SeasonTTS 1.0.0

SeasonTTS

SeasonTTS is a .NET text-to-speech library for Qwen3-TTS ONNX bundles.

The initial public release focuses on a simple application-facing API:

• instance-based preset voice generation through new CustomVoice(model) and Generate(...)
• instance-based voice cloning through new CloneVoice(model) and Clone(...)
• explicit initialization through Initialize() so model loading stays outside the hot inference path
• standard WAV byte output suitable for saving directly as .wav
• local model directory loading without requiring a Python runtime

The current implementation targets the pre-exported Qwen3-TTS ONNX bundles published by elbruno on Hugging Face. SeasonTTS is an independent project and is not affiliated with Qwen, Alibaba, or ElBruno.QwenTTS.

Naming

• Package name: SeasonTTS
• Main runtime APIs: SeasonTTS.CustomVoice, SeasonTTS.CloneVoice
• Repository: SeasonRealms/SeasonTTS

Origin

• SeasonTTS is built from a simplified and performance-tuned codebase derived in part from ElBruno.QwenTTS.
• Source files copied or adapted from ElBruno.QwenTTS retain their original attribution headers where applicable.
• SeasonTTS restructures the public API around two direct entry points: CustomVoice and CloneVoice.
• The current release also separates initialization from inference so repeated synthesis avoids repeated model loading.

Features

• Qwen3-TTS CustomVoice inference from .NET
• Qwen3-TTS Base voice cloning inference from .NET
• Explicit Initialize() step for model/session setup
• Returns standard RIFF/WAVE bytes instead of binding to a specific playback stack
• Supports local ONNX model directories
• Supports preset speaker selection via QwenVoicePreset
• Supports instruct style control on compatible 1.7B CustomVoice models
• Supports reference-audio cloning, with optional referenceText for ICL mode when the Base bundle includes tokenizer12hz_encode.onnx

Current API

public class CustomVoice : IDisposable
{
    public CustomVoice(
        string model,
        Func<SessionOptions>? sessionOptionsFactory = null,
        Func<SessionOptions>? vocoderSessionOptionsFactory = null);

    public Task Initialize(CancellationToken cancellationToken = default);

    public Task<byte[]> Generate(
        string text,
        QwenVoicePreset voice = QwenVoicePreset.Ryan,
        string language = "auto",
        string? instruct = null,
        CancellationToken cancellationToken = default);
}

public class CloneVoice : IDisposable
{
    public CloneVoice(
        string model,
        Func<SessionOptions>? sessionOptionsFactory = null);

    public Task Initialize(CancellationToken cancellationToken = default);

    public Task<byte[]> Clone(
        string text,
        Stream referenceAudioStream,
        string? referenceText = null,
        string language = "auto",
        CancellationToken cancellationToken = default);
}

Install

NuGet packaging is planned, but the first public release is source-first.

For now, either:

• reference the SeasonTTS project directly
• copy the published source into your solution
• wait for the upcoming NuGet package

Supported Models

SeasonTTS currently targets these three ONNX model repositories:

Direct model pages:

• elbruno/Qwen3-TTS-12Hz-0.6B-CustomVoice-ONNX
• elbruno/Qwen3-TTS-12Hz-1.7B-CustomVoice-ONNX
• elbruno/Qwen3-TTS-12Hz-0.6B-Base-ONNX

These repositories are public. In normal direct-download usage, HF_TOKEN is usually not required.

Model Download

You have two practical options:

1. Download via GitHub Actions

The repository includes:

• .github/workflows/download-qwen-tts-onnx.yml

This workflow supports three choices:

• qwen3-tts-12hz-0.6b-customvoice
• qwen3-tts-12hz-1.7b-customvoice
• qwen3-tts-12hz-0.6b-base

It downloads the full selected Hugging Face repository snapshot and uploads it as a workflow artifact.

