Requesting Advice for "Old" UTAU User

[ali_cee]

To preface this, I've been using UTAU since around 2012 (my first VB was CV with no oto.ini around that time, lol). I've been using OpenUTAU because it's a bit more intuitive and unfortunately due to visual issues becoming more functionally useful for me. The VB I'm using is CV from 2013 that I created, and I'm using it for sort of a personal project I guess? It's a bit hard to explain... （；´Д｀）

Essentially, I'm running into the issue that the VB I recorded while tunable and not uh... completely terrible is missing some key recordings for certain songs. While this wouldn't be a problem I just... don't sound like that anymore... and can't emulate the voice that I used prior. I'm worried that the maturing of my voice will derail the project a bit, as it won't sound similar or close enough to be overlooked.

For Vocal Reference:
The Original VB in OpenUTAU with USTX I made & "tuned" (Not the best, but I was very tired) [This is part of the 'Project']
My actual/current voice. Speaking Japanese for at least some idea how it might sound? [Japanese was a Google Translation, may not be accurate.]

I was also looking into recording VCV because, I am now an adult, and can functionally understand how to do so. But I'm getting a bit confused because I'm seeing a lot of different reclists floating around. I think I've got one that I want to try to use (Single Reclist - No Nonsense Japanese VCV), but there are a few things I'm not familiar with. I don't recall using OREMO ever before (my original VBs were recorded using a mic & Audacity splitting/noise reduction), and I'm seeing that there's now automation for oto.ini files (which I've struggled to get operating properly even with ANSI conversion). It doesn't help that I'm trying to understand if there's a staple English VCV/Reclist (or if I should even bother/attempt it).

I guess I'm looking at possibly where to Re-"start" with this VB (Advice, Recommendations), and if anyone has any tips/tricks that might simplify the process. This way I can focus on the project at hand again.

I'm sorry if this isn't the right place to post this, I was debating either UTAU Discussion or UTAU Help. I figured this might be the best spot.
(*_ _)人

 

[Kiyoteru]

If preserving the old tone of voice is an absolute requirement for you, you can use the original recordings to create a model with something like Diff-SVC or sovits. Then, after recording a new voicebank, you can convert the tone of your recordings to sound like the old one. If this isn't essential you can disregard this step.

For recording voicebanks, you can get plenty of reclists and base otos from this website.

JP Reclists | Wasteland UTAU

wastelandutau.neocities.org

To record, I recommend using OREMO or RecStar, which will let you open the recording list file and will help automatically save each recording with the correct filename and format.

After adding the basic oto.ini to the voicebank you'll need to fine tune and adjust it to fit. I highly recommend using vLabeler.

Home

An open source voice labeling application

vlabeler.com

For japanese voicebanks, I recommend using vLabeler's built-in OTO template generators instead of a base OTO.ini file.

 

[ali_cee]

If preserving the old tone of voice is an absolute requirement for you, you can use the original recordings to create a model with something like Diff-SVC or sovits. Then, after recording a new voicebank, you can convert the tone of your recordings to sound like the old one. If this isn't essential you can disregard this step.

For recording voicebanks, you can get plenty of reclists and base otos from this website.

JP Reclists | Wasteland UTAU

wastelandutau.neocities.org

To record, I recommend using OREMO or RecStar, which will let you open the recording list file and will help automatically save each recording with the correct filename and format.

After adding the basic oto.ini to the voicebank you'll need to fine tune and adjust it to fit. I highly recommend using vLabeler.

Home

An open source voice labeling application

vlabeler.com

For japanese voicebanks, I recommend using vLabeler's built-in OTO template generators instead of a base OTO.ini file.

Yeah, it's not an absolute requirement but I'm concerned about it. I guess I can create a smaller VCV bank first and try to mimic it as closely as I can with the resources you sent (though OREMO seems to refuse to work now, converting the hiragana to ANSI/UTF-8 symbols so I'll check RecStar). And if all else fails I'll try out Diff-SVC or sovits. I was hoping I wouldn't have to go as far as using an SVC or sovits but it is what it is ¯\_(ツ)_/¯

I did however try (and fail) at using the CV -> VCV YT tutorial (utilizing the UST files and UTAU original sounds to create VCVs). She sounded kind of worse than her CV, ngl.

I also hopped on a Discord to get feedback/advice, one of the recommendations was using the g- flag. So if I can get a VB at least somewhat similar in tone then I might be able to try that.

Thank you though for the resources, at least this helps get me on a more grounded platform to jump from! ^^