-
Notifications
You must be signed in to change notification settings - Fork 0
FAQ
Frequently Asked Questions about Hiro UST Generator.
Hiro UST Generator is a tool that automatically creates UST/USTX files (used in UTAU voice synthesis) from Japanese lyrics. It handles:
- Automatic phonemization of Japanese text
- Intelligent melody generation
- Voice-specific pitch ranges
- Accent pattern support
- Fine-tuned parameter adjustment
No! The default settings work well for most cases. The tool handles:
- Scale selection
- Pitch range assignment
- Note duration calculation
- Accent pattern application
Just enter your lyrics and click Generate!
Currently, Japanese is the primary language. Romaji input is supported (e.g., "kyakkya" for "きゃっきゃ").
[Check your license terms]
A: Windows only currently. macOS and Linux support coming in future versions.
A: OpenUtau (recommended) for best results. Avoids character encoding issues with UTAU.
A: See Installation guide.
A: Yes, Python 3.8 or higher. Download from python.org
A: Yes! It works on Windows, macOS, and Linux.
A: About 100 MB for the application and dependencies.
A: In the "🎵 Song Lyrics" section, type Japanese text:
[Verse 1]
きゃっきゃ うれし
[Chorus]
いたみ いたみ
You can use:
- Hiragana: あいうえお
- Katakana: アイウエオ
- Romaji: a i u e o
A: It controls randomness. The same seed always produces the same melody.
Use the 🎲 button to try different melodies with one click.
A: Multiple options:
- Change Seed: Click 🎲 Randomize button
- Change Scale: Try different musical scales
- Change Accent: Different accent patterns = different melodies
- Adjust Curve: Bias melody up or down
- Disable Motif: Turn off "Motif Memory" for more variety
A:
- Voice (Soprano/Alto/Tenor/Bass): Sets the key range (high/low)
- Scale (Major/Minor/Pentatonic): Determines which notes can be used
A: Click "💾 Save Preset" to save current settings. Load them later with "📂 Load Preset"
A: Not in the current version. Settings are live, so adjust and preview before generating.
A:
| Format | Best For | Stability |
|---|---|---|
| UST | UTAU (Windows) | |
| USTX | OpenUtau (all platforms) | ✅ Stable (recommended) |
A: Use USTX + OpenUtau for best results. UST in UTAU sometimes produces Chinese characters.
A: Only with UTAU/OpenUtau which support UST format. OpenUtau recommended for stability.
A: Japanese pitch patterns that make words sound natural:
- Heiban: Flat tone
- Atamadaka: High → Low
- Nakadaka: Low → High → Low
- Odaka: Low → High
Setting the correct accent makes lyrics sound more natural.
A: Generally:
- Greetings: Atamadaka
- Common words: Heiban or Atamadaka
- Polite speech: Nakadaka
- Formal: Odaka
Use preview to test!
A: It remembers pitch patterns from previous notes and continues them, making melodies more coherent and less random.
A: It generates intelligent melodies based on:
- Text structure
- Phoneme characteristics
- Voice range
- Musical scale
Sounds more natural than pure random selection.
A: It's the lead-in time (in milliseconds) before a note starts. Adjust for smoother transitions between notes.
Typical values: 20-40ms
A: How much the current note overlaps with the previous one. Creates smoother legato transitions.
Typical values: 5-15ms
A: Volume level of the note.
Range: 0-200 Typical: 80
A: The pitch envelope preset that affects how the note sounds:
- Pop: Balanced
- Opera: Sustained
- Whisper: Soft
- Belt: Powerful
Experiment to find your preference!
A: Use "Length Variation" (0.0 = fixed, 1.0 = maximum randomness) and "Base Length" (default 240 ticks).
A: How the tool represents sounds:
- Japanese: Uses Japanese phonetic system
- Hepburn: Romanization system
- Wapuro: Keyboard input romanization
- English: English phonetic system
Most users should use Japanese mode.
A: Yes, click "👀 Preview Phonemes" to hear individual sounds.
A: The basic unit of sound in Japanese. Examples:
- "あ" = 1 mora
- "きゃ" = 1 mora
- "きゃっきゃ" = 3 moras (きゃ + っ + きゃ)
A: Check:
- Python 3.8+ is installed
- All dependencies installed (
pip install -r requirements.txt) - Tkinter is available on your system
- Try the command line test in Installation
A: Try:
- Click "Gen" again
- Choose different settings
- Save to different location
- Check disk space
A: Check:
- Lyrics are valid Japanese
- Phoneme mode is set correctly (use Japanese)
- System audio is working
- Try a different preview text
A: Save your settings as a preset:
- Configure all parameters
- Click "💾 Save Preset"
- Name your preset
- Load it next time with "📂 Load Preset"
A: Try different settings:
| Problem | Solution |
|---|---|
| Too random | Enable "Motif Memory" |
| Too robotic | Enable "Lyrical Mode" |
| Wrong scale | Change "Scale" setting |
| Wrong key | Change "Voice" setting |
| No emotion | Enable "Vibrato" envelope |
A: Usually 1-5 seconds for a typical song. Depends on:
- Number of notes
- Computer speed
- Complexity of settings
A: Check:
- Disable "Lyrical Mode" for faster generation
- Disable "Motif Memory"
- Use simpler scale (pentatonic vs chromatic)
- Close other applications
A: Not in the current GUI version. CLI interface coming soon.
A: Works with UTAU 2008+ and OpenUtau (all versions).
A: No, VOCALOID has its own file format (VSQx). Use UTAU or OpenUtau.
A: Generates UST files compatible with any voice bank. You assign the bank in UTAU after generation.
- Check: User Guide
- See Examples: Usage Examples
- Browse: Architecture for technical details
Hiro UST Generator v0.2.0
Windows Only | Best with OpenUtau | Quick Start • User Guide • Home
- Home - Start here
- Quick Start - 5-minute guide
- User Guide - Complete manual
- FAQ - Q&A
- Architecture - How it works
- Usage Examples - Code samples
Need help? → FAQ
Ready to use? → User Guide