DeepVocal is a free to use singing synthesizer application software developed by Boxstar. It is the successor to the Sharpkey Galaxy software, which was a successor to the Sharpkey software.
History[]
The editor and its first vocal, Xia Yezi, were announced on June 7th, 2019, through a video on Boxstar's bilibili and Weibo.
- Version 1.0.6 was released on July 15th, 2019. This was the original public release of DeepVocal.
- Version 1.1.0 was released on August 9th, 2019. This update brought improvements to voice models.
- Version 1.1.2 was released on August 11th, 2019. This update made voice models crash less and accept more selections, fixing the common "unvoiced section exists in vowel" bug in most cases.
- On September 9th, alpha testers were requested to work on gender and CC and beta-test it before it was released. This was version 1.1.15.
- Version 1.1.16 was released on September 19th, 2019. This included the gender feature and CCs, listed in the Features section below.
- On January 20th 2021, Version 2.1.0 was released. This was also known as DeepVocal2. This update included vocal engine improvements making general vocals sound better. However, this update also renders all Independent Symbol notes generated with the 2.1.0 DeepVocal Toolbox to have no audio. DeepVocal 2 backwards compatible with 1.1.16 DeepVocals.
The original logo
The original thumbnail
Voices[]
All previous voices released for Sharpkey are also available for use in DeepVocal, except for Akane Inari and Suoyun Rila, who are available only in the Sharpkey Editor and the Sharpkey Galaxy editor respectively.
A list of all DeepVocal voices on the wikia can be found here.
Features[]
DeepVocal at first had the same features as its predecessor, the Sharpkey Galaxy Editor. It has added multiple features through updates since.
- Volume - Changes the volume
- Pitch - Changes the pitch
- Breathiness - Changes the breathiness
- Gender - Changes the formant of the voice, making it sound more masculine or feminine
- Phoneme - Changes the timing of phonemes and transitions
- Timbre - Changes which pitch is used for the note
Additional features have been implemented, such as the implementation of CCs. If one was to place an unvoiced consonant in both the Unvoiced Consonants tab and the Independents tab of the DeepVocal Toolbox, then when a note was entered containing that consonant it would play the proper VX and then the independent. Further updates to refine this system are planned.
Controversy[]
Controversy first struck DeepVocal over its terms of use, in which it was implied that a character could not be released for both DeepVocal and another synthesizer. This was due to a mistranslation in the English EULA. It has since been updated to reflect the proper meaning, being that one cannot use the exact same art piece one used for their released DeepVocal for another synthesizer. This has many workarounds, however, such as releasing an UTAU with that art first and then releasing your DeepVocal. Regardless, this has never been enforced.