Code-Inc/supertonic
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

supertonic 3

Browse Source
This commit is contained in:
haeon
2026-05-06 23:09:06 +02:00
parent 6fc89ea89e
commit 0a98c9f127
47 changed files with 530 additions and 411 deletions
Download Patch File Download Diff File Expand all files Collapse all files

294
README.md
View File

@@ -1,93 +1,51 @@
# Supertonic — Lightning Fast, On-Device TTS
# Supertonic — Lightning Fast, On-Device, Accurate TTS
[![v2 Demo](https://img.shields.io/badge/🤗%20v2-Demo-yellow)](https://huggingface.co/spaces/Supertone/supertonic-2)
[![v2 Models](https://img.shields.io/badge/🤗%20v2-Models-blue)](https://huggingface.co/Supertone/supertonic-2)
[![v3 Demo](https://img.shields.io/badge/🤗%20v3-Demo-yellow)](https://huggingface.co/spaces/Supertone/supertonic-3)
[![v3 Models](https://img.shields.io/badge/🤗%20v3-Models-blue)](https://huggingface.co/Supertone/supertonic-3)
[![v2 Branch](https://img.shields.io/badge/v2-release%2Fsupertonic--2-lightgrey)](https://github.com/supertone-inc/supertonic/tree/release/supertonic-2)
[![v1 Demo](https://img.shields.io/badge/🤗%20v1%20(old)-Demo-lightgrey)](https://huggingface.co/spaces/Supertone/supertonic#interactive-demo)
[![v1 Models](https://img.shields.io/badge/🤗%20v1%20(old)-Models-lightgrey)](https://huggingface.co/Supertone/supertonic)
<p align="center">
<img src="img/supertonic_preview_0.1.jpg" alt="Supertonic Banner">
<img src="img/Supertonic3_HeroImage.png" alt="Supertonic 3 Banner">
</p>
**Supertonic** is a lightning-fast, on-device text-to-speech system designed for **extreme performance** with minimal computational overhead. Powered by ONNX Runtime, it runs entirely on your device—no cloud, no API calls, no privacy concerns.
**Supertonic** is a lightning-fast, on-device text-to-speech system designed for local inference with minimal overhead. Powered by ONNX Runtime, it runs entirely on your device—no cloud, no API calls, no privacy concerns.
### 📰 Update News
- **2026.04.29** - 🎉 **Supertonic 3** released with **31-language support**, improved reading accuracy, fewer repeat/skip failures, and v2-compatible public ONNX assets. [Demo](https://huggingface.co/spaces/Supertone/supertonic-3) | [Models](https://huggingface.co/Supertone/supertonic-3)
- **2026.01.22** - **[Voice Builder](https://supertonic.supertone.ai/voice_builder)** is now live! Turn your voice into a deployable, edge-native TTS with permanent ownership.
<p align="center">
<img src="img/voicebuilder_img.png" alt="Voice Builder" width="600">
</p>
- **2026.01.06** - 🎉 **Supertonic 2** released with multilingual support! Now supports English (`en`), Korean (`ko`), Spanish (`es`), Portuguese (`pt`), and French (`fr`). [Demo](https://huggingface.co/spaces/Supertone/supertonic-2) | [Models](https://huggingface.co/Supertone/supertonic-2)
- **2026.01.06** - 🎉 **Supertonic 2** released with 5-language support. The v2 code path is preserved on the [`release/supertonic-2`](https://github.com/supertone-inc/supertonic/tree/release/supertonic-2) branch.
