how Hawking controlled his voice

In his later years, Stephen Hawking selected letters on a screen using a sensor that responded to movement in a cheek muscle; the computer sent his completed sentences to a speech synthesizer. Around 2011, his text-entry rate had fallen to roughly one or two words per minute. This did not mean the synthesized voice spoke only one word per minute: it voiced a sentence after he had composed it. A 2014 software upgrade made input easier and improved his communication rate. The illustration is AI-generated and is not a photograph of Hawking’s equipment.