Suggestions For Development Of Vocal Synthesis Program?

PrinceofHades

A wandering soul
Defender of Defoko
Alright. Let's get this party started.
I love coding. I'm going to college to major in Software Development in the fall. And UTAU looks like it's never gonna be updated again.
And after reading @SarCaustic 's eye-opening thread on that subject, I made a decision.

I'm going to create a new vocal synthesis program that will both replace and IMPROVE on UTAU.
Now, let me clarify that: The program that I want to create will be compatible with old UTAU voicebanks (cuz who wants to rerecord everything just so it will work with a new program? Not me.) AND provide new features that UTAU does not possess that are needed/would be awesome to have.

So. Here's what I need help with.
1. What should I use to code this program? (i.e., Visual C#, JavaFX, etc.)
I'm wonderful with languages, and programming languages are no different, so if you suggest one I don't know, no problem. Chances are I can learn it quickly.
EDIT: I've done some research, the decision is between JavaFX and QT with C++.
While my dad uses Linux, nobody in my house has a Mac. I need a Mac and Linux user to tell me which they think would work best. I intend for this to be cross platform, with no differences between the Windows, Mac, and Linux versions.

2. Name. What one earth do I call this thing?

3. Priority features. It's not going to look pretty (yet!) but what features are the MOST important that we absolutely CANNOT do without?

4. Beta-testers. I would prefer to do all the coding myself (although I may ask for help with concepts I do not understand) so what I need are people I can rely on to, when I have a new stable build, test it out and make sure I didn't miss any pesky bugs.

5. An answer to this question: Do I need a Defoko/default voice? Because I'm thinking in the early stages, it would be best to work mostly on making sure we have basic rendering capabilities and tuning, and so being able to choose/add a different voice wouldn't be much good if we couldn't do that.

I will update this when needed. (Such as, first stable build, new questions, or when I need to decide which feature to add first.)
In addition, this is a long term project. It may be months before I have a stable build to show you. I refuse to accept anything less than the best I can do.

If requested, I'll add contact information.
 
Last edited:

Zoku

making doper vocaloid music than the rest
Defender of Defoko
1) C++ (or any OS universal coding language)
2) Worry about that later. That should be an afterthought.
3) Engine, tempo changes, pitch changes, blending, GUI, configuration of voices
4) I could beta test whenever possible
5) You should have some sort of vocal to work with.
 
  • Like
Reactions: PrinceofHades

bio

VocalSynth Enthusiast
Supporter
Defender of Defoko
I'm open to beta testing.
Also I am open to creating voicebanks for your program. (maybe a custom dictionary?)
As @Zoku said, use C++, and name should come later.
 

Zoku

making doper vocaloid music than the rest
Defender of Defoko
^^The dictionary should come after the basics are set, as building a dictionary would open a whole other can of worms, including phonemes, approximation of lengths of notes depending on the VB, as well as other variables.
 
  • Like
Reactions: PrinceofHades

PrinceofHades

A wandering soul
Defender of Defoko
Thread starter
1) C++ (or any OS universal coding language)
2) Worry about that later. That should be an afterthought.
3) Engine, tempo changes, pitch changes, blending, GUI, configuration of voices
4) I could beta test whenever possible
5) You should have some sort of vocal to work with.

C++ does sound good, based on what I know about it.
You're right, name should be an afterthought. I tend to bite off more than I can chew.
I'll probably start with the GUI, engine, and configuration of voices. Tempo and pitch changes be easier to implement after I make sure that we can render notes. Can't change pitches or tempo if we have no notes.
Thank you! :smile:
That's what I thought. I'll grab a monopitch CV for that purpose (or just throw in a voicebank that only has vowels. You know what, we'll start with the vowels. That way, I can implement the voicebank and note editing with a minimum amount of samples to deal with. Then, once the basics of that are set, I'll work with consonants.)

I'm open to beta testing.
Also I am open to creating voicebanks for your program. (maybe a custom dictionary?)
As @Zoku said, use C++, and name should come later.

Thank you!
I was planning on using my own UTAU as the first default for simplicity's sake, but thank you for the offer.
As @Zoku said, we are nowhere near ready for a custom dictionary. That is going to be a hell of a lot work. While I do intend to implement a dictionary for ease of use, that's probably going to happen months and months and months later. ^^;
 

Similar threads