What do you think of an open source music vocal synthesizer?

What do you think of an open source music vocal synthesizer?

  • That sounds like a terrible idea.

    Votes: 0 0.0%
  • What's open source?

    Votes: 0 0.0%

  • Total voters
    20
  • Poll closed .

WendytheCreeper

(>☉ ͡ヮ☉<)
Defender of Defoko
I sometimes wonder about an open source program that's like Vocaloid or UTAU and the like. While in UTAU we can make our own voicebanks, wouldn't it be nice if other programmers could alter the code and everyone could chip in to make a better program?

What I mean is, it'd be super nice if there were a synthesizer that were not only free, but others can improve upon the source code and the program freely, kinda like programs such as GIMP or Blender or Linux. I'm pretty sure with most of the synthesizers that I know of, it'd be against their terms of use to try to alter the code(do correct me if I'm wrong)

I'm not really planning on making anything like this(I don't know an ounce of code, and it's rather ambitious), but I was wondering what others think of the idea, and if there are any such programs out there.
 
  • Like
Reactions: na4a4a

수연 <Suyeon>

Your friendly neighborhood koreaboo trash
Supporter
Defender of Defoko
I'd be all for it if there were a synthesizer that was free, allowed making banks, allowed for other languages besides English and Japanese, and allowed source code to be improved upon. I admire the fact that Ameya made UTAU possible and that Cadencii tried to bridge the Vocaloid-UTAU gap, but the development of both simply isn't satisfactory and complimentary to the sheer growth that the fandom has gone through since its inception.
 

nirereru

Teto's Territory
It'd be super cool to have an open source synth that anyone could improve. I almost wish UTAU was that way.
 

shugo974

Momo's Minion
It would be nice to have, expectantly for the users who mostly use Linux as their everyday operating system. Although I read that someone was working on it, though last I checked there wasn't any news about it for over a year :/
 

XJM

Teto's Territory
UTAU is pretty open even without an Open Source license. You can create other things like resamplers or wavtools and I do think it is possible to access the source code though you may have to pay to gain access to it and it would probably be strictly overseen by Ameya or whoever looks at that stuff.

Though I like the idea of an Open Source software like UTAU and VOCALOID. You would likely need a good team to keep the project rolling. It would be cool to see an Open Source Vocal Synth.
 

WendytheCreeper

(>☉ ͡ヮ☉<)
Defender of Defoko
Thread starter
UTAU is pretty open even without an Open Source license. You can create other things like resamplers or wavtools and I do think it is possible to access the source code though you may have to pay to gain access to it and it would probably be strictly overseen by Ameya or whoever looks at that stuff.

Though I like the idea of an Open Source software like UTAU and VOCALOID. You would likely need a good team to keep the project rolling. It would be cool to see an Open Source Vocal Synth.

This is why it'd be very difficult to have an open source program...many people turn to open source simply because they don't have the money to afford the big program. We may choose GIMP over PS because PS is so expensive. UTAU and some other synths fill the niche for a free alternative for Vocaloid, so one would wonder "why another voice synth?" That, and I'm pretty sure the average fan has no interest in getting involved in the nitty-gritty of of programming.

It makes me wonder: how many fans out there of Vocaloid and UTAU and stuff primarily use Linux? Even though Mac users are gaining some love with things like UTAU-Synth and Vocaloid Neo, what about the Linux users? There's not a huge crowd of them...I can name only a few Linux based artists off the top of my head(Vasco Basqué, David Revoy, and Tyson Tan come to mind) and those are just visual artists there. I can't name a single person who is important to the Vocaloid/UTAU/whatever people who use Linux primarily...
 

MillyAqualine

Ritsu's Renegades
Defender of Defoko
It'd be totally great ! and if it was featuring an Unicode patch, that'd be even much more wonderful <3
 

WendytheCreeper

(>☉ ͡ヮ☉<)
Defender of Defoko
Thread starter
I've heard of Cadencii before, though I've never gotten it to work for me. I've never heard of QTau though.
 

isengaara

Teto's Territory
Defender of Defoko
eCantorix is free software, and unlike UTAU it is able to sing in German. Creating your own voice preset is fairly simple. It is not easy to install eCantorix, but once you get eCantorix working you can synthesize songs in almost any language.

