____________________________________________________
All sprite images from publicly released voicebanks can be found through this link for covers and stand-in images in case the main image is not found in the original voicebank's files.
____________________________________________________
WH1SH-P3R (ウィスパー) || 'Whisper' misspelled
Gender || Non-Binary [They/Them]
Age || Unknown
Sexuality || Pansexual
Type || Monster (Demon) + Doll + Furry (Rabbit)
Personality || Unhinged and Unpredictable
Vocal Type + Range || Androgynous / Alto
____________________________________________________
Created in 2016 with the idea of an experimental FURloid UTAU, each bank is created with the experimental intention of recording different vocals from one voice provider leading to vast choices provided in the library.
WH1SH-P3R had never experience what is like to be human so they consistently adapt to 'different personalities' to accommodate. Enough so, you can't recognise if they are the same person.
In headcanon alternative universe, WH1SH-P3R can be found in multiple separate vessels as their own person.
____________________________________________________
Fun Facts:
WH1SH-P3R is not the first Bruneian UTAU however, all or most of UTAUs created in Brunei had been considered obsolete or inactive, rendering this UTAU to be the most active by updates.
WH1SH-P3R is considered to be the first DiffSinger to be made from a Bruneian voice provider with high English fluency with a slight accent that can vary.
They are partnered with Akuma Taikino // 大気のアクマ as her girlfriend canonly.
____________________________________________________
SORROW (formerly known as B1N4RY)
Depicted as carefree and mischievous in personality. Sometimes even awkward depending on the scenario. They don't care enough to understand people but is particularly interested in interactions with humans; wishing to form a personality around these people. They are the only one who can speak English more fluent other than AGONY and DS. They're considered the main design for UTAU, wearing a red dress and torn off bright pink jacket. They have an uneven short hairstyle.
AGONY
Depicted as quiet and gentle however will not hold punches if they are threaten. Working as a nurse makes them feel obsessive to help others whenever they can. They can speak English just as fluent as SORROW however their accent spectrum is slightly British. They are shown as a nurse with wrapped bandages covering majority of their face except their polycoria eye. They have a short low ponytail.
WRATH (alternatively known as DESPAIR)
Depicted as passive-aggressive and arrogant though they are mindful who is around them. Their attitude can be almost comparable to a rebellious teenager despite being the same age as the other appends (excluding REGRET). They are shown to have a slight fancy outfit with high end socks, rebelling with majority of the usual designs. They have low ponytail. They are often seen with REGRET as a similar dynamic to siblings with them being the older sibling in headcanon. Their design may seemed similar to EL3CT0's design due to similar tones.
REGRET
Note: This append version of WH1SH-P3R excludes canon romantic relationships and is separated from canon due to their child/teen-like appearance and vocals. This is to prevent illegal usage of the voicebank and design of the character, please look into Terms of Service for more information.
Depicted as child/teen-like with a kind and naive attitude. Despite that, they've shown to be very stubborn and cling on people whenever they want which often results in situations of getting lost from home. They are seen with a blouse with a red pinafore. The only significant removal was the notable polycoria in their left eye being replaced with a hollow hole. Their hair is shown to be short and childish compared to SORROW's hairstyle. They are often seen with WRATH as a similar dynamic to siblings with them being the little sibling in headcanon.
SOLEMN
Depicted as a flamboyant and brooding for a doll featuring being half-human and rabbit. Despite the demon's vessel being the odd one out, they are very androgynous in feature. Inside of relationship headcanons, they do not associate with any other appends due to their vessel appearance making them one of the more unique cases allowing flexibility of any skin tone colour. Pale and Tan are considered the default skin colours upon depiction.
DS (DIFFSINGER) (alternatively known as DISSOCIATION)
Depicted as very cocky, reckless and flirty compared to other variants, making them the most disliked among the regular UTAU appends of the DISORDERED series by headcanon. They can speak English more fluently and mimics both types of vocals (feminine and masculine). Despite being similar as SORROW's and WRATH's appearances, they are a completely different entity that takes on a half-formal suit with a jacket. Their uneven hair covers their right eye and has a low ponytail.
[Most banks are listed in MediaFire but we are trying to transfer all voicebanks over to Bowlroll when possible! Please be patient. Discontinued banks will not be listed in this website.
