Question about TTS voice laughter, + symbols breaking SpeechSynthesizer subtitles

Copper Contributor

When the SaraNeural TTS reads "haha" it makes a laughing sound. However, it is not pitched up accordingly - I pitch the voice up by 25% in SSML, but the laugh sounds like the default pitch. This makes it sound jarring and out of place.


A solution I'm looking for would be to either pitch up the laughter, or for the laughter to instead just be read as "haha". I assume this issue is present with other sounds such as groans that the TTS may attempt to do.



As for my second problem, the tts reads text with emojis such as :smiling_cat_face_with_heart_eyes: correctly, however the SpeechSynthesizer returns completely broken word-for-word subtitles for the text.



Thanks for any help!







0 Replies