For a more natural Text to speech output, it is necessary to add additional commas so that there are micropauses and the text sounds more like a human voice. Grammatical commas are already in the text. Additional commas are needed.
My text to speech engine (Microsoft Speech studio, Azure) makes perfect voiceover when I add commas to my texts manually.
I want to automate this task, for example using ChatGPT or another way. But so far it’s not working. If you have solved a similar problem, share your experience.
I'm using the free web version of ChatGPT-3.5. If I find a solution, I will use the API in the future.
What I tried to do:
1. I gave ChatGPT a set of my sentences (about 300), the original version and the version with additional commas, which I placed manually. I asked to find my pattern (set of rules) of how I add commas. This works well, but only once. Immediately after this ChatGPT adds commas well. But after that the second text can no longer be processed well. Apparently he “forgets” the rules he finds.
2. I asked ChatGPT to give me the found rules explicitly. And I added rules to each prompt, before the text. Works poorly, commas are in the wrong places as a result.
3. I tried to make up rules for additional commas myself. It hasn't worked out yet. There are either too few or too many commas.
4. I asked ChatGPT how to correctly compose a prompt for ChatGPT to place additional commas for text to speech. The essence of the answers: "...keep experimenting."
What kind of help do I expect:
Maybe someone has a ready-made prompt for this task. Or a set of rules for commas TTS for the English language.
Maybe there are a Python scripts (or in another language) for placing commas for tts. Or ready-made applications or services for this task.