The quality of the input decides the quality of the output. These simple practices make the voice clearer and steadier, in Text to Speech and in Audio Studio alike.
Write what you want to be heard
Numbers and symbols are read more reliably when written out:
- Instead of Dr. write Doctor.
- Instead of 100% write one hundred percent.
- Instead of moknah.io/docs write moknah dot io slash docs, or whatever you want spoken.
Sentence and line length
- Give the voice at least a sentence or two; very short texts can sound uneven.
- In Audio Studio keep lines between 3 and 20 words for the best results.
- Paste a whole text and every paragraph becomes its own line.
For Arabic in particular
- For Standard Arabic choose Arabic Tashkeel so every word is pronounced correctly.
- For dialects, write in your dialect, choose a voice that speaks it and keep Basic processing.
- Add diacritics yourself to an unclear word, and write numbers and abbreviations as words.
Common problems
- The audio sounds flat: raise Temperature a little, use an emotion mode with a tag such as [calm], or pick a voice that fits the content.
- The result is unstable: lower Temperature and Expressiveness, and give the voice a full sentence or two.
- A word is pronounced wrong: try Arabic Tashkeel, add diacritics, or try another voice.
- The same text sounds a little different each time: that is expected, because the voice performs the text fresh each time; lower Temperature for more consistency.
More in the FAQ and troubleshooting and the Text to Speech documentation.