Method (400) is to receive text input data (302) for conversion to synthetic speech (160) and to detect the presence of a hot word (130) assigned to the user device (110). It involves using a trained hot word recognition model (320) to determine whether the pronunciation of text input data contains hot words. Hotwords are configured to initiate a wakeup process on the user device to process one or more other words that follow the hotword and / or the hotword in the audio input data. When the pronunciation of the text input data contains hot words, the method also generates an audio output signal (304) from the text input data and provides the audio output signal to the audio output device (118) for audio output. Includes outputting a signal. The audio output signal is configured to prevent the wakeup process on the user device from starting when captured by the user device's audio capture device.
展开▼
