UTSU - A cross-platform vocal synth frontend

Mougeki Mero

Defoko's Slaves
Defender of Defoko
One more question: When "$direct=true" is set, does UTAU also ignore the consonant length, preutterance, and overlap from the OTO file? If so, I'll be able to implement it without much trouble.
Nope, UTAU still takes the lenght and oto settings into account. ^^'
 

Lethe

Ruko's Ruffians
Supporter
Defender of Defoko
Thread starter
@Kiyoteru
Thanks for testing! My mac is having similar problems with Wine, so it may be an issue with High Sierra.

Nope, UTAU still takes the lenght and oto settings into account. ^^'
Okay, I have a good idea what the feature does now! Will add it to my task list.

im having problems opening USTs.
Most likely there was an error parsing the UST. Can you send me a copy of the UST you tried to open? Also, can you tell me your operating system (Windows 10, Mac High Sierra, etc) and which voicebank you were using?
 
  • Like
Reactions: Mougeki Mero

heta-tan

Genderless Goon
Global Mod
Supporter
Defender of Defoko
This seems really nice so far but I noticed that there doesn't seem to be a way to edit the vibrato or flags. Do you have plans on adding these features in the future?

Keep up the good work!!

*edit* I mean the flags for an individual note not the whole UST
 

Dangosan

Jellie Bellie Pete Rat Gummie Candie
Defender of Defoko
UTSU won't even open under Fedora 27. Instead, it starts to whine about UtsuMain.
 

Yoichi-Masaki

Ritsu's Renegades
Defender of Defoko
a few things after testing the current version on windows 7

1: most usts won't open. Aria opened. it did not play well, but it opened
2: macres likes to crash alot
3: i can't get other reasmplers to work
4; no oto limit as far as I can tell, which is amazing
5: i was in macress crashing hell for about 20 min after trying to play all of the aria ust. It doesn't stop after closing utsu T.T I thought it would never end


also sugjestions

1: if possible let the tuning be more like pitch trace plugin.

I'm super excited to see more of this
 
P

partial

Guest
I am super envious of anyone knowledgeable enough to try to create something new in the field of vocal synthesis programs/front ends. I've always wanted to contribute to the vocalsynth world in this way but coding and I simply do not mix, I couldn't even pass my visual basic class in high school and my final was a card matching game that didn't even work. Pretty sure I followed a guide for it, too.

I am cheering you on from the sidelines and paying close attention to this thread! It seems interesting so far. Maybe I'll try UTSU soon!
 
  • Like
Reactions: cloudyyskie

Dangosan

Jellie Bellie Pete Rat Gummie Candie
Defender of Defoko
I tested it on Ubuntu 16.04 and it's still not cooperative. Maybe it's OpenJDK's fault.
 

Lethe

Ruko's Ruffians
Supporter
Defender of Defoko
Thread starter
Progress report: I added source code to github for two new programs, one for reading .frq files and one for creating them. All the visuals and core features of the voicebank editor are done, but it needs a few more days of fine-tuning before I can release a new binary. I'm also including as many bug fixes as I can!

@heta-tan
This seems really nice so far but I noticed that there doesn't seem to be a way to edit the vibrato or flags. Do you have plans on adding these features in the future?
Yes! In terms of priority, they're below the voicebank editor and plugin loader, but above everything else.

@Dangosan
UTSU won't even open under Fedora 27. Instead, it starts to whine about UtsuMain.
I tested it on Ubuntu 16.04 and it's still not cooperative. Maybe it's OpenJDK's fault.
Huh. My friend who tried it on Ubuntu didn't report any problems. Looks like I'll need to download Fedora 27 and Ubuntu 16.04 on VirtualBox and test them out. Can you send me the exact error text that gets printed when UtsuMain fails?

