HANKO PROJECT: Voice Provider Audition
Project Overview
Hello everyone! Do you like Hatsune Miku and other vocal synthesis characters? Are you a singer? Would you like to know what it would sound like to have your own voice go through a vocal synthesis software? This project is made for you! We are looking for a voice provider for the upcoming UTAU vocal "Hanko" ー please keep reading for more information!
Who is Hanko?“Hanko” is a vocal synthesis character ー similar to Hatsune Miku ー first created for the free vocal synthesis software UTAU, and eventually released later for DeepVocal as well. Hanko is a young adult going by she/they pronouns. She’s 165cm tall (172 with her heels!). She’s a singer and keytar player who loves orange cats and drinking lattes!
Like many vocal synthesis characters, Hanko is a blank state so everyone can express their creativity freely through her, and the information provided about her character by her creator is no more than a creative base. Hanko can be anything you want her to be!
Hanko has existed for a while and never really got her opportunity to make it, which is something we hope to change through rebranding, both in design and voice ー and we need your help !
What is UTAU?UTAU is a free vocal synthesis software, often compared to the wider-known commercial VOCALOID software. Select the voice bank of your choice, write notes, input lyrics, and play to hear your vocal sing what you made!
UTAU also allows its users to create their own voice banks for free, being its greatest strength back in 2008 and up to this day ー resulting in a wide variety of voices and characters from all over the world! Some popular UTAU vocals include Kasane Teto, Yamine Renri, Namine Ritsu, etc.
UTAU, just like VOCALOID, uses what we call concatenative synthesis, which consists of using flat recording strings containing different series of consonants and vowels for the software to mash everything together to recreate smooth transitions from consonants to vowels and vice-versa. It may sound complicated, but it’s quite simple in practice!
If you are curious about how UTAU sounds, here are a few examples:
Who are we?We are a team of two friends and vocal synthesis enthusiasts from France, doing this as a passion project and out of attachment for Hanko’s character, willing to do everything to see her shine:
-
Nobu (Twitter @ nobu_222_)・20・Any pronouns
Creator of Hanko, manager of the project, illustrator and vocal synthesis software user since 2017.
“Hii, I’m Nobu ! The owner of this project along with Orion, I always loved vocalsynth since I was really little. Hanko is a character that I cherish a lot and would love to give her another chance to get the spotlight! thanks to anyone who’d be interested in this project with the both of us.” -
Orion (Twitter @ orionhime_)・19・They/them
Previous voice provider for Hanko, adviser for the project, singer, illustrator and vocal synthesis software user since 2015.
“Hello everyone! Orion here! I’ve loved vocal synthesizers for so long, and Hanko is very dear to me. I’ve stepped down from my role as her voice provider to focus on recording for my own characters, but I still want to see Hanko shine and get her own chance in the vocal synth scene! We hope you’ll all love her as much as we do, and thank you for considering our project.”
Feel free to contact any of us to ask questions!
Requirements, What we ask for, ExpectationsAs in any project out there, we will both ask for a few requirements from applicants and we have eventual expectations for the future to be taken into consideration by applicants. We ask all applicants to read this section well!
We first of all ask for good microphone quality: as little background noise as possible and clear sound. If you don’t have a good microphone yet and plan to purchase a new one soon enough, feel free to show us what you’re capable of still!
We expect applicants to have a good Japanese accent and to know how to pronounce words in the language correctly. We obviously need the applicant to be able to speak English. If you are proficient in other languages (at least confident enough to be able to sing in those languages), feel free to tell us: it’s a very valuable bonus!
For Hanko’s voice, we would like a youthful, bright and powerful female voice, slightly more on the low end, as well as good singing skills ー the best example we found is Ikaruga Luca from THE [redacted] Shiny Colors, voiced by Kawaguchi Reina. Here is a playlist of examples of her voice!
As her voice provider, we’ll be asking you to record the UTAU voice bank from a precise list of Japanese lines. Here are a few explanations to help you understand what we expect of the final applicant:
The list is ~35 lines long. We will ask you to record it multiple times, on different pitches and with different expressions.
The list will each time be recorded on a fixed pitch (ex. C4, E4, etc). We will decide which pitches to record based on the chosen voice provider’s range.
Every line is a continuation of syllables to be said continuously on a fixed pitch (ex. n_ka_ki_ku_ke_ko_ko, n_ta_ti_tu_te_to_to, etc). We’ll have you record some extra samples as well (breaths, vocal fries, etc)
For Hanko to have a wider range of emotions, we want her to have other expressions aside from her normal tone: a softer one, and a more powerful one. Each expression will be recorded on the same pitches as the base normal tone.
We would also like to have the chosen voice provider record a single pitch in falsetto. Pitch is once again to be decided depending on range.
The voice provider will be provided a software made to record the voice bank more efficiently.
We would like to record 4 pitches for each expression, resulting in a total of 13 pitches. A single pitch approximately takes less than 10 minutes to record when you’ve got how to record it down! We will of course be here to assist the voice provider through the recording and to answer all their questions.
If you want to understand how UTAU recordings sound like beforehand, we recommend you to download a CVVC UTAU voice bank yourself and browse through the files (they’re all raw WAVs, so feel free to listen!) We recommend Koharune Ami Buffet Lunch.
While for now, we only consider making a voice bank for UTAU, we would eventually love to pay the voice provider picked through this audition to sing songs for a dataset to be used for a DiffSinger voice bank; DiffSinger is an AI engine, resulting in much more realistic possibilities than the concatenative UTAU engine. As we absolutely do not want to be involved with unethical AI, a DiffSinger voice bank will only be made with consent from the voice provider : we want every applicant to tell us their stance on the idea of providing data for an AI vocal in their audition. You’re also free to contact us and ask questions about it!
The use of AI in vocal synthesis is only to make vocals sound more realistic ー it doesn’t involve generating music or melodies and still requires user input and tuning to be used. We refuse to source data from non-consenting singers for our project.
If you are curious about how DiffSinger sounds, here are a few examples:
When you are logged in, you can comment, add submissions, create projects, upvote, search open roles, and way more. Login here.
“Hanko is a young adult going by she/they pronouns. She’s 165cm tall (172 with her heels!). She’s a singer and keytar player who loves orange cats and drinking lattes!”
“For Hanko’s voice, we would like a youthful, bright and powerful female voice, slightly more on the low end, as well as good singing skills ー the best example we found is Ikaruga Luca from THE [redacted] Shiny Colors, voiced by Kawaguchi Reina. Here is a playlist of examples of her voice!”
Tell us about yourself in your audition note, and don’t forget to tell us about any other languages you’re proficient in as well as your stance on the eventuality of recording for a DiffSinger voice bank in the future.
- female young adult
Please state your name and sing something in Japanese! It can be anything you want ー show off your singing skill and vocal range!
Comments
Public Submissions
