Character encoding error when using Benghali language fonts and the composers

[Alan23548972pnq5]

I am trying to create a document in Benghali that needs to be accessible 
 
When recieving the translation, I copy into indesign and export the PDF (no design just as plain text to get this working!)
using the standard composer it corruupts the formatting of the letters (in in design) however it passes accessibility in the PDF.

When i use the Adobe world ready composer it corrects the formatting issue in indesign but when it fails character encoding in the PDF

Ive tried this with 4 fonts  Noto Sans Benghali, Adobe Benghali, Myriad Benghali and Tiro Bangla all have the same issues and am really stuck with a deadline looming 

Any Suggestions?

[Abhishek Rao]

Hi @Alan23548972pnq5,  

Thanks for sharing the details. I can imagine how frustrating this must be. Could you please confirm the version of InDesign you're using and the operating system details? Also, let me know if the issue occurs with all your Bengali files or just a specific one. If possible, a small sample INDD file and a quick screen recording showing the encoding issue would really help me test it on my end and discuss it with the team.  

Looking forward to your update.  
Abhishek

[Alan23548972pnq5]

Hi Abhishek 

im using InDesign 2025 i think its 20.3 x 64 this is from the about Indesign... in the help section 

Windows Version :

This issue is with all the files im trying to create 
Ive created a 2 files with 2 different composer - one standard and one World ready - i then generate the PDF's from these - ill attach them here 
the files have 4 versions of the same copy all in different fonts and there are 2 versions the only difference is the composer 

ive attched a screen grab of the error on the world ready composer file being tested in acrobat for accessibility 

in this file the words are correct but fail accessibility in the other PDF (attached) it passes accessibility but it has changed / corrupted the formatting of the words so it doesnt read correctly 

Thanks for your support - hopefully we can find a fix 🙂

Alan

[Anubhav M]

Hello @Alan23548972pnq5,

Would you mind trying the suggestions I shared in this community post (https://adobe.ly/3Yozt3u) to see if they help with your workflow?

Looking forward to hearing from you.

Anubhav

[Alan23548972pnq5]

Hi Anubhav 
i have followed the suggestion but it didnt make a difference - still having the error!

[Abhishek Rao]

Hi @Alan23548972pnq5,

Thanks again for sharing the files. I tested with your files, and I could see that the character encoding error appears in Acrobat when using the World-Ready Composer, even though the text looks visually correct in InDesign.

Since this seems specific to how Acrobat is handling the exported Bengali text, I'd recommend posting this in the Acrobat Community as well for further insights:
https://adobe.ly/4lVY4GY

In parallel, I've raised this with the InDesign product team and will keep you posted as soon as I hear back.

Thanks for your patience, and I'll follow up soon with any updates!

Abhishek

[Alan23548972pnq5]

this is also in the Acrobat community channel as well - no joy as yet! 

[Joel Cherney]

I tried a few different ideas before I was able to export a PDF of your Bengali text without this character encoding error. The one that worked, in the end, was turning off the HarfBuzz character shaping engine, by going into the Advanced Type section of the Preferences and checking the "Enable Legacy Character Shaping Engine." I've attached a PDF that I made from your file (Benghali test file world ready composor.indd) that passes that part of the Accessibility Check in Acrobat. 

[Alan23548972pnq5]

Hi Joel 
Thanks you so much for trying this - i followed your instructions but when exporting i still get the error - could you send me the indesign file that you managed to output this from so i can compare other settings to see if this is in combinationwith something else?

thanks
Alan

[Joel Cherney · Accepted solution]

No problem! Let me tell you what I tried. I might be remembering some of the details incorrectly, but here's basically what I did:

1) First I looked at your INDD and saw that it really wasn't set up for accessibility. So I first made four paragraph styles, one for each of  your test fonts. I made sure that each one was marked as "Bangla (India)" and that each paragraph was mapped to the P tag in the Export Tagging.

2) Then, at export, I chose "Interactive PDF" and chose "Bangla (India)" in the Language dropdown in the Advanced section of the "Export to Interactive PDF" dialog.

3) That didn't work, so I made more coffee. While doing that, I realilzed that almost every time I've had a problem with any kind of complex script encoding recently, it's been the fault of HarfBuzz. So I went back to my INDD, turned on the Legacy Character Shaping preference, and exported a PDF. That worked fine with no character encoding failure.

4) I then said to myself, "I wonder if it's all HarfBuzz' fault." So I went back to your original INDD, turned on the Legacy Character Shaping pref, made no other changes that I recall, and exported a fresh PDF, with the language "Bangla (India)" selected in the  "Export to Interactive PDF," which I then attached to my last post, here. So I decided "Oh, all that other accessibility work was pointless, Alan just needs to turn on the Legacy Character Shaping Engine." At that time I didn't remember that I'd left the language setting in the  "Export to Interactive PDF" dialog, so that may well be the missing piece you'd need to make this work. Honestly, it might be worth testing with the language set in the "Export to Interactive PDF" dialog, but without the Legacy Character Shaping Engine, to see if it really is just declaring the language that is resolving the issue. 

That being said, I encourage you to not bother to try testing accessibility in PDF output from InDesign without doing all of the prepwork. Something that passes muster in Acrobat might fail when you run it through CommonLook or PAC 2024. If you're only using Acrobat to test, then something that passes muster in Acrobat might fail when it makes it to the end user. So: I'd suggest that you make a full set of paragraph styles, declare language everywhere you can, and maybe map your styles directly to H1 and P tags in the Export Tagging section (I don't know how well the Automatic setting works, I've just been specifying them manually). 

[Alan23548972pnq5]

Joel 

this worked it was the languaf=ge setup in the styles that i hadnt completed! 

thanks for your help on this - if i can ever return the efforts - let me know! 

Alan