Also with neocities, I need to use the CLI tool to upload the website which is (a lot) more annoying than the github workflow. I might try to make a workflow that does this anyway, but I am not sure that that’s possible.

Additionally I’ll need to delete all of my youtube shorts and instagram reels, which is sad. I will have to spend an annoyingly long time editing them and making the link at the end accurate.

Thank you for reading. I will work to update all information about me to link here instead.

I know, I still need to do the bigger resource sections with CVVX documentation and stuff. I’m working on it, it’ll just be a while. Sorry!

Thanks for reading!

Aditionally, I finally published Kobayashi Umeko (previously known as Magnesuin Mari)’s demo song. Watch it here!

The voice provider, Mari, drew the thumbnail with her character, as well as the standing artwork, which is vidible on the character’s page!

Aditionally, standing artwork now exist for En’ne Jeho and Girlthing! All of this artwork is made by the voicebanks’ respective voice providers, which I find very fun!

Also, I’m working on a large post about CVVX/VCVX/CVX as well as a resources section. Most likely, those will all become available within this week.

I also did a lot of restructuring of the site (mainly changing all of the demo songs (ow) so that they’re all the same now), so there may be issues. Please let me know if you find any, and I will fix them ASAP!

Thanks for reading!

Suigin Koora Updates & “CVVX” Phonemizer

With the release of Suigin Koora (and OIL) on NeoDIVA, I took a break from UTAU. I think I’ll start doing stuff again, I have just been incredibly busy recently.

I recorded a new JA CVVC voicebank for Suigin Koora. It has multiple pitches and voice colors. A demo is not yet available because I have only OTOed the first pitch, C3. Eventually, it will be more like this, with different voice colors having differet pitches. (an X represents a subbank)

The NT version is a joke, as it was recorded with me plugging my nose to make it more nasally.

It will have the following extra phonemes:

• l (English l)
• wr (English r)
• rr (rolled r)
• M (M variant of ん (before b/p))
• NG (NG variant of ん (before k))
• ff (english f (Japanese f is more accurate to how it should be))
• dj (ぢ’s consonant)
• v Ex (exhale)
• v In (inhale)
• v Cm (complex breath (out then in))
• v Fr (vocal fry)

Originally I was only planning to have it use a presamp.ini file to manage the extra phonemes, but it didn’t work properly with the JA CVVC & VCV Presamp phonemizer. After converting it to romaji, I could get it to generally work with the ZH CVVC Presamp phonemizer, but some things didn’t work. In the end, I decided ot commission cadlaxa to make one for me. It uses a YAML dicionary. It is called the JA CVVX phonemizer. It is still in development, but eventually I will release it alongside my voicebanks (with some restrctions). So far it seems really good! Ideally, it’ll act similar to DeepVocal’s phonetic dictionary(or dictionaries..) which will make it much more customizeable. When I get my hands on it, I’ll make a blog post describing how it works and stuff, as well as making all of my currently-public voicebanks work with it. I’m not certain about the name, but I think it makes the most sense.

Also (yes there is more lol), Suigin Koora DeepVocal has been put on halt for a while. I am working on many other projects, and it isn’t a very high priority for me. If I do it again (which I will, honestly I kinda don’t like how “stiff” CVVC V2 sounds), I will use either v3 recordings or completely new ones.

Kikyuune Aiko DeepVocal

Speaking of DeepVocal, I am making the official Kikyuune Aiko JA DeepVocal voicebank! It’s been a really fun project, and I am nearly finished with it (other than t phonemes, which haven’t been recorded yet). Aiko is one of my favorite voicebanks ever, as she was the first UTAU voicebank I ever really used. I think DeepVocal sounds really cool and I’m so excited for it to release!

Tsukumo Shion DeepVocal

Yeah I’ve been a little obsessed with deepvocal..lol

I started working on a CV DeepVocal port of Tsukumo Shion, before realizing that I should definitely ask permision before releasing it, so I stopped development and sent Missile (the person who made the vb) an email asking for permission. If they respond/give me permission, I will conitnue development. Also DV doesn’t like CV apparently lol, it “works” but is just kinda weird.

Demo Songs

I have upcoming demo songs, I just am working on too many things at the moment ot make much progress on them. Here’s a list.

• Suigin Claire - Overload (Kirisame)
• Suigin Koora JA CVVC V3 - Electric Angel (Yasuo)
• Suigin Koora EN C+V & S-CVVC & GrayGlish & DiffSinger (EN) - Human (Flavor Foley)
• Suigin Shelly EN C+V - Weathergirl (Flavor Foley)
• Suigin Chris EN C+V - Electric Weekend Zone (Flavor Foley)
• OIL DiffSinger (EN) - It Just Keeps Going On and On (English translation of “It Just Keeps Going”/”Madamada Tsudzuku Yo” (Kaguo))
• Magnesium Maru CVVC - Amedama (Popeyo)
• Magnesium Mari VCV - Onomatope (Maigo Hanyuu)
• Max EN C+V - Metronome (Circus-P)
• Girlthing JA CVVC - Honeymoon Un Deux Trois (Daketen)
• Mizuki Mae JA CVVC - Darling (Suda Keina)
• Scrunge JA CVVC - Lagtrain (Inabakumori) or Telecaster B Boy (Three) (we haven’t decided yet)

Thank you for reading!!