- **2025.12.10** - Added `supertonic` PyPI package! Install via `pip install supertonic`. For details, visit [supertonic-py documentation](https://supertone-inc.github.io/supertonic-py)
- **2025.12.10** - Added [6 new voice styles](https://huggingface.co/Supertone/supertonic/tree/b10dbaf18b316159be75b34d24f740008fddd381) (M3, M4, M5, F3, F4, F5). See [Voices](https://supertone-inc.github.io/supertonic-py/voices/) for details
- **2025.12.08** - Optimized ONNX models via [OnnxSlim](https://github.com/inisis/OnnxSlim) now available on [Hugging Face Models](https://huggingface.co/Supertone/supertonic)
- **2025.11.24** - Added Flutter SDK support with macOS compatibility
### Table of Contents
## Quick Start
- [Demo](#demo)
- [Why Supertonic?](#why-supertonic)
- [Language Support](#language-support)
- [Getting Started](#getting-started)
- [Performance](#performance)
- [Built with Supertonic](#built-with-supertonic)
- [Citation](#citation)
- [License](#license)
Install the Python SDK and generate speech immediately. On the first run, Supertonic downloads the model assets from Hugging Face automatically.
## Demo
```bash
pip install supertonic
```
### Raspberry Pi
### Python
Watch Supertonic running on a **Raspberry Pi**, demonstrating on-device, real-time text-to-speech synthesis:
```python
from supertonic import TTS
https://github.com/user-attachments/assets/ea66f6d6-7bc5-4308-8a88-1ce3e07400d2
# First run downloads the model from Hugging Face automatically.
tts = TTS(auto_download=True)
### E-Reader
style = tts.get_voice_style(voice_name="M1")
Experience Supertonic on an **Onyx Boox Go 6** e-reader in airplane mode, achieving an average RTF of 0.3× with zero network dependency:
text = "A gentle breeze moved through the open window while everyone listened to the story."
wav, duration = tts.synthesize(text, voice_style=style, lang="en")
https://github.com/user-attachments/assets/64980e58-ad91-423a-9623-78c2ffc13680
### Chrome Extension
Turns any webpage into audio in under one second, delivering lightning-fast, on-device text-to-speech with zero network dependency—free, private, and effortless:
https://github.com/user-attachments/assets/cc8a45fc-5c3e-4b2c-8439-a14c3d00d91c
---
> 🎧 **Try it now**: Experience Supertonic in your browser with our [**Interactive Demo**](https://huggingface.co/spaces/Supertone/supertonic-2), or get started with pre-trained models from [**Hugging Face Hub**](https://huggingface.co/Supertone/supertonic-2)
## Why Supertonic?
- **⚡ Blazingly Fast**: Generates speech up to **167× faster than real-time** on consumer hardware (M4 Pro)—unmatched by any other TTS system
- **🪶 Ultra Lightweight**: Only **66M parameters**, optimized for efficient on-device performance with minimal footprint
- **📱 On-Device Capable**: **Complete privacy** and **zero latency**—all processing happens locally on your device
- **🎨 Natural Text Handling**: Seamlessly processes numbers, dates, currency, abbreviations, and complex expressions without pre-processing
- **⚙️ Highly Configurable**: Adjust inference steps, batch processing, and other parameters to match your specific needs
- **🧩 Flexible Deployment**: Deploy seamlessly across servers, browsers, and edge devices with multiple runtime backends.
## Language Support
We provide ready-to-use TTS inference examples across multiple ecosystems:
| Language/Platform | Path | Description |
|-------------------|------|-------------|
| [**Python**](py/) | `py/` | ONNX Runtime inference |
| [**Node.js**](nodejs/) | `nodejs/` | Server-side JavaScript |
| [**Browser**](web/) | `web/` | WebGPU/WASM inference |
| [**Java**](java/) | `java/` | Cross-platform JVM |
| [**C++**](cpp/) | `cpp/` | High-performance C++ |
| [**C#**](csharp/) | `csharp/` | .NET ecosystem |
| [**Go**](go/) | `go/` | Go implementation |
| [**Swift**](swift/) | `swift/` | macOS applications |
| [**iOS**](ios/) | `ios/` | Native iOS apps |
| [**Rust**](rust/) | `rust/` | Memory-safe systems |
| [**Flutter**](flutter/) | `flutter/` | Cross-platform apps |
> For detailed usage instructions, please refer to the README.md in each language directory.
tts.save_audio(wav, "output.wav")
print(f"Generated {duration:.2f}s of audio")
```
## Getting Started
@@ -107,18 +65,30 @@ Before running the examples, download the ONNX models and preset voices, and pla
> - Generic: see `https://git-lfs.com` for installers
```bash
git clone https://huggingface.co/Supertone/supertonic-2 assets
git lfs install
git clone https://huggingface.co/Supertone/supertonic-3 assets
```
### Quick Start
Some language examples need native runtimes:
- **Go**: install the ONNX Runtime C library. On macOS, `brew install onnxruntime` is enough; the Go example auto-detects Homebrew paths.
