Fluid Vocal Synthesis System: UTAU .NET front-end

Oda Hikaru

Teto's Territory
It's good to breath for a while. Smell the roses, finish your pending works and when those works feels fine agan you come back. ^^
 

Ezkori

Teto's Territory
I saw this on a forum and I can not wait for this to be a thing. I despise the way the UTAU editor looks and works, and I think that this is 100% needed in this community.
 
  • Like
Reactions: KNΞMΛTCS

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
Since I didn't see any progress on the project since a long time, I'm posting to say that I forked the project on github and I'm working on fixing the issues of the current code in order to continue the project in the place of the creator
The migration will take some time though because there's a lot of bad code practices I noticed in the code so I'll redo a lot of things the "correct way" in order to make the code more modular and intuitive

I am planning too, to migrate the project to .NET Core, which would make it cross platform, but it will take some times before reaching this stade
 
  • Like
Reactions: Buck and Kiyoteru

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
In fact, I ended up starting immediatly by porting all the projects slowly to .NET Core, with the AvaloniaUI api for the UI in order to make cross platform development in the future easier
actually, oto didn't have that much difficulty to go to .net standard, wavmod either, but I'm working on FluidSys since this one contains some UI code to remake with avalonia since it previously was using WinForms which is a Windows only technology
 

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
I think that I'll try too to use the source code of World4Utau in order to create a resampler that would work well with it like, that would allow really fast rendering of vocals, which would be a lot better to work with it
Because what I always hated with UTAU is how slow it is to render vocals everytime, my wish would be to make something like Piapro or Vocaloid, where it renders in real time pretty quickly the vocals
 

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
That sounds really great, UTAU takes AGES to render, even when I use 'updated' versions of the wavtool and resamplers
That's because of the way UTAU renders vocals, it renders the vocals by calling the resampler executable note by note, it's a slow process, which is why resampler.dll tends to be faster, even though the quality of resampler.dll is...meh
I feel like it's possible to make a system that can render in real time, it will just require to think about how I'll do it.
But first of all, I must think about making something that works by fixing the software first lol
 

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
Yep. There are a lot of people who want huge things to be changed with how UTAU works - especially since it's been roughly 5/6 since the last Windows update (Using Win-UTAU here). There are a lot of things that could be fixed, but only Ayame/Ameya can do that and it's clear they no longer have interest in the program
Yep, I'm planning to do a big rewrite of the software, I feel like I want to make a kind of "UTAU 2.0" so with added features and performance
I'll just have to learn more about the way everything works since I don't have issues with programming in itself but I never worked on audio stuff programming in the past, once I'll have fixed the issues of the software and made sure everything works as intended, I'll start by trying to add the feature to render multiple utaus at once (a bit like piapro where we have one track per vocaloid and they all sing together)
 

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
I would love that. One request, if you're going to. Could you make it high contrast/add a high contrast mode for those who are hard of sight? One of my main issues with UTAU is it's really not accommodating to me. I have nystagmus, which means I can't focus, especially with black on white.
Since the whole fluid interface is black with white text, I wonder how I could do to make it easier to see
 

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
That sounds like the best way to go about it. With FVSS, though. there is the whole 'no one can distribute Defoko' thing. So, the question is - will it come without a voicebank, like Cadencii, or will it come with a different Default?
I dunno, Well, someone could probably volunteer to make the official default voicebank, but anyway, atm there's a lot more to think about than simply the default vb
 
  • Like
Reactions: Mitt64 and Kiyoteru

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
Well, someone could probably volunteer to make the official default voicebank
Feel free to contact me regarding this whenever you're ready. St. Defoko's has provided a voicebank for MelodiaSynth and we would be willing to assist with the development of other vocalsynths as well.
 
Last edited:

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
I'll post about it when it will be the focus, actually my main focus is to make it usable before adding a default vb, since having an unusable default vb would be a bit weird
 
  • Like
Reactions: Kiyoteru

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
After some thinking I ended up thinking about something,
Since FVSS wasn't optimized since the beginning and wasn't planned to be made on .NET Core at all, wouldn't it be better if I litterally remade from scratch a new one inspired by FVSS ?

unless people would prefer simply FVSS itself
It would be a bit longer at first but easier in the long run to make a completely new one since I would perfectly know the code I wrote, and I would be able to organize everything nicely
 
  • Like
Reactions: Nohkara

Pikachuk(SeleDreamsP)

Teto is love, Teto is life (Rin too)
Defender of Defoko
It probably would be better
Well, in this case, I'll probably program the new software in C++ instead and use the Qt api for the UI, because it would be more efficient in C++, and I could even port it to phones, so it would even allow people to use UTAUs on phones
I will post a new thread related to the new project to ask people about what kind of UI they would prefer for the workflow etc
 

Similar threads