@Yoichi-Masaki
Thanks so much for the detailed testing and bug reports! Thanks for the icon design as well! I'll forward that on to Sour, it should be small enough to fit inside an icon.
1: most usts won't open. Aria opened. it did not play well, but it opened
2: macres likes to crash alot
The USTs probably have parsing errors, and macres is probably getting stuck in an infinite loop somewhere. To debug this, I'll need to be able to replicate the error on my own computer. Can you sent me a link to the exact Aria UST you used, as well as a link to the voicebank(s) you played it with? Also, can you send me links to some of the USTs that failed to open?
3: i can't get other reasmplers to work
They worked on Windows 10, so this may be a Windows 7 issue. One thing you can try is run UTSU as a .jar file from the command line. All of the logs from the resamplers and wavtool will be printed into your terminal, so you can see exactly what type of errors they're throwing.
4; no oto limit as far as I can tell, which is amazing
Nope, there's no oto limit!
5: i was in macress crashing hell for about 20 min after trying to play all of the aria ust. It doesn't stop after closing utsu T.T I thought it would never end
aaah sorry!
TODO(Lethe): Cancel all in-progress resamplers and wavtools when UTSU closes.
1: if possible let the tuning be more like pitch trace plugin.
If possible, UTSU will support the pitch trace plugin itself! I'll try to implement some cool pitch editing into the program as well, though.

@partial
Thanks for your support! I'll try my best to get UTSU into a state where you can make great things with it.
:smile:
 

Dangosan

Jellie Bellie Pete Rat Gummie Candie
Defender of Defoko
Error: Could not find or load main class com.utsusynth.utsu.UtsuApp
 

Lethe

Ruko's Ruffians
Supporter
Defender of Defoko
Thread starter
Error: Could not find or load main class com.utsusynth.utsu.UtsuApp
I downloaded Ubuntu and tried for myself. It turns out Utsu is not compatible with OpenJDK, as OpenJDK does not support some libraries that Utsu relies on. However, I confirmed that Utsu works with Oracle Java (see download instructions here). My beta tester didn't catch this because he was *already* using Oracle.
 
  • Like
Reactions: sangv

Yoichi-Masaki

Ritsu's Renegades
Defender of Defoko
Ah! I forgot to respond to this! here is the aria ust. as for the other ust's I tried to open, I can't recall because there were quite a few. I know I tried to open a bunch of English ones
 

Attachments

  • ARIA_main - Copy.zip
    22.1 KB · Views: 2
  • Like
Reactions: Lethe

Lethe

Ruko's Ruffians
Supporter
Defender of Defoko
Thread starter
It's finally time for a new release! As usual, you can go here to find new Windows/Mac installers and a new JAR executable.

Updates:
  • Voicebank editor added!
    • Can edit name/author/description of a voicebank.
    • Can edit pitch suffixes for a multipitch bank.
    • Can edit oto configs, duplicate them, and delete them.
    • Table-based and waveform-based oto editors.
    • Can right-click on oto table row to regenerate the .frq file.
    • Can right-click the "Frq" column header to regenerate all .frq files.
    • Can right-click the waveform to play a raw lyric.
    • Can save voicebank in a format readable by UTAU and UTAU-Synth.
    • Unicode lyrics automatically left out of oto.ini files used by UTAU.
  • Multiple tabs added!
    • Each tab can hold a song (.ust file) or a voicebank.
    • Unsaved voicebank changes will apply to any songs using that voicebank.
    • Close a voicebank without saving, and songs will reload the voicebank from saved version.
    • Warning appears when closing a tab without saving.
    • Zooming in or out applies across all tabs. Same for language changes.
    • Stops user from opening same voicebank/file in more than one tab.
  • Bug fixes.
    • Less lag than before when editing long USTs.
    • Any running resamplers/wavtools will be destroyed when Utsu closes.
    • Notes no longer remain highlighted when removed from highlighted section.
    • Invalid notes can no longer be highlighted.
    • Removed lots of parsing errors that stopped people from opening USTs.
    • Removed several bugs in writing USTs and voicebanks.
Here's a picture of the new voicebank editor:
voicebank_screenshot.png


Future plans:
Some voicebank editor features (editing .frq files, switching voicebank image, and a tree view of .wav files that users can drag to make new oto entries) are planned but de-prioritized for now. The remaining "core" features IMO are plugins, note properties, and a detailed vibrato editor. I'll work on those first.
 
P

partial

Guest
It's finally time for a new release! As usual, you can go here to find new Windows/Mac installers and a new JAR executable.

