It Wasn't Hatsune Miku: the Untold Story of the Real First Vocaloid
By James H. Sterling
The core technology originated in March 2000 at the Music Technology Group of Pompeu Fabra University in Barcelona, Spain. Working in close partnership with Yamaha Corporation, a research group spearheaded by Japanese engineer Hideki Kenmochi set out to solve one of audio engineering's hardest problems: concatenative vocal synthesis. Synthesizing spoken words was already possible, but singing demanded pitch control, emotional resonance, vibrato, and seamless transitions between consonants and vowels.
Yamaha codenamed this experimental initiative Daisy. The moniker paid direct homage to "Daisy Bell", the 1892 song that researcher John Kelly and speech synthesis pioneer Max Mathews programmed an IBM 7094 computer to sing at Bell Labs in 1961. Kenmochi spent three years refining the algorithmic backend. His system sliced recorded human vocals into individual phonemes, diphones and triphones, and stored them in a structured library. When a composer entered notes on a digital piano roll and typed English lyrics, the software stitched those microscopic sound fragments together in real time.
By early 2003, Yamaha had a functional, ground-breaking program: the Vocaloid 1 engine. However, Yamaha made a strategic corporate decision that dictated the next two decades of digital music. Rather than marketing and packaging end-user voice software directly, Yamaha chose to license the core engine out to third-party sample developers. The company would provide the software shell, while independent music companies would hire vocal talent, record phoneme libraries, and sell the resulting commercial voicebanks.
James Sterling reports on renewable energy developments, climate policy, ecological conservation, and green tech innovations around the globe.
