Antonella Pecchioli 

Unfortunately PDF files are just tricky!  We also had problems with IRIS technically, it was a 3rd party technology we paid for and it didn't always work well.  So in 2022 you'll see we removed the PDF filetype altogether that worked with IRIS and have replaced it with something new.  Unfortunately the new filetype doesn't OCR at all, but we did create a free plugin for 2022 that does a reasonable job most of the time.  You can read about it here:

 PDF Assistant for Trados 

The principle with this is to use the app to convert the file to a DOCX first. Then you can tidy up the PDF as needed, and I recommend TransTools to help with this:

https://www.translatortools.net/products/transtools

Ideally, I think if you get a lot of PDF files to work on, it may be worth investing in something like Adobe Acrobat ( https://www.adobe.com/acrobat.html ) or Abbyy FineReader ( https://pdf.abbyy.com/ ) as these tools provide better control over the editing of PDF files so that you can prepare them properly for translation first.

Thank you Paul for your clear answer. Do you think that the app IRIS can create troubles (as from its installation appear error messages...) and it's better if I remove it? 

Antonella Pecchioli 

Antonella Pecchioli said:

Do you think that the app IRIS can create troubles (as from its installation appear error messages...) and it's better if I remove it? 

I wouldn't be 100% certain, but I have seen problems in the past with OCR.  Although the PDF itself can also be the cause of how successful you would be with this tool.  They are simply not designed to be translated... it's always better to get the original source.  I know that's not always possible so all CAT tools have tried to do something to address this... some try to retain formatting, some just extract text etc.  But in the end my own opinion is I would always take a tool that that designed to handle PDFs and create a file that can be properly translated first.

In my own life I actually use two free tools for managing this stuff.

• NAPS - Not Another PDF Scanner to scan in documents I get as hard copy
• FreeOCR - a small tool that uses Google Tesseract and has support for OCR'ing many languages

Not even close to the capabilities offered by some of the commercial tools but they suit my needs and it's often a lot faster to simple get the text, remove the line breaks... all of which FreeOCR will support quite nicely... and then just recreate the layout in Word myself.  Normally doesn't take long and the effort at the start means the translation process is not only problem free, but the final file is perfect.  So effort at the start means no effort needed at the end.

Dear Paul, 

in line with the topic, I have several questions:

1. So there is no point in installing IRIS OCR on Trados 22 SR1, as far as I can understand? Actually, in my case, the option does not even appear in the PDF settings window (does anyone else have this same problem? I restarted Trados after the installation , but still it wouldn´t show up).

2. I see, that you are recommending commercial programs like Adobe Acrobat and FineReader, however, Adobe Pro´s results are not extremely good:

For example, a desktop screenshot is recognized as follows:

And even to obtain this result, we need to perform file conversion outside Trados...

Then, according to Google, PDFelement is the best OCR program nowdays. What is your opinion on it? Would you recommend investing in it?

3. As for the FreeOCR you mentioned, it is 100% online, isn´t it? (blocked on the company´s equipment due to the permissions issue). Is the tool below the one you referred to:

Thanks in advance!

 Yulia Kovalskaya 

I think we all have questions about PDF files!!!

Yulia Kovalskaya said:

So there is no point in installing IRIS OCR on Trados 22 SR1, as far as I can understand?

Correct... no point.

Yulia Kovalskaya said:

Actually, in my case, the option does not even appear in the PDF settings window (does anyone else have this same problem?

Everyone!  This is because the new filetype doesn't use IRIS and we replaced the old one.

Yulia Kovalskaya said:

Then, according to Google, PDFelement is the best OCR program nowdays. What is your opinion on it? Would you recommend investing in it?

I'm not familiar with it.  Even though this thread is only a few months old I actually added to my choices in my personal life.  I still use NAPS and sometimes FreeOCR, but I now choose two other tools:

1. PDF-XChange Editor - https://pdf-xchange.eu/pdf-xchange-editor/ 
   Mainly because I wanted a good PDF editing capability.  I haven't played with the OCR yet so I took a screenshot of your image (that already was a poor resolution), saved as a PDF and tried it - 
   Also not perfect and you'd still need to do some work removing superflous tags before processing but I was quite impressed with this quick test.  It also has a lot of other very nice features.  I use Adobe at work on my work laptop but I found that too pricey for my needs at home.  PDF-XChange Editor is a lot cheaper, it's perpetual, and I actually find it a lot more functional and user friendly than Adobe too.
2. ChatGPT
   for many things I receive where I want the text it's brilliant for screenshotting the image and asking for the text, and the translation ;-)

Yulia Kovalskaya said:

As for the FreeOCR you mentioned, it is 100% online, isn´t it?