If there is any issues with the voicebanks, please contact K1RA so they can update accordingly.
All voicebanks are optimised to be used in classic UTAU and OpenUTAU. Other softwares may vary.]
[ Facing own identity ]
"A series of Japanese voicebanks that associate gender as their core elements. This voicebank series ran from 2016 to 2023.
Majority of the old images and personality of WH1SH-P3R from GENDERED series does NOT reflect their current design and personalities so please do not combine the same with the DISORDERED series. Some of the GENDERED series' designs in this website is a remastered version of 2016 which can be used anytime for their respective voicebanks.
All voicebanks in this series will be forbidden to use in certain scenarios, please read Terms of Service for more information."
"EL3CT0 is the non-vessel form of WH1SH-P3R making their appearance different from the main line. Can be considered as an alter ego however is still considered as the same character."
DEMONS' ORIGIN CONCEPT/LORE
"The original Vihilines (demons) before they fused were unique to different tones and octaves.
The Vihilines in the GENDERED series are known as
- CH41-L1N (a demented mermaid)
- R3D-R34P3R (red-riding hood aspected fox)
- H4N5 (demented tv head)
In the DISORDERED series, only one Vihiline exist as EL3CT0 due to fusion under B1N4RY (which leads to the creation of successor banks in DISORDERED series).
CH41-L1N is the main Vihiline who takes control of the vessel which leads their appearance found more dominant in EL3CT0's design.
However, it is not clear whether R3D-R34P3R's or H4N5's personalities and appearance ever appeared more often but it's more noticable throughout different appends in DISORDERED series."
[ Mentally distressing to the core ]
"A series of Japanese voicebanks that associate trauma as their core elements. All banks in this series utilises 8-mora VCV reclist containing glottal stops and extra words for singing extras like English with an exception of DOWNPOUR and SOLEMN banks using CVVC 8-mora reclist by SalemWasteland.
All VCV banks in this series can be combined in order to create a multiexpression-append bank."
[SORROW]
[DOWNLOAD]
(Natural Bank)
"Neutral sound with an androgynous tone. Lack of emotions during recording."
vcv sorrow_Track4.wavvcv wrath_Track4.wav[SOLEMN]
[DOWNLOAD]
(Experimental KIRE bank)
"An experimental multiappend-pitch bank with 5 pitches.
The higher the note, the more cutesy it is, the lower the note, the more mature they sound.
Comes with a prefix.map."
A3 - Whisper
C4 - Neutral
F4 - Nasal
A4 - Tense
D5 - Nasal High-Pitched
cvvc solemn_Track4.wavDesign by BlinkkyDrop
"These voicebanks are under the DISORDERED series (and some under languages) however are only exclusive to this website.
This section is a bonus for those who want to play around with unique styled VBs that are fully completed but doesn't necessarily fit very well with the other sectors' idea or themes. This does not include scrapped, private or discontinued voicebanks made.
Some voicebanks will not have new sprites/images be provided if the bank vocals are similar to another earlier released voicebank or not likely present with a new sprite for any reasons."
[DESPAIR - VCV JP]
[DOWNLOAD]
(Emotional ver. of WRATH)
"Power sound in androgynous + emotional tone."
An attempt to belt vocals but it resulted similarities to WRATH. Less monotonous; more expressive than the original.
despair_Track4.wav[DOWNPOUR - CVVC JP]
[DOWNLOAD]
[not recommended solo but you do you]
"Growl sound; severely exaggerated."
Never done death growling resulting in the unfortunate lost of vocals during recording.
Most difficult voicebank to use but it does have a better effect when it sings high tempo songs.
There isn't any fixing of frequencies while oto-ing.
downpour_Track4.wav[MALAY SOLEMN - CVVC]
[DOWNLOAD]
"On par with SOLEMN JP, this VB is to commemorate the new version of spoon's Malay Reclist 2025 ."
Comes with 4 pitches and a prefix.map created by spoon.
[ A3, C4, F4 & B4 ]
Sample MALAY SOLEMN.wav[ Expand your region ]
"A series of voicebanks that allow singing of different languages outside of the main language of Japanese.
Each of these language voicebanks may have different tones during recording and will not sound similar to majority of the current Japanese banks unless stated otherwise.