- **Java**: use a JDK, not just a JRE. On macOS, `brew install openjdk@17` works.
- **C#**: targets .NET 9 and allows major-version roll-forward, so .NET 9 or newer runtimes can run it.
Then run the Python example:
**Python Example** ([Details](py/))
```bash
cd py
uv sync
uv run example_onnx.py
```
This generates `outputs/output.wav` using the default preset voice.
### Other Runtime Examples
<details>
<summary><b>Run Supertonic in other languages and platforms</b></summary>
**Node.js Example** ([Details](nodejs/))
```bash
cd nodejs
@@ -182,95 +152,130 @@ cd ios/ExampleiOSApp
xcodegen generate
open ExampleiOSApp.xcodeproj
```
- In Xcode: Targets → ExampleiOSApp → Signing: select your Team
- Choose your iPhone as run destination → Build & Run
In Xcode: Targets → ExampleiOSApp → Signing: select your Team, then choose your iPhone as run destination and build.
</details>
### Technical Details
- **Runtime**: ONNX Runtime for cross-platform inference (CPU-optimized; GPU mode is not tested)
- **Runtime**: ONNX Runtime for cross-platform inference
- **Browser Support**: onnxruntime-web for client-side inference
- **Batch Processing**: Supports batch inference for improved throughput
- **Audio Output**: Outputs 16-bit WAV files
## Performance
## Performance Highlights
We evaluated Supertonic's performance (with 2 inference steps) using two key metrics across input texts of varying lengths: Short (59 chars), Mid (152 chars), and Long (266 chars).
Supertonic 3 is designed for practical on-device inference: compact enough to run locally, while staying competitive with much larger open TTS systems.
**Metrics:**
- **Characters per Second**: Measures throughput by dividing the number of input characters by the time required to generate audio. Higher is better.
- **Real-time Factor (RTF)**: Measures the time taken to synthesize audio relative to its duration. Lower is better (e.g., RTF of 0.1 means it takes 0.1 seconds to generate one second of audio).
### Reading Accuracy
### Characters per Second
| System | Short (59 chars) | Mid (152 chars) | Long (266 chars) |
|--------|-----------------|----------------|-----------------|
| **Supertonic** (M4 pro - CPU) | 912 | 1048 | 1263 |
| **Supertonic** (M4 pro - WebGPU) | 996 | 1801 | 2509 |
| **Supertonic** (RTX4090) | 2615 | 6548 | 12164 |
| `API` [ElevenLabs Flash v2.5](https://elevenlabs.io/docs/api-reference/text-to-speech/convert) | 144 | 209 | 287 |
| `API` [OpenAI TTS-1](https://platform.openai.com/docs/guides/text-to-speech) | 37 | 55 | 82 |
| `API` [Gemini 2.5 Flash TTS](https://ai.google.dev/gemini-api/docs/speech-generation) | 12 | 18 | 24 |
| `API` [Supertone Sona speech 1](https://docs.supertoneapi.com/en/api-reference/endpoints/text-to-speech) | 38 | 64 | 92 |
| `Open` [Kokoro](https://github.com/hexgrad/kokoro/) | 104 | 107 | 117 |
| `Open` [NeuTTS Air](https://github.com/neuphonic/neutts-air) | 37 | 42 | 47 |
<p align="center">
<img src="img/metrics/s3_vs_measured_wer_range_voxcpm2.png" alt="Supertonic 3 reading accuracy compared with measured model ranges and VoxCPM2">
</p>
> **Notes:**
> `API` = Cloud-based API services (measured from Seoul)
> `Open` = Open-source models
> Supertonic (M4 pro - CPU) and (M4 pro - WebGPU): Tested with ONNX
> Supertonic (RTX4090): Tested with PyTorch model
> Kokoro: Tested on M4 Pro CPU with ONNX
> NeuTTS Air: Tested on M4 Pro CPU with Q8-GGUF
Across measured languages, Supertonic 3 stays within a competitive WER/CER range against much larger open TTS models such as VoxCPM2, while preserving a lightweight on-device deployment path. Asterisked languages use CER; the others use WER.