There is also the Fesival Speech Synthesesis System which has its own Singing Mode. Both Sinsy and MAGE pHTS use the well known HTS engine which is free software. The free verion of Sinsy does not ship with an English Voicebank, only Japanese is supported.

As I use Trisquel GNU/Linux I don't install any nonfree software. Currently I can confirm that eCantorix and Sinsy can be used on GNU/Linux. I also use Cadencii with vConnect-STAND for which I created a PPA. If you don't have Cadencii you could aso use rosegarden or musescore.[/URL]
 

WendytheCreeper

(>☉ ͡ヮ☉<)
Defender of Defoko
Thread starter
eCantorix is free software, and unlike UTAU it is able to sing in German. Creating your own voice preset is fairly simple. It is not easy to install eCantorix, but once you get eCantorix working you can synthesize songs in almost any language.

There is also the Fesival Speech Synthesesis System which has its own Singing Mode. Both Sinsy and MAGE pHTS use the well known HTS engine which is free software. The free verion of Sinsy does not ship with an English Voicebank, only Japanese is supported.

As I use Trisquel GNU/Linux I don't install any nonfree software. Currently I can confirm that eCantorix and Sinsy can be used on GNU/Linux. I also use Cadencii with vConnect-STAND for which I created a PPA. If you don't have Cadencii you could aso use rosegarden or musescore.[/URL]

That German sample is pretty nice! While it sounds robotic, it's pretty smooth, though I don't don't know any German so I can't say anything about pronunciation.

I should really play around with Cadencii. I can't really get it to actually make sound though.
 

isengaara

Teto's Territory
Defender of Defoko
Getting a new program to sing is not easy, for me It took about three month to get eCantorix running, so I can render my own songs.
 
  • Like
Reactions: WendytheCreeper

na4a4a

Outwardly Opinionated and Harshly Critical
Supporter
Defender of Defoko
UTAU is pretty open even without an Open Source license. You can create other things like resamplers or wavtools and I do think it is possible to access the source code though you may have to pay to gain access to it and it would probably be strictly overseen by Ameya or whoever looks at that stuff.

Though I like the idea of an Open Source software like UTAU and VOCALOID. You would likely need a good team to keep the project rolling. It would be cool to see an Open Source Vocal Synth.
But the editor and the formats aren't open, which is the stinker.

If those were open, someone could of made native phoneme or vc support a thing already. (And that would take it so much further)
 
  • Like
Reactions: WendytheCreeper

WendytheCreeper

(>☉ ͡ヮ☉<)
Defender of Defoko
Thread starter
But the editor and the formats aren't open, which is the stinker.

If those were open, someone could of made native phoneme or vc support a thing already. (And that would take it so much further)
Now that I reflect on the great things we're doing in UTAU now, with VCCV just around the corner(or at least in the near foreseeable future), imagine what we could do if UTAU were completely open source?

System support could be so much better, since there could be more people making a better UTAU-Synth, and perhaps it could bring the attention of Linux users too, which seem to be kinda nonexistent as far as people who use vocal synths go.

What if we could change how the oto.ini works, too? How could we implement features that could benefit certain types of samples? CVs? VCs? VCVs? And anything else?

And how could we change how UTAU renders voices? What flags could we improve on? Could we change what makes up an .ust?
 

na4a4a

Outwardly Opinionated and Harshly Critical
Supporter
Defender of Defoko
If otoing had vc support, that would pretty much satisfy a lot of people (even though otoing vc isn't that big of a deal really, just can't stretch the vowel....but then again...why would you need to?)

If we could use a pre-established dictionary (like that used for tts on your computer already) and just have that sort of drive the samples, that would be pretty great imo.

What would be nice is more continuity with the formats....better handling of files between platforms...

Edit: like the microsoft speech dictionary. Just take the vowels and cross reference them to the closest vowel in the voicebank.
A lot of what we have now is thanks to reverse engineering and such (from what I've read) and open sourcing would take that layer of pain out of the workflow.
 
Last edited:
Similar threads

Similar threads