Last evening I forgot to make an update about this, but NeoDiva released the DiffSinger mode for both OIL and Suigin Koora! You can download them from NeoDiva’s website! Their singer pages are now updated.

Additionally, I am working on a new JA CVVC voicebank for Suigin Koora. It has more voice colors (regular, power, whisper, and NT (recorded while plugging my nose)) as well as covering a total of 3 pitches. Eventually there will be a chart with all of them.

Also, I got ruby to work on my laptop, that way I can edit this website lol

Since it’s 2026 now, I want to share some of my goals with you all.

• I want to release all of Suigin Koora’s voicebanks, including the DeepVocal (once I make it better).
• I want to expand my YouTube channel, mainly by posting YT Shorts when I upload new videos, and to give previews of my new voicebanks (I have already uploaded quite a few!)

Also, some small updates:

If you look at Suigin Koora’s page, you’ll see that they now have a DeepVocal voicebank. I ported the CVVC 2.0 recordings to it and reconfigured it entirely. It still needs a lot of work of course, but it is cool to see how it sounds. The configuration process felt much more professional than UTAU/vLabeler’s, I’m not really sure how to explain that, but it feels like it was meant to be used to make commercial voicebanks and not just hobbyist projects.

As you may have noticed, the entire structure of the website has changed a bit, and I think that is a very, very good thing. Voicebanks specific to characters (for example, Suigin Koora VCV V1) now has its own page, seperate from Suigin Koora’s!

OIL and Suigin Koora DiffSingers!

Finally, I can show you demonstrations of the 2 DiffSinger Models I made! NeoDIVA did the training and pretty much everything, Oliver (my friend) recorded for OIL and I recorded for Suigin Koora! I did all the labeling.

I also made Suigin Koora’s Demonstration on my youtube channel (it’s a short).

Eventually, I will do some English demo song, but I haven’t decided what to do yet.

Mizuki Mae

My friend has finally recorded for Mizuki Mae, and we are going to release the voicebank publicly! It will still be a while though, since we still need to finalize the OTO. We are going to do Darling (by Suda Keina) as the main demo song, and Alien Alien as another song (Tuned by another friend)!

Lisatopian Reclist Things

I wrote an entire reclist and base OTO for the Lisatopian conlang. It’s a bit of a nightmare lol, but it’s very high quality!

Some of them, at least. The French voicebank has not been completely OTOed, and the toki pona and other ones haven’t even been recorded yet. Currently, the JA CVVC 1.0 and 2.0, JA VCV 1.0, EN GrayGlish 1.0, EN C+V 1.0, and EN S-CVVC 1.0 are available.

Thank you <3

Kouga

I Finally released OIL!

He is yet another voicebank that I have made! See his page when you get a chance!

Eventually, I will release the demo song (It Just Keeps Going, by Kaguo), but I have many other projects that I am working on for now.

Even further in the future (maybe not?), NeoDIVA will train and release OIL’s DiffSinger model, which will primarily be in english. OIL’s VP (Oliver) and I are working on an English translation of the same demo song used for Japanese to be used as the demo song for the English voicebank, it is a very fun process!!

Additionally, I updated En’ne Jeho CVVC V1 to 1.1.

In this version, I fixed some issues with the README.md and applied noise reduction to the samples which you can read about in the last post.

Suigin Kurea Updates

I also recorded and OTOed a VCV voicebank for Suigin Claire. I still need to explain to the voice provider what releasing the voicebank means, but I would predict that she will allow it to be released. If it does get released, I will make another update and stuff.

Mari Updates

Another thing, I recorded and OTOed a VCV voicebank for Mari. The voice provider does not want me to release the voicebank publicly, but we have multiple ideas for demo songs which I am excited to work on!

Thank you so much for reading!!

-Kouga

I tested UtaUtaUtau’s Threaded Noise Removal Python script on En’ne jeho’s voicebank (since that was the only one I had installed on my laptop). The difference wasn’t very noticeable when you listen to the audio with this voicebank since it has very little background noise, but some of my newer ones (Suigin Claire, Ely, Oil, etc.) have quite a lot. I am going to run this script on En’ne Jeho as well as all of my other voicebanks because I think it will make them more consistent. It should also help some of them sound better.

EDIT: Releasing soon, I will make another post when it does!

Example image: