Strange text conversion for text to speech

Hello Heroes!

We have a strange problem with how Storyline converts Swedish when publishing. We don't want to create mp3 files for every slide in the course we use JavaScript and the 'SpeechSynthesisUtterance' function. It works fine for almost everything. But when we grab the text from the html output, some letters or combinations of letters are converted to symbols. For example, the small letter 'f' is converted to a hollow square. The letters 'fi' is converted to 'î€'. 

The text is shown correctly when looking in the course but in the output,  it is converted.

Can we turn this feature off in any way? Or do You have a complete list of all the letters and combination of letters we need to swap before reading the text?

I have attached a *.story fil with the problem. Click the "ear" in the bottom of the slide. At least in Edge and Swedish language the problem appears.