Text to speech vs speech to text: what's the difference?

Text to speech (TTS) turns written text into spoken audio: you type or select words, and a computer voice reads them aloud. Speech to text (STT) does the opposite: it listens to spoken words and writes them down. In short, TTS gives text a voice, and STT turns a voice into text. Both are everyday features on phones and computers, and both are important accessibility tools.

What is text to speech?

Text to speech, also called speech synthesis, takes text and produces a voice reading it. You meet it when a sat nav reads out directions, when a screen reader describes what’s on a phone screen, or when an app reads an article aloud.

For people who can’t speak, or can’t speak reliably, TTS is a form of AAC (augmentative and alternative communication). They type what they want to say and the device speaks it for them. That’s what MyVoice does: you type, tap Speak, and your words are read aloud in a natural voice. Our guide to what AAC is explains the wider picture.

How text to speech works

In plain terms, a modern TTS system does two jobs:

1. Works out how the text should sound. It expands numbers and abbreviations (“Dr” becomes “doctor” or “drive”), works out pronunciation, and decides on rhythm and intonation, such as rising at the end of a question.
2. Generates the audio. Neural networks, trained on many hours of recorded human speech, produce a voice with natural-sounding pacing and emphasis.

Older systems stitched together fragments of recorded speech, which is why they sounded choppy. Today’s neural voices are far smoother, and many are hard to tell apart from a person.

What is speech to text?

Speech to text, also called speech recognition, voice typing or dictation, listens to speech and writes it down. You use it when you dictate a message, ask a voice assistant a question, or watch automatic captions on a video.

How speech to text works

A speech recognition system also does two main jobs:

1. Recognises the sounds. It breaks the audio into tiny slices and works out which speech sounds are most likely being made.
2. Picks the most likely words. A language model uses context to choose between similar-sounding options. “I walked around the block” is more likely than “I walked around the clock”, so that’s what it writes.

Accuracy depends on audio quality, background noise, accent and how clearly someone speaks. Modern systems handle a wide range of voices well, but they can still struggle with some speech differences, which matters for people whose speech is affected by a health condition.

TTS vs STT: side by side

Common uses of text to speech

• Speaking for people who can’t. AAC users, people with motor neurone disease, people recovering from a stroke, and people with temporary voice loss use TTS to talk. If you’ve lost your voice, our guide on how to communicate without speaking has practical tips.
• Reading support. Hearing text read aloud while following along helps many people with dyslexia and other reading difficulties.
• Screen readers. Blind and partially sighted people use screen readers such as TalkBack on Android and VoiceOver on iPhone, which rely on TTS.
• Listening on the go. Articles, documents and emails can be listened to while walking or commuting.
• Content creation. Creators use TTS for voiceovers and narration.

Common uses of speech to text

• Captions for people with hearing loss. Live captioning turns speech into on-screen text in conversations, calls and videos.
• Hands-free typing. People with arthritis, repetitive strain injury, paralysis or limited hand movement can write by speaking.
• Note-taking and transcription. Meetings, lectures and interviews can be transcribed automatically.
• Voice control. Voice assistants and voice commands rely on STT to understand what you’ve said.

Who benefits from each?

The simplest way to think about it: TTS helps people whose challenge is speaking or reading, and STT helps people whose challenge is hearing or typing.

Someone with a voice condition may type and use TTS to speak. Someone who is deaf may use STT captions to follow what others say. Someone with limited hand movement may use STT to write and TTS to listen back. Many people use both at different times.

If you’re exploring either for a health reason, a speech and language therapist, audiologist or occupational therapist can help you find the right setup. This article is general information, not medical advice.

Where text to speech and speech to text work together

Some of the most useful tools combine both directions:

• Live captions with typed replies. On recent Pixel phones, Live Caption can caption a phone call and let you type a response that is read aloud to the caller.
• Live Speech on Apple devices. Apple’s Live Speech lets people type what they want to say and have it spoken during calls and in person, and Live Captions can transcribe what others say.
• Conversation apps. Some assistive apps are starting to combine listening and speaking, so a person who can’t speak can see what the other person said and reply by text to speech. We wrote about this in exploring assistive text to speech in 2026.
• Translation. Translation apps listen in one language (STT), translate, then speak the result in another (TTS).

FAQs

Is text to speech the same as speech to text?

No. Text to speech reads text aloud. Speech to text writes down what is said. They’re opposite directions of the same idea.

What is TTS used for?

Mostly reading text aloud, speaking for people who can’t speak, supporting reading, powering screen readers and creating voiceovers.

Is dictation text to speech or speech to text?

Dictation is speech to text. You speak, and your words appear as text.

Can one app do both?

Yes. Phones have both built in, and some apps combine them. Most apps still focus on one direction, though. MyVoice is a text to speech app focused on speaking clearly in a natural voice. If you’d like to try TTS on your phone, our guide to using text to speech on iPhone and Android is a good place to start, and you can hear our voices on the voice sampler.

← Back to blog