A MEMS speaker that includes a control unit and multiple MEMS elements that include a membrane positioned in a first plane, a blind that is positioned in a second plane and a shutter that is positioned in a third plane. The control unit is configured to control the multiple MEMS elements to operate in an audio output mode or an ultrasonic output mode. The MEMS speaker is configured, when operating in the ultrasonic output mode, to oscillate at least one of the membrane, blind and shutter thereby generating an ultrasonic signal without audio-modulating the ultrasonic signal. The MEMS speaker is configured, when operating in the audio output mode, to oscillate the membrane thereby generating the ultrasonic signal and oscillate at least one of the shutter and the blind thereby modulating the ultrasonic acoustic signal to generate an audio signal.