Tempest Ōkaminoto / テムペスト 狼のと - Introduction
Tempest is a virtual vocalist for the UTAU Singing Synthesis Tool. He sings in English and Japanese, using different voice libraries.
Currently, he has 2 upcoming voice libraries. One of which is English "UNISING" which uses X-SAMPA phonemes to transcribe different sounds into notes into the piano roll. this uses the common UTAU format of CVVC for english.
The other voice library - Japanese - is a VCV style voice library, with aliasings for CV and CVVC Japanese utilisation.
These are currently unavailable, but stay tuned for them! He currently has three different voicebanks available for public download, however.
VCV is the most highly regarded method for UTAU in Japanese, as it is very easy to make it sound smooth and natural. This voicebank is, as of current, the highest quality on this site. The voice library is lighter in tone and more of character than the CVVC voicebank option is. A future powerskale/kire voicebank is scheduled, though the VCV currently provided is functional, despite having a lower quality configuration. (Auto generated by Moresampler).
JP VCV "Light"Download
English "ARPASING" (Multiple)
Arpasing is an UTAU method designed by Kanru Hua - it's based on diphones and transitions every phoneme individually, allowing for extremely smooth and seamless results. The voicebank quality is not the highest however, with click sounds being present. In addition, the ARPASING method can sometimes get messy on intricate phoneme groups, as well as faster tempo sections.
Below are the downloads for to different arpasing voicebanks for Tempest. One of which followed a "British english" reclist, which leads to more natural-sounding pronunciation. The other, is a reclist with every single possible transition present between each phoneme - which means you can definitely make him sing "voice". Both of these reclists were written by myself (HitCoder).
Arpasing GB-ENG 240717Download
Japanese CV (2016 / Early 2017)
"CV" refers to "consonant-vowel" transitions - the voicebank is choppy and slightly robotic as a result. The tone is particularly soft in addition, but this leads to an overall unique aesthetic. Much of the pronounciation is incorrect, namely palatalised consonants (like "kya" or "gyo").
The voicebank is not of a high quality, where the CVVC style voice library listed below is of a better quality. However, this voice library is completely configured, so for a general audience would be more suitable.
Japanese CVVC (Late 2017)
Tempest's Japanese CVVC voice library is available for download, but is currently not completed (and may not be completed due to becoming deprecared, with a future VCV style voice library around the corner to come in place of it). However, if you are looking for a more mature tone for Japanese, and also a smoother and more realistic vocal, this may be more suitable for you. Some samples are missing, and the configuration as of writing this is not completed, but for the most part configuration can be manually updated if need be.
The overall voicebank quality is higher than that of the CV style Japanese voicebank, but neither of which are preferred over one another due to the CV style voice library being completed, while the CVVC is incomplete.
English CVVC (Early 2017) (NOT RECOMMENDED)
Tempest's English CVVC (2017) voicebank is additionally unfinished, and is not recommended for usage. It is available for legacy reasons, in the case of stylistic choice, or user preference. The configuration for this voicebank is currently incomplete, and the quality of the voice library is much below the current desired standard. Be aware that this is currently an obselete product.
In addition, this voicebank did not follow a standard recording script, but was instead recorded using audacity running in the background, referencing a phoneme list in microsoft excel. The voicebank was recorded on a headset microphone, which most likely was a Turtle Beach X12 headset microphone.