Text-to-Speech (TTS)

Text-to-speech (TTS), or speech synthesis, is the technology that turns text into spoken audio. Older generations sounded robotic because they stitched together recorded fragments; current systems use neural networks that generate the sound wave from scratch, learning intonation, pauses, and emotion from thousands of hours of real speech. In the best tools, the result is hard to distinguish from a human narrator.

Features that set modern platforms apart:

• Voice and language variety, including regional accents.
• Voice cloning: creating a synthetic voice from samples of a real one, useful for consistent narration, and ethically sensitive (it requires the consent of the person being cloned).
• Style control: adjusting speed, emphasis, pauses, and emotion.
• API access: generating audio programmatically, to give chatbots and assistants a voice.

Uses range from accessibility (screen readers, audio description) to content production: video narration, audiobooks, podcasts, courses, and automated phone systems. It is the reverse path of *speech-to-text* (transcription).

Concrete example: a content creator writes the script for an educational video, pastes the text into a tool like ElevenLabs or Murf, picks a voice, and downloads the finished narration in minutes: no studio or microphone required.