### Real-time Factor
### Supertonic 2 to Supertonic 3
| System | Short (59 chars) | Mid (152 chars) | Long (266 chars) |
|--------|-----------------|----------------|-----------------|
| **Supertonic** (M4 pro - CPU) | 0.015 | 0.013 | 0.012 |
| **Supertonic** (M4 pro - WebGPU) | 0.014 | 0.007 | 0.006 |
| **Supertonic** (RTX4090) | 0.005 | 0.002 | 0.001 |
| `API` [ElevenLabs Flash v2.5](https://elevenlabs.io/docs/api-reference/text-to-speech/convert) | 0.133 | 0.077 | 0.057 |
| `API` [OpenAI TTS-1](https://platform.openai.com/docs/guides/text-to-speech) | 0.471 | 0.302 | 0.201 |
| `API` [Gemini 2.5 Flash TTS](https://ai.google.dev/gemini-api/docs/speech-generation) | 1.060 | 0.673 | 0.541 |
| `API` [Supertone Sona speech 1](https://docs.supertoneapi.com/en/api-reference/endpoints/text-to-speech) | 0.372 | 0.206 | 0.163 |
| `Open` [Kokoro](https://github.com/hexgrad/kokoro/) | 0.144 | 0.124 | 0.126 |
| `Open` [NeuTTS Air](https://github.com/neuphonic/neutts-air) | 0.390 | 0.338 | 0.343 |
<p align="center">
<img src="img/metrics/supertonic2_vs_3_comparison.png" alt="Supertonic 2 and Supertonic 3 comparison">
</p>
<details>
<summary><b>Additional Performance Data (5-step inference)</b></summary>
Compared with Supertonic 2, Supertonic 3 reduces repeat and skip failures, improves speaker similarity across the shared-language set, and expands language coverage from 5 to 31 languages. It keeps the v2-compatible public ONNX interface, so existing integrations can move to v3 with the same inference contract.
<br>
### Runtime Footprint
**Characters per Second (5-step)**
<p align="center">
<img src="img/metrics/runtime_cpu_gpu_latency_memory.png" alt="Supertonic CPU runtime compared with GPU baselines">
</p>
| System | Short (59 chars) | Mid (152 chars) | Long (266 chars) |
|--------|-----------------|----------------|-----------------|
| **Supertonic** (M4 pro - CPU) | 596 | 691 | 850 |
| **Supertonic** (M4 pro - WebGPU) | 570 | 1118 | 1546 |
| **Supertonic** (RTX4090) | 1286 | 3757 | 6242 |
Supertonic 3 runs fast on CPU, even compared with larger baselines measured on A100 GPU, and uses substantially less memory. The open-weight fixed-voice setting does not require a GPU, which makes local, browser, and edge deployment much easier.
**Real-time Factor (5-step)**
### Model Size
| System | Short (59 chars) | Mid (152 chars) | Long (266 chars) |
|--------|-----------------|----------------|-----------------|
| **Supertonic** (M4 pro - CPU) | 0.023 | 0.019 | 0.018 |
| **Supertonic** (M4 pro - WebGPU) | 0.024 | 0.012 | 0.010 |
| **Supertonic** (RTX4090) | 0.011 | 0.004 | 0.002 |
<p align="center">
<img src="img/metrics/model_size_comparison.png" alt="Model size comparison">
</p>
</details>
At about 99M parameters across the public ONNX assets, Supertonic 3 is much smaller than 0.7B to 2B class open TTS systems. The smaller model size is a practical advantage for download size, startup time, and on-device inference.
