how did stephen hawking talk

Stephen Hawking spoke using a computer-based speech system that he controlled with tiny movements of his cheek muscle, which selected letters and words on a screen and then sent them to a voice synthesizer that produced his famous robotic voice.

How his system basically worked

• Hawking had amyotrophic lateral sclerosis (ALS), which gradually paralyzed almost all voluntary muscles, including those needed for normal speech.

• He communicated through a specialized computer mounted on his wheelchair, originally controlled by a hand switch and later by a single cheek muscle.

• The computer ran software that showed letters, words, and commands on a screen; he would “click” his cheek at the right moment to choose what he wanted.

• Once he built a sentence, the computer passed it to a speech synthesizer that turned the text into his distinctive electronic voice.

The cheek sensor and cursor system

• As his hand movement declined, engineers added an infrared or pressure-based sensor that detected tiny twitches in one cheek muscle.

• A cursor would scan through options (letters, word groups, or menu entries), and a cheek twitch at the right time would stop the cursor and select that item.

• This method is naturally slow, so the system relied heavily on prediction and shortcuts to keep his effective speaking rate usable.

Word prediction and speed

• Over time, companies like Intel and SwiftKey helped upgrade his system with powerful predictive text, trained on his emails, lectures, and books.

• The software learned his style, so he often only needed to type a few letters before the full word or even the next likely word appeared.

• With prediction and custom shortcuts, he could reach up to around 15 words per minute in optimal conditions; without prediction, he could drop to about 1 word per minute.

Why his voice sounded “robotic”

• Hawking’s voice came from a hardware speech synthesizer (famously from a company called Speech Plus) that used an older, mechanical-sounding synthesis method.

• Later, more natural voices became available, but he deliberately kept the original synthetic voice because it had become part of his identity and how the world recognized him.

• That same voice was used for decades in his lectures, documentaries, and interviews, making it one of the most recognizable computer voices in the world.

How he did interviews and lectures

• Because generating sentences was slow, Hawking usually prepared answers and full lectures in advance and saved them on his computer.

• For TV or public appearances, producers often sent questions ahead of time so he could compose responses before recording.

• During a talk, he could send pre-written sentences to the synthesizer one by one, giving the impression of a more natural, continuous presentation.

TL;DR: Hawking talked by twitching a single cheek muscle to control a computer, which used predictive text to build sentences and then fed them into a speech synthesizer that produced his iconic robotic-sounding voice.

Information gathered from public forums or data available on the internet and portrayed here.