Z Idol Talker v1.1.0β Update Released

Character Driven TTS Software from Z Vocal Project!

Hello! Long time no see (It’s been a super busy couple weeks) but I’ve decided to make up for my absence by finally releasing an update to Z Idol Talker I said would be added shortly after launch, alongside another goodie. Feel free to report any bugs you encounter in the new version through this google form.

Full patch notes can be viewed here.

Z Idol Talker Lip Flap System Implemented!

Starting from version 1.1.0β onwards, your voicebanks will now sync mouth movements to the audio, provided they have a banner_open.png file in their file system. Specifically, the mouth will open if the volume of the sound is more than half the average volume. Adjustments to this threshold will be made based on user feedback, but I think it works pretty well for the time being.

Idolgotchi System Implemented!

As a bonus, Z Idol Talker V.1.1.0β also introduces the “Idolgotchi” system, finally giving use to one of the “reserve” buttons. From 1.1.0β onwards, this texture can be modified by setting reserved1.png OR idolgotchi.png. For full disclosure, this system is loosely based off some ramblings in one of the Maghni AI dev streams, it just seemed like a fun thing I could add while I worked on the lip flaps.

The Idolgotchi system allows users to feed your voicebanks a whopping FIVE foods. By default these foods are Dorayaki (Sweet), Lemon (Sour), Coffee (Bitter), Pretzel (Salty), and Hamburger (Savory). Voicebanks that do not have custom dialogue set will refer to each food very generically as it’s flavor-type. It is up to you as the voicebank to creator to write interesting dialogue for your characters when they interact with each food item.

The default foods

To set the custom dialogue for your voicebank, create a text file named “food.txt” and add entries for each food item in Sweet-Sour-Bitter-Salty-Savory order. Extra lines will be thrown out, and too few lines will likely crash the program.

To modify the food your voicebank has, add an 80×80 png with the name of the taste you want to replace. As long as you add custom dialogue the image does not have to match the taste to make sense, however their classifications are still important for matching dialogue to images.

The possibilities are truly limitless, and I hope you all will enjoy the new level of customization it gives for your voicebanks. I may gamify it a bit more in the future.

The foods for Defoko (Top) and Ethel (Bottom)

Introducing Ethel and Ambrose!

In order to test the lip flap system, we have created two new voicebanks (both voiced by Dunder lower your expectations) with designs that can make proper use of the new feature.

Ethel and Ambrose are a pair of time travelers who are doing a horrible job at blending in in the past. The pair have undergone surgery to drastically lower their ages, but their voices give away that they’re not really in their 20s. I’m not sure if they will be involved in Z Vocal Project’s story to any meaningful extent but I hope you will enjoy them regardless!
I’m not sure if they’ll get character pages, so for the time being I’ll ramble about them here.

Ethel has been modified to run off of electricity, but she’s still mostly human. The giant clock she wears isn’t a time machine or anything, she’s just that bad at hiding she’s from the future. The wires coming out of her head are live and she is fully unaware of the dangers that come with that.
Ambrose runs off of food like a normal person, but if you starved him he wouldn’t die. I don’t have a reason as to why he’s immortal yet, but that’s the gist of him. He has a burn scar on the left side of his face that he covers up with his hair. The right side of his head is shaved as he wears a cochlear implant (I don’t know if I drew it right). There is no implant on the other side, but that does not necessarily mean he can hear on that side.

What updates are coming in the future?

To touch on stuff from the last article, Streamer mode is probably not happening anymore. I still really want to get Japanese working but I’m just sort of wasting colab credits each attempt.
What I think I will be working on next are voicebanks based off of the Japanese Voice Actress, PJS, and TYC corpuses. I want to try and take the Japanese input audio, and successfully convert it into an English output audio.
I’m also working a bit more on Dunder Chinese and Dunder Polish. I didn’t know where to bring it up in this article but Dunder (The English bank based on my normal voice) has also released for Z Idol Talker today. I decided it wasn’t HQ enough for me to worry about it.

Thanks for reading!

Dunder_CV ACT II will release eventually I promise, probably for the 7th anniversary of ACT I at this point. I’m also working on a game demo somewhat in the Z Vocal Project canon, it was meant to release in February but I got very busy.