This chapter opens with a case study of a 1979 Japanese song featuring vocoder, illustrating how robotic singing voices have been perceived and aestheticized. It also briefly highlights creative pioneers like Wendy Carlos, who worked with vocoders, and introduces similar vocal effects such as the Sonovox. The chapter examines the historical significance of Homer Dudley’s groundbreaking inventions—the vocoder and the voder—as an innovation in signal processing and a commercially nonviable product, respectively. Dudley’s approach to signal processing was built upon Helmholtz’ perceptual theories, laying the groundwork for spectral analysis. The chapter also introduces the inception of the source-filter model of speech, a theoretical model that proposes vocalization is generated by the vocal cords (the source) and then modified by the vocal tract (the filter). This process was often modeled using electrical circuits, which contributed to the development of spectral analysis methods for speech. The influence on singing synthesis of 1952 Gunnar Fant’s OVE formant synthesizer is traced through the development of Music and Singing Synthesis Equipment (MUSSE). Fales’ account of the role of formants in aural perception is discussed, followed by an overview of Mori’s influential concept of the uncanny valley and its applications in both audio and audiovisual media. The chapter concludes with two contrasting case studies of film music, illustrating how context and expectation influence uncanny valley effects in mediated experiences. One case examines how altering formants using Vocaloid enhanced uncanny valley-type effects in an animated film by Satoshi Kon, while the other looks at how the unremarkably human-sounding voice of HAL9000 in 2001: A Space Odyssey was made eerily unsettling.