Skip to content
SVSynthetic Voice Notes

Reading list

  1. What text-to-speech is and how it turns words into sound

    A plain-English look at how a text-to-speech system reads typed words aloud, and where the front-end has to guess, such as with numbers and abbreviations.

  2. Neural voices compared with older robotic voices

    Why some synthetic voices sound human and others sound robotic, with the main methods behind older and newer voices explained in plain words.

  3. SSML: the markup that controls pauses, speed and emphasis

    What SSML is, which tags set pauses, pitch, speed and spelling-out, and the mistakes that make a voice sound strange or fail altogether.

  4. Fixing names and terms a synthetic voice gets wrong

    Practical ways to stop a text-to-speech voice mangling your business name, product names and acronyms, from respelling to SSML to lexicons.

  5. Voice cloning explained: what it is and what it needs

    What voice cloning means, the difference between quick and professional clones, and the recorded permission makers ask for before they build one.

  6. Consent and licence terms for synthetic voices

    What to check before using a synthetic voice in videos, adverts and phone messages. Who agreed, what the licence allows and where official rules apply.

  7. Audio formats, sample rates and file sizes from TTS tools

    Which audio file types text-to-speech tools produce, what sample rate and compression mean, and how to match a file to a phone system or editor.