|💻||Technology article work in progress.
For information on how to help, see the guidelines. More subjects categorized here.
NetVocaloid was an online vocal synthesis service. Users could use synthesize singing voices on a device connected to the Internet by executing the VOCALOID™ engine on the server. This service can be used even if the user does not actually own VOCALOID™ software in such device. The service had been available in English, Japanese and Spanish and simply required a internet service and a method to interact with its services.
It was set up to function as a bonus service for users to make better use of the VOCALOID™ software. The service, also featured the original version of VocaListener. On April 28, 2012, the service was closed. It would later be replaced with VOCALOID NET. The main difference between the two services is that NetVOCALOID catered for both English and Japanese, while VOCALOID NET catered only for Japanese.
NetVOCALOID had catered to all voicebanks for all VOCALOID2 vocals, both English and Japanese. It also allowed access to all of them via its Network. However, commercial releases that accessed this network were limited for the general public and all voicebanks were restricted.
- Synthesizing audio data of VOCALOID™ singing from sequence data of melody parts written in specialized XML files [*1].
- Running effects (Reverb/Compressor) on audio data of synthesized singing.
- Encoding audio data of synthesized singing as compressed audio files (MP3/FLAC) [*2].
- Providing the above functions with HTTP protocol-based WebAPI.
Although NetVocaloid offers the synthesis service, it doesn't provides the tools for the creation or edition of the required XML file for the synthesis, relying in the tools an provided by the companies that make use of the service.
The first version of VocaListener was available via the service.
VocaListener is a singing synthesis system that automatically estimates parameters for singing synthesis from a user's singing voice with the help of song lyrics. Its creators are Tomoyasu Nakano and Masataka Goto. Originally it was set up to work with NetVocaloid and would work with VOCALOID™. It would work with Japanese VOCALOIDs, though English demonstrations of some of the services did exist.
AIST (National Institute of Advanced Industrial Science and Technology) developed the system in which users can adjust the limits to where the original singer can normally sing to where the VOCALOID can better match the original singer's inflections and details to the point of sounding like the original singer.
In demonstrations, Hatsune Miku and the Append voicebanks were used not only separately but blended together; this added many layers to the singing output that a single voicebank alone could not achieve, resulting in a much more human-like tone.
- Demonstrations of the follow VOCALOIDs: Hatsune Miku, Kagamine Rin(Act1), KAITO, MEIKO, Kagamine Rin(ACT2), Kagamine Len(ACT2), Gackpoid, Megurine Luka, Megpoid, Kiyoteru Hiyama, Kaai Yuki, SF-A2 miki, Hatsune Miku(Append), Lily, VY1, Gachapoid, Nekomura Iroha, Utatane Piko, Kagamine Rin(Append), Kagamine Len(Append), VY2
VocaListener was never made available for download. This version was restricted to research groups only.
One of the advantages this service offered was vocal rendering without the need to buy the voicebank or VOCALOID engine itself. So long as you had a legit method of accessing the NetVOCALOID service, you had a means to render vocals.
How voice rendering worked was similar in each case, but below is an example using the ' Let's sing with Miku ♪' app:
- In the app 'Let's sing with Miku ♪' users would type in lyrics that they intended Hatsune Miku to sing, for example "Sakura".
- Once completed they would send off the data via whatever method was connected to the NetVOCALOID service, in this case the ' Let's sing with Miku ♪' app. "Sakura" would be sent to the service to be rendered.
- The data "Sakura" would automatically be rendered by the NetVOCALOID service with the correct voicebank, Hatsune Miku.
- The user would then receive the completed "Sakura" file from the NetVOCALOID service with Miku singing "Sakura" as requested.
Many of the English and Japanese VOCALOID2 vocals were available while the service existed, and it even had voicebanks that were not declared such as the vocal for 'Metal Gear Solid: Peace Walker'.
The user did not have much control over the vocals and it was not always the best method of creating vocal results, but it was not impossible to get samples for use in music and write entire songs using the service. Though the user did not own a voicebank, the results of the services was theirs.
Smart Phone' Service Track Record
Netvocaloid is used in two mobile phone services, allowing send singing messages. These ones are Utaou (Sing with Miku) and Keitai Gackpoid, allowing you use the voices of Hatsune Miku and Camui Gackpo respectively.
Let's sing with Miku ♪
It was also used in promotion for Let's sing with Miku ♪.
Mobile Phone Gackpo
It was used in promotion for Mobile Phone Gackpo.
Metal Gear Solid: Peace Walker
NetVocaloid was one of options available as DLC for Metal Gear Solid: Peace Walker. This option allowed to customize the phrases sung by Metal Gear ZEKE during the game. The user could use either text/lyric input or phoneme input for edit the pre-established tracks available within the game. 8 additional tracks were available for in the different DLC packs available for the game. This is one of the few examples of VOCALOID-flex that was available for use by the general public. This option was shut down on April 28, 2012 when Yamaha removed the NetVocaloid service from their website.
NetVocaloid was utilized during the 2nd edition of PlanB, a collaborative project sponsored by Ballantine's where the people can help to create the lyrics for a famous artist.
Voctro Labs, S.L. provided the tools that allowed to the participants listen their written lyrics sung by a virtual female vocalist through the official website. For that the company created Ballentunes, an application that allowed to the user edit the lyrics accord the pre-established melody created by La Oreja de Van Gogh, in an easy and intuitive interface through the website. Ballentunes then transforms the input lyric into a pre-tuned VSQ, for be synthesized through NetVocaloid's service. After this, the users receive a mp3 file of their entry which they can listen and share, if they wish to.
The contest ended with "Otra vez me has sacado a bailar" chosen as winner.