jheysonjhoelmamanihancco-lang/audirs-vox-lab

An open-source, Rust-powered voice training companion that turns raw vocal practice into a visual, measurable, and safe journey of self-discovery. Inspired by the phonetics-first philosophy of audirs, VocalForge reimagines what a modern voice trainer can be — a studio, a mirror, and a mentor, all in one.

• Overview
• The Philosophy Behind VocalForge
• Why VocalForge Exists
• Feature Highlights
• Visualization Engine
• Safety-First Training Model
• Responsive Interface
• Multilingual Support
• Always-On Guidance
• Architecture & Tech Stack
• Performance Notes
• Accessibility & Inclusive Design
• Privacy & Data Ethics
• Roadmap
• Contributing
• Community Guidelines
• SEO & Discoverability Notes
• Frequently Asked Questions
• Disclaimer
• License

VocalForge is a Rust-based desktop and web companion for vocalists, voice actors, speech therapists, gender-affirming voice practitioners, and curious learners who want to understand their instrument from the inside out. It listens, analyzes, and paints a live picture of your voice — formant frequencies, vocal weight, pitch contour, resonance balance — and translates those numbers into visuals you can actually feel.

Where most tools stop at a pitch meter, VocalForge builds an entire atelier: a workspace where your breath, resonance, and tone become shapes on a canvas. Think of it as a blacksmith's forge for the voice — heat, pressure, and patience turning raw material into a crafted instrument.

Every voice is a fingerprint. No two are identical, and no single "ideal" exists. VocalForge was built on three principles:

1. Visual clarity beats guesswork. If you can see your formants shift, you can learn to control them.
2. Safety is not optional. Pushing the voice without feedback is how strain happens. VocalForge coaches you to train within sustainable boundaries.
3. Progress is personal. Metrics are yours, for your own journey — never for comparison against someone else's ideal.

The project borrows conceptually from earlier explorations in formant and pitch visualization, then expands them into a full-spectrum training environment.

Most voice tools are either clinical and cold, or playful but shallow. VocalForge sits in the sweet spot: rigorous enough for serious practice, warm enough to use daily. It exists because:

• Learners deserve real-time feedback without a coach peering over their shoulder.
• Voice work is deeply personal, and often deeply emotional — the tool should be gentle.
• Existing software often ignores safety, weight, and resonance in favor of raw pitch.
• Rust gives us real-time performance without the memory overhead of heavier stacks.

• 🎚️ Live formant tracking (F1, F2, F3) with smooth interpolation
• ⚖️ Vocal weight estimation and drift detection across a session
• 🎵 Pitch contour mapping with note-name and cent-offset readouts
• 🧘 Guided warm-up and cool-down routines
• 📈 Session history with trend lines and progress snapshots
• 🖥️ Responsive interface that adapts from wide desktops to narrow tablets
• 🌐 Multilingual support with community-contributed locale packs
• 🕛 Round-the-clock assistance through in-app guidance and community channels
• 🎨 Themeable visualizations (light, dark, high-contrast, and colorblind-safe palettes)
• 🔒 Local-first audio handling — your recordings stay where you decide

The heart of VocalForge is its rendering pipeline. Signal data is captured, windowed, and transformed, then projected into a set of coordinated views:

A 2D scatter field where each point is a moment in time, mapped by F1 and F2. As you speak or sing, a comet trail draws your resonance path. Hovering reveals the exact vowel neighborhood you're hovering through.

A gentle vertical gauge that shifts from light to heavy, giving immediate feedback on how much mass your voice is carrying. Sudden jumps are flagged with a soft pulse rather than an alarm.

A ribbon-like line that rises and falls with your pitch. It doesn't judge; it simply shows. Optional target overlays let you trace toward a reference contour without pressure.

A subtle meter that reflects breath support indirectly through stability and amplitude variance — a proxy, not a medical measurement.

Each view is independently toggleable, and layouts can be saved as presets.

VocalForge treats your voice the way a luthier treats a violin. The safety model includes:

• Load monitoring — cumulative vocal effort is estimated across a session and gently capped.
• Rest reminders — configurable intervals to pause, hydrate, and reset.
• Strain heuristics — indicators that suggest when weight, pitch, or volume trends signal fatigue.
• Session budgets — daily time thresholds you define yourself, respected by the app.

