The DefoCollection

This project is a collection of UTAU Voicebanks inspired by the default library packaged with UTAU. The AquesTalk-F1 voice and the AquesTalk software as a whole have both evolved a great deal since UTAU's Default Voice was released and this project seeks to showcase the changes that have been introduced to both the AquesTalk and UTAU softwares. This collection has a total of 5 Voicebanks for sampled from different versions of AquesTalk, which can be used alone or together. Each Voicebank is sampled from a different "Phont" or voice found within AquesTalk and has support for a wide variety of features that have become commonplace for modern UTAU voicebanks and also includes basic english support.

This project is still in development. This page is also,,, unfinished so like,,, shhh don't ask about it yet

About The Collection's Voicebanks:

Two of the available banks are sampled from the AquesTalk1 software and provide new features to the voice while sticking closely to the tone of the original Default Voice. These Voicebanks have had support for English vowels added through manual formant shifting in iZotope RX

The AquesTalk-F1 RetroModern Voicebank is sampled from the 20th Anniversary Edition of AquesTalk1 and has improved vowel clarity along with a near complete reworking of consonants. This library uses the most recent version of the AquesTalk software and features added support for the consonants "kw" and "gw", which are exclusive to this library

The AquesTalk-F1 VintageClassic Voicebank is sampled from the original edition of AquesTalk1 and is an attempt to faithfully recreate the tone of the default library packaged with UTAU with only minor splicing to correct weak or incorrect pronunciations. This library uses the oldest version of the AquesTalk software and features an extra liquid R consonant, which is exclusive to this library

Two of the available banks are sampled from the AquesTalk2 software and provide a new range of expressiveness to the voice. These Voicebanks have had support for English vowels added by creating custom formants in PhontDesigner

The AquesTalk-F1 HardDrive Voicebank uses the PhontDesigner Software by AQUEST to alter the tone of the base Phont, AquesTalk2-F1b, into a strong, tense vocal. This library features EVEC-Style Growl samples, which may run into rendering issues with some resamplers

The AquesTalk-F1 SoftWare Voicebank uses the PhontDesigner Software by AQUEST to alter the tone of the base Phont, AquesTalk2-F1b, into a weak, whispering vocal. This library may run into rendering issues, with some resamplers being unable to detect the F0 pitch

The final UTAU voice library is sampled from the AquesTalk10 software and provides a wider vocal range. This Voicebank has had support for English vowels added by assigning new formants with the Praat Vocal Toolkit

The AquesTalk-F1 MultiThread Voicebank is sampled from AquesTalk10 and has five pitches (D3, G3, B3, E4, and A4) to cover the reasonable range for the engine. It is a relatively soft bank and is the only AquesTalk engine to have native acoustic data above 4kHz. While the tone of this library is generally considered to be the cleanest, due to the lack of SRC upsampling, it may sound out of place when used alongside other voicebanks from the collection

All UTAU Voicebanks for the collection contain the following features:
Full CVVC Support for Japanese, including Vowel-Vowel transitions.
Endbreaths and Natural Ends
Support for [tyu], [dyu], [hu] and [fyu]*
CVVC support for English Vowels: [AA, AE, AH, AO, AX, EH, ER, UH and UW]**

* Not all AquesTalk Engines natively support the [fyu] or [hu] samples. If the sample is missing in the specific engine, the sample has been  synthesized manually through splicing or devoicing, respectively.
** English Vowels do not have support for Vowel-Vowel transitions

Usage Guidelines:
These voice libraries were configured and are managed by Jackhuwubenak and they do not have any affiliation with AQUEST or Ameya/Ayame.
Credit to Jack Huwubenak or the DefoCollection Project is not required for the use of these libraries, but is appreciated
Credit to AQUEST is required for any use of these libraries
Redistribution is not allowed except in the instance that this page is no longer available
These banks or their samples may not be used for any public-facing projects involving AI and may not be ported to any TTS software, doing so is against AQUEST's Terms of Use

In accordance with Ameya's Guidelines for the Default Voice, it is recommended to refrain from using the Defoko character or name with these libraries, the name "Defoko" and "Utane Uta" belongs to the original CV version packaged with UTAU and managed by Ameya and should not be associated with third party voicebanks made from other versions of AquesTalk.
In accordance with AQUEST's Personal Use License,  libraries in this collection may not be used for commercial purposes, including monetized uploads.
If you are interested in using these libraries commercially, please purchase the necessary licenses at AQUEST's online store. Please note that they do not ship overseas and you would need a proxy 

Contact:
I can be contacted through Twitter or Discord at @jackhuwubenak or via email at jackhubenak@gmail.com
If you use any of these libraries, you're welcome tag me in any posts or just shoot me a dm with a link! I'd be delighted to see what y'all do with them
Also, if you notice anything wrong or screwy in the banks, let me know so I can correct it for future versions of these libraries