### Natural Text Handling
## Demo
Supertonic is designed to handle complex, real-world text inputs that contain numbers, currency symbols, abbreviations, dates, and proper nouns.
> **Try it now**: Experience Supertonic in your browser with our [**Interactive Demo**](https://huggingface.co/spaces/Supertone/supertonic-3), or get started with pre-trained models from [**Hugging Face Hub**](https://huggingface.co/Supertone/supertonic-3)
> 🎧 **View audio samples more easily**: Check out our [**Interactive Demo**](https://huggingface.co/spaces/Supertone/supertonic#text-handling) for a better viewing experience of all audio examples
### Raspberry Pi
Watch Supertonic running on a **Raspberry Pi**, demonstrating on-device, real-time text-to-speech synthesis:
https://github.com/user-attachments/assets/ea66f6d6-7bc5-4308-8a88-1ce3e07400d2
### E-Reader
Experience Supertonic on an **Onyx Boox Go 6** e-reader in airplane mode, achieving an average RTF of 0.3× with zero network dependency:
https://github.com/user-attachments/assets/64980e58-ad91-423a-9623-78c2ffc13680
### Chrome Extension
Turns any webpage into audio in under one second, delivering lightning-fast, on-device text-to-speech with zero network dependency—free, private, and effortless:
https://github.com/user-attachments/assets/cc8a45fc-5c3e-4b2c-8439-a14c3d00d91c
## Why Supertonic?
- **Blazingly Fast**: Optimized for low-latency, on-device speech generation across desktop, browser, and edge deployments
- **Lightweight**: Compact ONNX assets designed for efficient local execution
- **On-Device Capable**: Complete privacy and zero network dependency
- **Accurate Reading**: Improved reading stability with fewer repeat and skip failures
- **Expressive Tags**: Supports simple expression tags such as `<laugh>`, `<breath>`, and `<sigh>`
- **Flexible Deployment**: Ready-to-use examples across Python, JavaScript, browser, mobile, and native runtimes
## Language Support
Supertonic 3 supports 31 languages:
| Code | Language | Code | Language | Code | Language | Code | Language |
|------|----------|------|----------|------|----------|------|----------|
| `en` | English | `ko` | Korean | `ja` | Japanese | `ar` | Arabic |
| `bg` | Bulgarian | `cs` | Czech | `da` | Danish | `de` | German |
| `el` | Greek | `es` | Spanish | `et` | Estonian | `fi` | Finnish |
| `fr` | French | `hi` | Hindi | `hr` | Croatian | `hu` | Hungarian |
| `id` | Indonesian | `it` | Italian | `lt` | Lithuanian | `lv` | Latvian |
| `nl` | Dutch | `pl` | Polish | `pt` | Portuguese | `ro` | Romanian |
| `ru` | Russian | `sk` | Slovak | `sl` | Slovenian | `sv` | Swedish |
| `tr` | Turkish | `uk` | Ukrainian | `vi` | Vietnamese | | |
We provide ready-to-use TTS inference examples across multiple ecosystems:
| Language/Platform | Path | Description |
|-------------------|------|-------------|
| [**Python**](py/) | `py/` | ONNX Runtime inference |
| [**Node.js**](nodejs/) | `nodejs/` | Server-side JavaScript |
| [**Browser**](web/) | `web/` | WebGPU/WASM inference |
| [**Java**](java/) | `java/` | Cross-platform JVM |
| [**C++**](cpp/) | `cpp/` | High-performance C++ |
| [**C#**](csharp/) | `csharp/` | .NET ecosystem |
| [**Go**](go/) | `go/` | Go implementation |
| [**Swift**](swift/) | `swift/` | macOS applications |
| [**iOS**](ios/) | `ios/` | Native iOS apps |
| [**Rust**](rust/) | `rust/` | Memory-safe systems |
| [**Flutter**](flutter/) | `flutter/` | Cross-platform apps |
> For detailed usage instructions, please refer to the README.md in each language directory.
