Axfc is a jerk. Here's my post on how to get downloads from there to work.

Multi-pitch CVVC banks do not work properly with the shareware A for automatic button!! Any articles where I complain about CVVC banks being broken is my own fault for not figuring it out sooner!!

Wednesday, January 15, 2025

Wiki Wednesday #306 - Yori Igarashi / 五十嵐ヨリ

 whoo! Progress!

Wiki Wednesday #306 - Yori Igarashi / 五十嵐ヨリ

Yori is a shy twenty year old trans man who has a secret online life where he is a famous singer.

Who is Igarashi Yori?

Art from wiki


.

Hey guys! So, I've run into a bit of an issue. I can't do anything creative. I just don't know how to make up stories or make music.

For the most part, that's fine. There is no creativity needed for my current favorite hobby: labeling data for Diffsinger.

The issue is how I have this blog set up. In November, I'm going to do my best to push through and write a novel. But I don't just need a third person perspective novel - I also need this first person perspective feature. I have ideas... But I can't put them into words.

So instead, I'm going to put what I know about Diffsinger into words until I feel like I have enough leeway to not let this blog run out of content!

Diffsinger is crazy. It is amazing. I was always really lenient towards the realism of vocal synths. It took basically nothing for me to be convinced that Miku was a person. But I'm getting older and I'm getting exposed to more stuff and I'm losing that leniency. I can't help but pick out flaws that make me feel like it's not worth it to create those things. 

But Diffsinger?! Diffsinger changed all of that.

Something that I've sworn myself to is that I'll never take vocal synths seriously again. A lot of people take them seriously, but not me. So I just want to experiment and have fun! 

There's a lot of stuff to test and learn, which is exciting! But a big part of that is also not putting pressure on any one model to be perfect or anything. You can record a lot of data and have a Japanese model that mimicks your voice almost exactly, or you can throw together a six minute English model that isn't perfect but still kicks the butt of old school VOCALOID any day of the week! 

I'm always out there searching for people to record for me. I'm recording my own test banks, but what I really want is a ton of fun voices that wouldn't otherwise exist!

Though, maybe ironically, I won't actually feature Diffsinger voices on this blog. Even if there were hundreds of them to play with, I write this blog on my cheap little Windows tablet. It's barely strong enough to open OpenUtau! I know I could technically use it on a different PC, but I just really like the set up that I have. 

Also... The banks are usually 400 megabytes a piece! I have great Internet and can handle downloading that much, but it's still pretty crazy. You can train in a way that makes twenty megabyte models, but even then, all I know is that I don't know what any of that means or how you do it!

So, because I just don't have much brain ability now, I hope you all have at least one question answered by my upcoming talks about Diffsinger. The software is so much fun! 

How Is Yori Igarashi's bank?

Yori has a two pitch CV-VV bank. He has a nice voice!


Where Can I download Igarashi Yori?

You can find him on his official site. He is cool!

No comments:

Post a Comment