Nope... it's a desktop application: http://www.paperfile.net/

But note it's incredibly basic.  I like it because it's simple and fast and often served my purpose perfectly.  No substitute for a more professional tool though.

Ultimately there are no easy answers for PDFs.  All PDF files are different and depending on how they were created, how good the quality is, what the content is, what language the content is, the less consistent your results.  So it's often a case of knowing your tools and finding the most appropriate for the job.  Or encourage your clients to stop providing PDF files as much as possible!!

Paul said:

PDF-XChange Editor - https://pdf-xchange.eu/pdf-xchange-editor/ 
Mainly because I wanted a good PDF editing capability.  I haven't played with the OCR yet so I took a screenshot of your image (that already was a poor resolution), saved as a PDF and tried it - 
Also not perfect and you'd still need to do some work removing superflous tags before processing but I was quite impressed with this quick test.  It also has a lot of other very nice features.  I use Adobe at work on my work laptop but I found that too pricey for my needs at home.  PDF-XChange Editor is a lot cheaper, it's perpetual, and I actually find it a lot more functional and user friendly than Adobe too.

There are many options here, could you , please , confirm, which one I am to use... I think the first one. Nice that they offer this trial version (and the license price is also OK).

As for the GPTchat, I did not get it...how do you use it in the scanned docs recognition process?

Once again, many thanks for sharing your expertise, Paul!

Yulia Kovalskaya 

Yulia Kovalskaya said:

There are many options here, could you , please , confirm, which one I am to use... I think the first one.

I think so too.  If you want OCR you need at least the Plus version: https://pdf-xchange.eu/feature-overview.htm

Yulia Kovalskaya said:

As for the GPTchat, I did not get it...how do you use it in the scanned docs recognition process?

Two ways...

1. copy paste the image, or load the file, into the ChatGPT UI and use a prompt to tell it what you want to do.  I use this for small usecases where I just want the text... not going to be useful for large files or where I want formatting retained.
2. via the API.  But you need to figure that one out!

Interestingly we have just developed a solution here in the community using OpenAI API image recognition capability that will go live very soon that automatically takes the images that are pasted into the forum and creates the Alt-Text based on the image and the context in the thread.  It can provide a pretty good description of the image like this and this is important for blind users who work with screen readers since images are problematic unless users always fill in the Alt-Text when they post.... and I can tell you they don't!

Yulia Kovalskaya said:

many thanks for sharing your expertis

You're welcome, but don't confuse me telling you what I use outside of work as I don't translate or handle documents in the way you probably need in a professional setting.  So whilst PDF-XChange is a pretty good and professional tool, the rest of the stuff I mentioned that I use may not be appropriate at all.  So please don't shoot me if you test and find it's not helpful at all!!

Paul said:

automatically takes the images that are pasted into the forum and creates the Alt-Text based on the image and the context in the thread.  It can provide a pretty good description of the image like this and this is important for blind users who work with screen readers

Ohhh, sounds cool, however, I can hardly imagine the description of a Trados editor tab screenshot, for example....not THAT imaginative I am, will love to test it when available!!!!

Paul said:

So please don't shoot me if you test and find it's not helpful at all!!

No way, would be such a light way to get rid of me. I prefer to keep torturing you with my questions!!!

Loads of thanks, Paul, for all your help! We´ll try the OCR options you recommend!

Dear All, 

here am I again, the beginner of this thread, still with some questions on the topic... I have another PDF that doesn't charge in TRADOS and I still don't know what to do.

I have read the other questions and replies put by Yulia and Paul and downloaded NAPS2. Paul could you please tell me again what NAPS does? I mean, how NAPS can help me with my PDF that TRADOS doesn't read. I attach a screenshot. 

Could I please have a video demonstration?

Thanks a lot, 

Antonella 

I also attach a short video of the problem I am facing with the PDF. Thanks!