Notes:

• direct downloads from these public repositories usually work without HF_TOKEN
• if GitHub-hosted runners hit Hugging Face rate limits, setting HF_TOKEN as an optional repository secret can improve reliability

2. Download directly from Hugging Face

You can manually download the same ONNX bundles from:

• elbruno/Qwen3-TTS-12Hz-0.6B-CustomVoice-ONNX
• elbruno/Qwen3-TTS-12Hz-1.7B-CustomVoice-ONNX
• elbruno/Qwen3-TTS-12Hz-0.6B-Base-ONNX

Typical local layout:

Models/
  qwen3-tts-12hz-0.6b-customvoice-onnx/
  qwen3-tts-12hz-1.7b-customvoice-onnx/
  qwen3-tts-12hz-0.6b-base-onnx/

Quick Start

CustomVoice

This example matches the current sample app usage:

using SeasonTTS;

var model = @"../../../../../../Models/qwen3-tts-12hz-0.6b-customvoice-onnx";
var customVoice = new CustomVoice(model);

await customVoice.Initialize();

byte[] wavBytes = await customVoice.Generate(
    "Hello from SeasonTTS.",
    QwenVoicePreset.Vivian);

File.WriteAllBytes("customvoice.wav", wavBytes);

Example with 1.7B instruct control:

using SeasonTTS;

var model = @"../../../../../../Models/qwen3-tts-12hz-1.7b-customvoice-onnx";
var customVoice = new CustomVoice(model);

await customVoice.Initialize();

byte[] wavBytes = await customVoice.Generate(
    "Welcome to SeasonTTS.",
    QwenVoicePreset.Ryan,
    instruct: "Speak warmly and slowly with a clear presentation style.");

File.WriteAllBytes("customvoice-17b.wav", wavBytes);

CloneVoice

This example also follows the current samples/Sample/App.cs pattern:

using SeasonTTS;

var model = @"../../../../../../Models/qwen3-tts-12hz-0.6b-base-onnx";
var cloneVoice = new CloneVoice(model);

await cloneVoice.Initialize();

using var referenceStream = File.OpenRead("reference.wav");

byte[] wavBytes = await cloneVoice.Clone(
    "This sentence uses the cloned voice.",
    referenceStream);

File.WriteAllBytes("clonevoice.wav", wavBytes);

Optional ICL mode:

using SeasonTTS;

var model = @"../../../../../../Models/qwen3-tts-12hz-0.6b-base-onnx";
var cloneVoice = new CloneVoice(model);

await cloneVoice.Initialize();

using var referenceStream = File.OpenRead("reference.wav");

byte[] wavBytes = await cloneVoice.Clone(
    "This sentence uses the cloned voice.",
    referenceStream,
    referenceText: "The transcript of the reference audio.");

File.WriteAllBytes("clonevoice-icl.wav", wavBytes);

Why This API

This initial API is intentionally simple:

• the caller creates either CustomVoice or CloneVoice
• Initialize() performs the expensive model and embedding setup ahead of inference
• Generate(...) and Clone(...) return ready-to-save WAV bytes
• the host application remains free to choose its own recording, playback, storage, or transport flow

This shape matches the current sample integration style and keeps app-side code small.

Initialization Notes

The current public API expects the caller to initialize the instance explicitly before synthesis:

var customVoice = new CustomVoice(modelDir);
await customVoice.Initialize();

var cloneVoice = new CloneVoice(modelDir);
await cloneVoice.Initialize();

This avoids paying model-loading cost inside every inference call.

Result Format

Generate(...) and Clone(...) both return a byte[] containing a standard .wav file:

• sample format: 16-bit PCM WAV
• sample rate: 24000
• output is ready to save with File.WriteAllBytes(...)

Copyright And Attribution

SeasonTTS contains code derived in part from ElBruno.QwenTTS.

• SeasonTTS repository code is released under the MIT License
• portions derived from ElBruno.QwenTTS remain subject to the upstream MIT License and attribution requirements
• the upstream project and license text are listed in THIRD-PARTY-NOTICES.md
• copied or adapted files should retain their attribution headers

Project relationship:

• Upstream inspiration and source base: elbruno/ElBruno.QwenTTS
• This project simplifies the public API, restructures initialization, and continues development independently

License Split

• This repository code is released under the MIT License
• Pre-exported ONNX model bundles are not covered by this repository MIT license
• Model weights, tokenizer assets, and converted ONNX bundles remain governed by their own upstream terms and repository metadata

Recommended references:

• ElBruno.QwenTTS
• Qwen3-TTS official project
• Qwen official model collection

Disclaimer

SeasonTTS is an independent open source project. It is not affiliated with, endorsed by, or distributed by Qwen, Alibaba, or the ElBruno.QwenTTS project.