDF-Tools UI:

image.png

You may need to place it in a zip folder in order to upload it to the forums. Or you can send it to us by email <Support@PDF-XChange.com>, with a link to this forum post so we can connect the dots.

It would also help to have a screenshot of the preview image, before the issue appears in the file. Though I expect that may contain some sensitive information, perhaps you could blank it out with a cutout on the scanner? Otherwise, please do send this by email, for security sake, we can ensure the file will be handled with care, and deleted when the case is resolved.

Kind regards,

Why don't you just try a random Windows installation of XChange Editor 10 with a random piece of paper containing red text and do a random MRC compression with PDF Tools? If you do so, you will not find a setting which does NOT show this effect.

My suspicion is that the compression settings of MRC and JPEG are not adjustable at all. MRC is always on strongest compression and JPEG compression seems not to happen at all. Seems none of you tried this PDF tool before.

[Daniel - PDF-XChange]

Hello, xchangeit

I am asking you for this because I did test it with a random file. Even using multiple shades of red (and an orange, as your sample seems to be a very light red), yet experienced no adverse effects in the output, as compared to the black text. Today I did the same test with both "high" and "low" quality, and while the visual output is much the same, the file size and DPI is notably lower with the "low quality" option, so there is most definitely a difference there, even if its hard for the human eye to see.

We do our best to ensure that our products are well tested, but there are countless variables out there, not all of which can reliably be tested, as we simply do not have the resources to do so.

So please, help us to help you, and provide what little we ask for. It is not much, but reducing the amount of variables significantly speeds up the process of finding a solution. I cannot offer you our best if you will not give us what we need to work with.

Kind regards,

[xchangeit]

Hello Daniel,

please unzip and try this attached file. The .pdtex ending is not allowed in this forum.

PDF-Tools Exchange MRC.zip

zipped for forum reasons

(1.94 KiB) Downloaded 35 times

It comes from a standard Windows 10 64 bit 22H2 installation, from an XChange Editor 10.2 test installation, but it shows these effects, even with OCR switched off.

[Daniel - PDF-XChange]

Hello, xchangeit

Thank you for that, I am still seeing the same on this end, however, I will note that I am using build 10.3.1.387, not a 10.2 build, so you may want to update and see if that helps at all.

Beyond that, if you could also send a copy of the scanned original file as an image without any conversion to PDF, it may help us to reproduce it here. So far in testing, all of my files turn out like my image above.

Kind regards,

[xchangeit]

Hello, xchangeit

Thank you for that, I am still seeing the same on this end, however, I will note that I am using build 10.3.1.387, not a 10.2 build, so you may want to update and see if that helps at all.

Beyond that, if you could also send a copy of the scanned original file as an image without any conversion to PDF, it may help us to reproduce it here. So far in testing, all of my files turn out like my image above.

Kind regards,

As I wrote earlier, I also tried the 10.3.1.387 on another machine, but with the same result.
To avoid any misunderstanding: my process is scanning a printed paper as the first step, not only taking a file and compressing it.
So the attached file should be printed in colour and then the print be scanned with the defined tool.

[Daniel - PDF-XChange]

Hello, xchangeit

Thank you for that, printing this file out and testing with the same settings I was finally able to reproduce the issue. With that said, it seems that the root cause is simply the Scanner DPI prior to the MRC conversion. Scanning at a higher DPI (1200 in my test case) gave an acceptable result with both medium and high quality MRC compression. I have attached the "Medium" compression quality document from this test.

2024-08-302.pdf

(659.87 KiB) Downloaded 41 times

Low quality was a bit fragmented, so I will be bringing this to our Dev team for review, but I would be inclined to say that low quality output is expected when using low quality settings or input. Striking the correct balance is a manual process, and in this case, it seems you need a higher quality input, to account for the MRC compression settings.

Kind regards,

[xchangeit]

... it seems that the root cause is simply the Scanner DPI prior to the MRC conversion. Scanning at a higher DPI (1200 in my test case) gave an acceptable result with both medium and high quality MRC compression. ...
Low quality was a bit fragmented, so I will be bringing this to our Dev team for review, but I would be inclined to say that low quality output is expected when using low quality settings or input. ...

Hello Daniel,

I am really pleased you could reproduce my problem and I hope your developers will find the reason for it.

But seriously, scanning with 1200 dpi is no solution for this. It just minimizes the optical impression of that bug. Nowhere in document scanning they recommend a scanning resolution of more than 200-300 dpi for archiving documents. To be more precise, b/w scans perhaps need 300 dpi, since they have no graduations. 24 bit color scans only need 150-200 dpi to get a good OCR result.

A document size of 600 KB per page is not acceptable for archiving documents. With a simple JPG compression of a 300 dpi color document, you can achieve perhaps 200 KB or less with a quite good optical quality.

By the way, the bug is not only in PDF tools, but in your MRC compression in general. This is the result of a 300 dpi scan and saving it via optimized mode directly from XChange Editor, with MRC medium:

So I am really looking forward to the results of your developers.

[Daniel - PDF-XChange]

Hello, xchangeit

The team is now looking into this, I cannot promise any changes, but it will see a thorough investigation. for the time being, I hope my proposed workaround is helpful.

Kind regards,

[nehakakar]

One alternative could be to try out the JPEG compression settings again after updating to the latest version, as sometimes updates resolve these types of problems. You can also use online tool like JPEG Compressor that can help manage file sizes without losing image quality.

[Paul - PDF-XChange]