Making a voicebank using audio files?

[HAZPUNK]

(this first paragraph is mostly just unimportant context, you can skip to the second one for the actual post)

I've been getting into a lot of speech-synthesis and utau related things lately, trying (and failing miserably multiple times) to make my own speech synthesizer using python. Eventually I decided to try openutau, and I am amazed by how much better it is then trying to decipher an impossible language like cmake. I've used it enough to be comfortable with how the interface works and how to import voicebanks, and for a song I'm making, I have decided to make my own voicebank.

I've followed a tutorial on how to make my own using OREMO, which has been pretty helpful, up until getting to the actual recording itself. I have a reclist and bgm for the voicebank, however, it uses samples from some speech synth toy thing from 1978 (The speak-n-spell) which are tuned using FL studio to F4 (which has been the most convinent key to pitch correct to) using some effects. I got 2 samples done before realizing that OREMO doesn't allow dragging and dropping audio files, and afaik, importing samples at all. The same is true for Akorin and Recstar. Is there a software for importing audio files or a way to do this in OREMO/Akorin/Recstar? Thanks!

P.S this is my first post. Hello!

 

[heynotloid]

I think if you rename the file to the phoneme in OREMO's reclist, it will pick it up. What an amazing idea you have! When you finish it, I'd be happy to see it sing!

 

[SunnyWolves]

Oremo isn't necessary for creating voicebanks. All you have to do is name the sample and put it in a folder before using it with utau or openutau. You can also oto (label) it in a program such as SetParam

 

[HAZPUNK]

Oremo isn't necessary for creating voicebanks. All you have to do is name the sample and put it in a folder before using it with utau or openutau. You can also oto (label) it in a program such as SetParam

Sorry for the late response, thanks so much!

 

[SunnyWolves]

I must ask, why did you bother getting recording bgm if you're not recording the samples? Was it just a misunderstanding or because info sources urged you to?

 

[HAZPUNK]

I must ask, why did you bother getting recording bgm if you're not recording the samples? Was it just a misunderstanding or because info sources urged you to?

I thought reclists were needed to make an oto, to be safe I used it so everything would line up.

 

[SunnyWolves]

Nah, a reclist just means "recording list", aka what you need to record to make a voicebank. If you have the samples, you can move right on to oto.

 

[Halo]

You need the reclist, but as long as you make the samples according to a reclist it can be recorded, spliced, or sampled in anything. OREMO is just a stripped down recording program that makes sure your recordings come out how UTAU specifically expects them and saves time naming files haha, but it's only of benefit if you're recording.
Just remember to save your final files as 16 bit 44100khz mono WAV files! UTAU and OpenUtau both struggle with almost any other configuration If you plan on using a base oto also be sure to name the files accurately (probably best to copy paste from the list).

 

[HAZPUNK]

You need the reclist, but as long as you make the samples according to a reclist it can be recorded, spliced, or sampled in anything. OREMO is just a stripped down recording program that makes sure your recordings come out how UTAU specifically expects them and saves time naming files haha, but it's only of benefit if you're recording.
Just remember to save your final files as 16 bit 44100khz mono WAV files! UTAU and OpenUtau both struggle with almost any other configuration If you plan on using a base oto also be sure to name the files accurately (probably best to copy paste from the list).

Do I need the same amount of samples the reclist has?

 

[Halo]

Do I need the same amount of samples the reclist has?

UTAU is really freeform, so technically no, you can do anything you like. If you want it to work without issue and this is your first time editing an oto I would recommend it. That said, if you have other technical experience, it's pretty simple to just edit the lines in a premade base oto.ini that reference the files you don't have; it just may not sound as good or work without more manual editing, which is up to your own discretion.

If it doesn't come with a base oto.ini then you can genuinely just stop at any point and resume when you've either tested to your satisfaction or encounter a problem caused by the missing files.... I suppose technically you could even do that with a base oto.ini, you'd just risk more program crashes or infinite loading haha.