## Natural Text Handling
Supertonic is designed to handle complex, real-world text inputs that contain natural prose, punctuation, abbreviations, and proper nouns.
> 🎧 **View audio samples more easily**: Check out our [**Interactive Demo**](https://huggingface.co/spaces/Supertone/supertonic-3) for a better viewing experience of all audio examples
**Overview of Test Cases:**
| Category | Key Challenges | Supertonic | ElevenLabs | OpenAI | Gemini | Microsoft |
|:--------:|:--------------:|:----------:|:----------:|:------:|:------:|:---------:|
| Financial Expression | Decimal currency, abbreviated magnitudes (M, K), currency symbols, currency codes | ✅ | ❌ | ❌ | ❌ | ❌ |
| Time and Date | Time notation, abbreviated weekdays/months, date formats | ✅ | ❌ | ❌ | ❌ | ❌ |
| Phone Number | Area codes, hyphens, extensions (ext.) | ✅ | ❌ | ❌ | ❌ | ❌ |
| Technical Unit | Decimal numbers with units, abbreviated technical notations | ✅ | ❌ | ❌ | ❌ | ❌ |
@@ -300,33 +305,7 @@ Supertonic is designed to handle complex, real-world text inputs that contain nu
</details>
<details>
<summary><b>Example 2: Time and Date</b></summary>
<br>
**Text:**
> "The train delay was announced at **4:45 PM** on **Wed, Apr 3, 2024** due to track maintenance."
**Challenges:**
- Time expression with PM notation (4:45 PM)
- Abbreviated weekday (Wed)
- Abbreviated month (Apr)
- Full date format (Apr 3, 2024)
**Audio Samples:**
| System | Result | Audio Sample |
|--------|--------|--------------|
| **Supertonic** | ✅ | [🎧 Play Audio](https://drive.google.com/file/d/1ehkZU8eiizBenG2DgR5tzBGQBvHS0Uaj/view?usp=sharing) |
| ElevenLabs Flash v2.5 | ❌ | [🎧 Play Audio](https://drive.google.com/file/d/1ta3r6jFyebmA-sT44l8EaEQcMLVmuOEr/view?usp=sharing) |
| OpenAI TTS-1 | ❌ | [🎧 Play Audio](https://drive.google.com/file/d/1sskmem9AzHAQ3Hv8DRSZoqX_pye-CXuU/view?usp=sharing) |
| Gemini 2.5 Flash TTS | ❌ | [🎧 Play Audio](https://drive.google.com/file/d/1zx9X8oMsLMXW0Zx_SURoqjju-By2yh_n/view?usp=sharing) |
| VibeVoice Realtime 0.5B | ❌ | [🎧 Play Audio](https://drive.google.com/file/d/1ZpGEstZr4hA0EdAWBMCUFFWuAkIpYsVh/view?usp=sharing) |
</details>
<details>
<summary><b>Example 3: Phone Number</b></summary>
<summary><b>Example 2: Phone Number</b></summary>
<br>
@@ -352,7 +331,7 @@ Supertonic is designed to handle complex, real-world text inputs that contain nu
</details>
<details>
<summary><b>Example 4: Technical Unit</b></summary>
<summary><b>Example 3: Technical Unit</b></summary>
<br>
@@ -444,9 +423,8 @@ This paper describes the self-purification technique for training flow matching
This project's sample code is released under the MIT License. - see the [LICENSE](https://github.com/supertone-inc/supertonic?tab=MIT-1-ov-file) for details.
The accompanying model is released under the OpenRAIL-M License. - see the [LICENSE](https://huggingface.co/Supertone/supertonic-2/blob/main/LICENSE) file for details.
The accompanying model is released under the OpenRAIL-M License. - see the [LICENSE](https://huggingface.co/Supertone/supertonic-3/blob/main/LICENSE) file for details.
This model was trained using PyTorch, which is licensed under the BSD 3-Clause License but is not redistributed with this project. - see the [LICENSE](https://docs.pytorch.org/FBGEMM/general/License.html) for details.
Copyright (c) 2026 Supertone Inc.