Each one of the voicebanks in this section attempts to appropriate the right sounds to the language associated."
Sample CN.wavSample KR.wav[ENGLISH BETA - VCCV]
[DOWNLOAD]
[CZloid's American Reclist]
(this voicebank's oto.ini was never completed but you may request the current oto.ini through @ K1RA otherwise please use its predecesor, ARPAsing.)
Sample EN.wav[ENGLISH - ARPASING ]
[DOWNLOAD]
[Avocado's Word-Based Reclist](predecesor to VCCV English)
oto-ed || Ryankun12
arpasing.wav[MALAY - CVVC]
[DOWNLOAD]
[spoon's Malay Reclist 2021]
(outdated but functional; to make the bank somewhat compatible with 2025's reclist, replace larger [E] with [x] in the alias.)
Sample MALAY.wav[FRENCH - CVVC]
[PetitMot Reclist]
(work in progress // hiatus)
Sample FR.wav[ENGLISH (REGRET) - C+V]
[DOWNLOAD]
[Chevrefee's Reclist (Version 1.00)]
(A specialised ENGLISH bank to imitate one of the DISORDERED series.
Do note that the voicebank utilised the initial reclist which is version 1.00 during release, some of the words are different from the current C+V phonemizer provided so you have to tweak aliasing to get some words correct.)
Sample EN C+V.wav[Note: The DiffSinger banks shown are in BETA and the languages are separated from each other which means there is specific data for each (either more or lacking thereof).
In order for one to sing the other language in a song, you will need to download the specific banks for it.
In the future, there will be a chance that the quality may change overtime if it were to have an update or be fused together so please check the version of the DiffSinger voicebank written before downloading, thank you.]
WH1SH-P3R DS is a flexible androgynous DiffSinger voicebank that allows switch between natural feminine and masculine tones. Alto is the best word in terms of their dataset's range.
Currently there are only 2 languages that the DiffSinger has shown capabilities in singing on par with the voice provider. [JP & EN]
Regardless of what the voices colours were labelled in the voicebank, the main voice colour is always defaulted to an androgynous feminine tone which is better in higher octaves while the second voice colour is masculine androgynous tone which is better in lower octaves.
The only exception is the current version of EN DS (ver. 0.5.) where there are no voice colours however either of the vocals will activate according to the octave, ranges and expressions given so do experiment with it.
Utilising both with different tones is possible however will result in experimentation phases. Useful in solo-ing duets, choirs and layered harmonies.
Learning and Labelling || Mae Blythe
Trained along with pre-released Rewhi / Lewy with permission as a multispeaker with 11 minutes of Japanese singing data from the voice provider.
Range
[ estimation A3 - D4 ]
Both vocal modes ranges are mixed together during training allowing the vocals to be stronger in middle octaves however may not be effective in very high ranges due to voicer's dataset. Low ranges are possible to give an effect of a vocal fry.
WHISPER (JP)
ds w_MAINMORESAMPLER.wavMAIN (JP)
ds main_MAINMORESAMPLER.wavLearning and Labelling || Mae Blythe
Comes with some amount of data utilising Mae Blythe's reclists and English singing data to cover range and singing of the voice provider. It contains 1 hour and 30 minutes of average data.
Range
[ estimation E3 - D5 ]
Ranges are mixed together during training allowing the vocals to be much flexible with middle ranges however please be aware that the voicer's dataset does not allow any higher or lower than the average range stated otherwise it might just fail to work.
It is recommended to use this bank if you are experimenting with tones and octaves.
Learning and Labelling || Mae Blythe
Comes with some amounts of data utilising Mae Blythe's reclists and 15 minutes of English singing data to cover range and singing of the voice provider. Expressions are unstable.
FEMININE (EN) [ A3 - D5 ]
Higher octaves is recommended.
1_chorus.wavMASCULINE (EN) [ E3 - D4 ]
Lower octaves is recommended.
2_chorus_copy.wavNote: If any errors were to arise such as the DiffSinger's singing being strained, please play around with the tension curve or tone shift.
The recommendations of those octaves are only if it doesn't sound good in specific ranges.
[Reference Sheet]
"The jacket is actually removable as their UTAU counterpart, (SORROW)."
[Alternative Design]
"Can be used to replace the current design."