Updates:
  • Voicebank editor added!
    • Can edit name/author/description of a voicebank.
    • Can edit pitch suffixes for a multipitch bank.
    • Can edit oto configs, duplicate them, and delete them.
    • Table-based and waveform-based oto editors.
    • Can right-click on oto table row to regenerate the .frq file.
    • Can right-click the "Frq" column header to regenerate all .frq files.
    • Can right-click the waveform to play a raw lyric.
    • Can save voicebank in a format readable by UTAU and UTAU-Synth.
    • Unicode lyrics automatically left out of oto.ini files used by UTAU.
  • Multiple tabs added!
    • Each tab can hold a song (.ust file) or a voicebank.
    • Unsaved voicebank changes will apply to any songs using that voicebank.
    • Close a voicebank without saving, and songs will reload the voicebank from saved version.
    • Warning appears when closing a tab without saving.
    • Zooming in or out applies across all tabs. Same for language changes.
    • Stops user from opening same voicebank/file in more than one tab.
  • Bug fixes.
    • Less lag than before when editing long USTs.
    • Any running resamplers/wavtools will be destroyed when Utsu closes.
    • Notes no longer remain highlighted when removed from highlighted section.
    • Invalid notes can no longer be highlighted.
    • Removed lots of parsing errors that stopped people from opening USTs.
    • Removed several bugs in writing USTs and voicebanks.
Here's a picture of the new voicebank editor:
voicebank_screenshot.png


Future plans:
Some voicebank editor features (editing .frq files, switching voicebank image, and a tree view of .wav files that users can drag to make new oto entries) are planned but de-prioritized for now. The remaining "core" features IMO are plugins, note properties, and a detailed vibrato editor. I'll work on those first.
You are a saint for vocalsynth users, I want you to know.
I have absolutely always wanted to create something like this but I couldn't even pass my Visual Basic class in highschool...
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
Otsukare!
I'll assume for now that my previous feedback on the UST editor will still be taken under consideration in the future.
How does one open a voicebank in the VB editor? Everything is grayed out, so I'm not sure what I'm meant to select.
Will the audio visualizer at the bottom also have a spectrogram mode, and zoom controls? Spectrogram view is essential to me, since a waveform isn't always clear enough.
I'm looking forward to note properties and vibrato control!
 
  • Like
Reactions: HIRATELIER

Lethe

Ruko's Ruffians
Supporter
Defender of Defoko
Thread starter
@partial
You never know, it could have just been a bad experience in that one class. If you do want to get back into programming, I've heard good things about Codecademy.

I'll assume for now that my previous feedback on the UST editor will still be taken under consideration in the future.
No worries, all the features you recommended are still on my task list!

How does one open a voicebank in the VB editor? Everything is grayed out, so I'm not sure what I'm meant to select.
Are you trying to open .utau files maybe? Utsu only supports directory-based voicebanks right now, so you would need to basically copy the contents of the .utau voicebank into a new folder and select that folder when you open a voicebank.

Will the audio visualizer at the bottom also have a spectrogram mode, and zoom controls? Spectrogram view is essential to me, since a waveform isn't always clear enough.
Is that what the .spef files are for? Sure, I'll add spectrograms and zoom controls to the task list, at the same priority as the other voicebank features.

Utsu creates both (foldered) oto.ini files and a (non-foldered) oto_ini.txt file. All of the oto data will be saved in both formats except for the lyrics/filenames not supported by Shift-JIS, which will only be saved in the oto_ini.txt file. That way, UTAU-Synth will be able to read the entire voicebank while UTAU will be able to read all parts that can be rendered in Shift-JIS. Prefixmaps work the same way--Utsu will create a Unicode prefixmap file and a Shift-JIS prefix.map file. The description, name, and author, meanwhile, will be saved in Shift-JIS by default so that UTAU can read them. If any characters in those fields are Unicode-only, the files will be saved in Unicode instead.

TL:DR You can make Unicode voicebanks on Utsu and open them on UTAU-Synth, but UTAU won't be able to use them properly.
 
Last edited:

Similar threads