Forum Discussion
How to get Text to Speech to pronounce phonetics and all the individual letters sounds
Hello,
I am developing a course for English learners and I need some way of getting the text to speech to pronounce phonetics that are taught in phonics. By this, I mean short vowels such as ă, ĕ, ĭ, ŏ, ŭ, combined vowels like 'ou' and consonant sounds such as sh, ch, th, ow, ck as well as all the individual alphabet letter sounds. Is this possible, and if so, what do I type?
Thanks,
Joshua
13 Replies
Hi Joshua,
The Text-to-Speech feature works best with full words, but it seems to also recognize partial words when a vowel is involved.
For example, ă, ĕ, ĭ, ŏ, ŭ could be "ah, eh, ih, oh, ooh." However, "sh" doesn't work well because there's no vowel in the phonetic spelling.
Have you considered recording a human voice pronouncing these phonetics using the narration tool?
Hello Alyssa,
Thank you for your suggestion. Yes. I have considered it, but it defeats the purpose of getting the computer to say things for you. Also, it gets pretty tedious when I have to do the recording separately and edit it so that they connect with the other voices.
i tried to put Ahh for a long a sound and it doesn't work -- is there a way to have the letter A and it sounds like way without the w? I tried Ay and i says I.
Hi Christine,
Thanks for reaching out! It sounds like you're also using Text-to-Speech and are running into issues getting certain sounds with the letter "A."
What language are you using? If you're using English (US), I had no issues with using the letter "A" or "Ay" to sound like the ending in way using Joanna's Voice.
Using "Ahh," I noticed that the sound was short, like "uh." Perhaps you can try "Auh" and see if that works for you!
Hello, can you please suggest: How to write for ssml markup for short and long vowel sounds within words. For example, a short 'i' sound. . .to pronounce the 'i' as 'y'? If we write for phonemic without SSML markup - the 'lyve' spelling in text to speech appears in closed captioning. Using English (US) neural voices.
Hi Lori,
Happy to help!
You can check out this article to learn how to distinguish between homonyms using SSML tags. The article also contains some examples, so feel free to substitute the word "lead" with the word "live" and you should get similar results.
I was finally able to get Joanna to say "A" by writing it as "eigh"
Thanks for sharing your findings with the community, Kristen! 🎉
How about the word Dose...Jessica keeps saying dos.
Hi Jacob,
For the word 'dose', try using 'dows' instead. I tried on my end and the resulting audio is pretty close to the actual pronunciation of the word.
Hope this helps!
I think using
Speech Synthesis Markup Language (SSML)
in text to speech can help I tried some of it and it pronounced some phonemes but it much more patience and time to get a good final result if anybody interested tell me and I will send you what I did
For some reason SSML is not available for all languages available (seems like a beta release).
