How can I get a French voice to correctly pronounce English words.

Question

Hello,
In my work, I create training materials for technical engineers.
It often happens that I need to have technical terms that are often English spoken by a foreign language voice. Some Microsoft Azure speech voices effortlessly switch between two languages in a sentence, for example Dutch to English. But sometimes it is also very difficult as for example with a Dutch sentence with some German words. Then the voice tries to pronounce the German words in Dutch.
The same happens when going from French to English words, that always goes wrong, all English words are pronounced in French and the pronunciation then doesn't make sense.

Questions and
How can I get a French voice to correctly pronounce English words in a simple way? For example, a French sentence containing three English words.

Is there any way in the SSML file that I can change language codes to make this happen? If so is there a code list I can use?

If someone knows, please let me know. Thanks in advance.

Translated with www.DeepL.com/Translator (free version)

Answer 1

@René Tersteeg Yes, this should be possible with SSML voice element where multiple voices in different/same languages can be used.

For example, I have used English and Hindi to test the scenario:

<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="en-IN">  
    <voice name="en-IN-NeerjaNeural">  
        This is how you say "Thank you very much" in hindi.  
    </voice>  
    <voice name="hi-IN-SwaraNeural">  
        आपका बहुत बहुत धन्यवाद  
    </voice>  
</speak>  

Perhaps, the easiest way to test this is from Azure Speech studio audio content creation tool. Just enter the sentences in the tool and select the text that you want to be output in English and select a voice. Similarly, select the text that is in French and select the voice that is required. Save the file and hear the audio before deciding on the voice that sounds right. You can later create your own SSML elements in your applications and directly call the speech API to get the speech output.

Same SSML from the studio using the audio content creation tool.

If an answer is helpful, please click on or upvote which might help other community members reading this thread.