【1961 | IBM 7094 : First Computer To Sing】 Daisy Bell

Описание к видео 【1961 | IBM 7094 : First Computer To Sing】 Daisy Bell

http://VOCALOID.wiki (By VocaloP / ボカロP別)

BACKGROUND:
"Daisy Bell" was composed by Harry Dacre in 1892. In 1961, the IBM 7094 became the first computer to sing, singing the song Daisy Bell. Vocals were programmed by John Kelly and Carol Lockbaum and the accompaniment was programmed by Max Mathews.

This performance was the inspiration for the famous scene in the 1968 film 2001: A Space Odyssey in which the HAL 9000 computer sings the song as it is deactivated.


FURTHER READING:
YAMAHA's "Daisy Project":
While it did not become a product, the "Elvis Project" helped establish that a series of phonetics in a wide range of pitches would help build a synthesizer based on any model. YAMAHA agreed to help them start a fresh new project; it was at this point that Kenmochi Hideki joined. The first initial ideas came from him in Japan in 2000, with most of the research done at the Pompeu Fabra University and the development of the core signal processing libraries created in C++. YAMAHA itself was responsible for the product design and development of the actual product. It was pure collaborative research, and they did not think about selling at that time.

At the time, synthesizers would take days to produce good quality results, but the vocal would always sound inhuman and obviously generated by a machine or computer. The price was expensive as well. This meant that while all other parts of the music production were by then fully able to be recreated in a DAW, producing a good quality vocal performance meant hiring a human vocalist. So, the aim of the project was to provide a fast, low cost way of getting uncanny human-like vocals to give producers full control of music production. They used "Elvis" as the base model for ideas and set about to tackle two main problems:

how to process and transform singer recordings so that it would result in a performance of a given song sounding as natural as possible and provide the feeling of a continuous flow
how to process and transform the singer recordings so that it would result in a performance of a given song sounding as natural as possible and provide the feeling of a continuous flow

The VOCALOID™ project was originally codenamed "Daisy Project" ("DAISYプロジェクト" or "でいじぃぷろじぇくと"), a name taken from the song "Daisy Bell" and was at a prototype stage in March 2002. (EpR ) was developed as the first voice model and it allow the researchers to transform vocal timbres in a natural manner while preserving subtle detail. At first, "Daisy" could only say vowels like "ai (love)". Four months later, "Daisy" began to support consonants, with the first "complete word" being "asa" (morning).

Because YAMAHA itself could only provide limited vocals, they licensed the software out to various 3rd party studios. The first studio to join this project was Crypton Future Media, who were contacted in May 2002. YAMAHA then attempted to find English studios to support an English version, but the majority of responses to contact were negative. The first studio to enter development was Zero-G, joining in the fall of 2002, with PowerFX also joining that year. Thus, both English and Japanese voicebanks began development.

At the 6th anniversary of VOCALOID™, Hiroyuki Itoh noted that they received demos from Zero-G without warning of what seemed to be a male vocal singing. Since they came unexpectedly, they did not realize they were VOCALOID™ demos and thought they were some sort of prank.

"Daisy" was demonstrated at the 6th anniversary of VOCALOID™, where a file called "Fly Me to the Moon" was played, the file was originally created for 7/16/2002 when Crypton were shown the first demonstration in their Sapporo office. "Daisy" still had troubles with consonants at the time.

"Daisy" dropped as a name due to conflicts with copyrighting - despite attempts to change the name (such as translating it into Japanese), they ultimately could not register it

The only 4 known vocals for "Daisy" were; LEON, LOLA, HANAKO and TARO. LEON and LOLA were the only ones ever to be shown to the public, releasing as official voicebanks for the final VOCALOID software.

Комментарии

Информация по комментариям в разработке