Safety here is a partnership, not a gatekeeper. The tool suggests; you decide.

VocalForge is constructed with a responsive layout philosophy so the visualization canvas feels natural on a 13-inch laptop, a 27-inch monitor, or a tablet propped against a music stand. Panels reflow, fonts scale, and touch targets grow when input shifts from pointer to finger.

• Adaptive grid layout
• DPI-aware rendering
• Keyboard-first navigation
• Reduced-motion mode for sensitive users

Voice work is universal; language is not. VocalForge ships with internationalization baked in from day one, not bolted on later. Locale packs are simple, text-based, and community-editable. Interface language, phoneme reference sets, and help content can all be localized independently.

Current and planned locale coverage grows with contributors. If your language isn't here yet, the locale template makes it a weekend project.

Training shouldn't require waiting for support hours. VocalForge includes:

• Contextual in-app tips that appear when they're relevant
• A built-in knowledge base covering common questions
• Community forums and chat channels with global coverage
• Documentation that's written to be read, not skimmed

Guidance is designed to feel like a patient teacher, not a pop-up ad.

The Rust core keeps the CPU cool and the battery happy, while the visualization layer ensures animations stay fluid even on modest hardware.

• Lock-free audio thread for glitch-free capture
• Zero-copy buffers between DSP and render stages
• Adaptive frame pacing to match display refresh
• Idle-state throttling to conserve power
• Session snapshots compressed for fast reload

VocalForge is designed to run comfortably alongside a DAW, a video call, or a stack of browser tabs.

• Full keyboard operation
• Screen-reader-friendly labels and landmarks
• Colorblind-safe visualization palettes
• Adjustable contrast and motion settings
• Text scaling without layout breakage
• Captions and transcripts for any embedded media

Inclusion is a design constraint, not a checkbox.

Your voice is biometric data. VocalForge treats it that way.

• Audio is processed locally by default
• No telemetry without explicit opt-in
• Session history lives in a local store you control
• Export and delete are first-class features
• No hidden uploads, ever

2026 Q1

• Stable formant tracking across all target platforms
• Locale pack v1 release with community contributions

2026 Q2

• Adaptive training plans based on session trends
• Expanded vowel reference library

2026 Q3

• Multi-voice profile support for shared devices
• Enhanced breath analytics

2026 Q4

• Plugin bridge for DAW integration
• Public API for community visualizers

Roadmap items shift with community feedback. Priorities are living, not carved in stone.

Contributions are welcomed with open arms — code, docs, translations, design, and bug reports alike. Before opening a pull request, please read the contribution guide and code of conduct. Small, focused changes are easier to review and merge than sprawling ones.

Areas where help is especially valued:

• Locale packs
• Accessibility audits
• DSP accuracy testing
• Visualization presets
• Documentation clarity

Be kind. Be patient. Be generous with context. Voice training can be vulnerable work, and this community is a safe space for learners at every stage. Harassment, gatekeeping, and elitism have no home here.

VocalForge is written to be found by the people who need it: voice training software users, formant visualization enthusiasts, pitch tracking practitioners, Rust audio tooling developers, and gender-affirming voice communities. Keywords are woven in naturally — voice trainer, formant visualization, pitch tracking, vocal weight analysis, Rust audio application, real-time voice feedback, multilingual voice software — never stuffed, always in service of clarity.

Is VocalForge a replacement for a voice coach? No. It's a companion, not a substitute. A skilled coach brings judgment and empathy that software cannot replicate.

Does it record my voice? Only if you ask it to. Default behavior is analysis without storage.

Can it run on older hardware? Yes, with reduced frame pacing and simplified visualizations.

Will there be a mobile version? A responsive web target is planned; native mobile is on the long-horizon roadmap.

How do I request a feature? Open an issue with context and use case. The best requests describe a problem, not just a solution.

VocalForge is an educational and self-practice tool. It is not a medical device and does not diagnose, treat, or prevent any condition. If you experience persistent vocal pain, hoarseness, or discomfort, consult a qualified healthcare professional such as a laryngologist or speech-language pathologist. Training results vary by individual, and no specific outcome is guaranteed. Use responsibly, hydrate often, and listen to your body.

This project is released under the MIT License. See the full license text at MIT License.

Copyright (c) 2026 VocalForge Contributors.