• Replies 4 replies
  • Subscribers 140 subscribers
  • Views 2004 views
  • Users 0 members are here
Options
Related

Translating text within images?

Jessica H

Hello all,

Do any of you have a good solution for translating text within images?

We're in the process of transitioning to a Tridion Docs -> ManTra workflow, and we can't figure out the best way to translate text within screenshots.

Currently, we use FrameMaker, and we create text boxes on top of our images for each text element, which our translators can then pull into their translation tool and translate for all of our foreign languages (30+). But we can't figure out a way to do this easily in Tridion Docs. SDL suggested possibly SVG overlays, but we would love to hear a real-world example of how this might work.

As an example, here's a screenshot from one of our user guides. I've circled in red every instance of text that needs to be translated.

Screenshot of Trados Studio user guide with circled text elements. Status area shows 'Adult' and 'Non-Paced' with a warning 'Printer Door Open'. Parameter area displays vital signs. Message area has 'Sinus Rhythm' alert. Date and time are visible.

Thanks for any ideas or tips Slight smile

Generated Image Alt-Text
[edited by: Trados AI at 6:38 AM (GMT 0) on 5 Mar 2024]
emoji
Rate translation
×
Suggest better translation
×

  • Hi - I'm a little fascinated by your process. Why don't you (or your translators) just capture new, clean localized screenshots from the localized UI? That's what FICO (and I think most companies) do as far as I know. I don't know about SVG overlays, but it sounds like a good option for you if you plan to continue asking your translators to hack the screenshots. I really look forward to reading all the comments on this thread, by the way - I'm very curious how other companies are tackling this. It's definitely a big pain point for localization!
  • in reply to Elizabeth Gschwind

    We also capture screenshots of the localized UI. If the UI isn't translated, then the image in the documentation isn't translated, but we will include the UI commands and screen names translated in parenthesis. For example <uicontrol>File ( translation)</uicontrol>.

  • Thanks Elizabeth and Sue! Yes, I know it seems like a terribly clunky process Sweat smile Unfortunately, it comes down to time/schedule. The translations take a monumentally long time, and we do not have time at the end of the localization/translation process to then go in and take a bunch of new screenshots and update all the user guides. So, the translators are translating the UI strings and user guide text overlays simultaneously.

    I would love it if the schedule were adjusted to allow us time to take screenshots after the fact, but tech writers' wishes are pretty low on the totem pole Slight smile

      do you have anything to add?
  • in reply to Jessica H

    Hi, Jessica. Perhaps you could continue your graphic-building process outside of the CMS workflow, and add some automation to make it all work together. You mention that the translators are currently working with UI strings -- I assume that these are created by the developers and associated with some sort of corresponding data ID in their UI development framework. Could your tech doc team possibly build screen images that reference the same data ID for each label? There are ways to build data-driven images like this in Photoshop (and I assume probably other illustration software applications). For example, see:  https://helpx.adobe.com/photoshop/using/creating-data-driven-graphics.html. You could then automate the import of these data-driven images back to SDL using the web services API (or Powershell library that calls the same). I recognize that this would represent a fair amount of initial work, but once in place could potentially save a lot of time and cost because your translators would only be translating the UI strings once. One other note worth mentioning is that there are ways to also automate the export of your existing UI images and overlays out of FrameMaker using FrameScript or the FDK. Good luck!