EnglishJapanese

Important Notices & User Guide

Important Notices

BR-TTS Neural Network Edition
Important Notices & User Guide

Version: 1.0.0  |  Copyright (C) 2025 BR SYSTEMS  |  GPL v3
Product Overview

Product Overview

BR-TTS NNW is a high-quality Japanese voice synthesis system using the latest neural network technology. It generates natural read-aloud audio from just a few seconds of voice samples.

Main Features

  • Easy voice synthesis via GUI operation
  • Automatic conversion of numbers to kanji reading (e.g., 15 o'clock → juugo ji)
  • Selection from multiple voice samples (with preview function)
  • Playback speed adjustment (0.5x to 2.0x)
  • Network folder support
  • GPU acceleration support (NVIDIA CUDA)
  • Processing log display function

Technical Specifications

ItemDetails
Voice Synthesis EngineCoqui TTS (XTTS v2)
Supported OSWindows 10/11 (64-bit)
Recommended GPUNVIDIA GPU (CUDA compatible)
LicenseGPL v3
System Requirements

Operating Environment

i
Minimum Requirements
  • OS: Windows 10/11 (64-bit)
  • Memory: 8GB or more
  • Storage: 5GB or more free space
  • Python 3.11 environment (source code edition)
✓
Recommended Environment
  • Memory: 16GB or more
  • GPU: NVIDIA GeForce GTX 1060 or higher (CUDA)
  • Storage: SSD recommended
  • Commercial edition: Python not required
Installation

Installation Instructions

Commercial Edition (exe)

1

Run Installer

Run the installer received after purchase.

2

Follow Instructions

Follow the on-screen instructions to complete installation.

3

Launch from Desktop

Launch using the desktop icon created during installation.

Source Code Edition (GitHub)

1

Download Source Code from GitHub

https://github.com/oyamat/br-tts-nnw

2

Build Anaconda Environment

conda create -n tts_nnw0 python=3.11 -y
conda activate tts_nnw0
3

Install Required Packages

pip install -r requirements.txt
4

Run the Application

python tts_gui.py

Please refer to README.md on GitHub for detailed installation instructions.

Usage

How to Use

1

Launch Application

Start from the desktop icon (Commercial) or run python tts_gui.py (Source).

2

Select Voice Sample

Click "Select" button → select a voice sample (.wav file). If multiple are available, preview to select the optimal one.

3

Enter Text

Enter the text to be voiced in the text field. Turning on Auto Number Conversion will automatically convert "15 o'clock" to "juugo ji."

4

Synthesize Voice

Click "Synthesize Voice" button. Wait until processing is complete (a few seconds to tens of seconds).

5

Play & Save

Confirm with "Play" button → save with "Save As."

i
Voice Sample Recommendations
  • Format: WAV file (16-bit, 22050 Hz recommended)
  • Length: Approx. 3–10 seconds
  • Content: Clear recording with little noise, rich in emotional expression
License

License

✓
Open Source (GPL v3)

This software is licensed under the GNU General Public License Version 3 (GPL v3). See LICENSE.txt for details.
Source code is published at: https://github.com/oyamat/br-tts-nnw

i
Commercial Edition

BR SYSTEMS offers a commercial edition with support (¥9,800). The commercial edition includes exe format delivery, installer, manual, and email support. The software license itself is identical.
Details: https://www.brsystems.jp/tts-nnw-1/br-tts-neural-network-edition-1/

Voice Rights

Regarding Use of Generated Voice

✓
Copyright of Generated Voice

The copyright of voice generated by this software belongs to the customer. It can be freely used including for commercial purposes.

!
Important Notices
  • Rights processing for voice used in voice samples is the responsibility of the customer
  • Generating voice that infringes on the rights of others is prohibited
  • Use for illegal purposes is prohibited
Troubleshooting

Troubleshooting

!
Application Won't Launch
  • Check that GPU (CUDA) drivers are correctly installed
  • Check that Python environment is correctly set up (source code edition)
!
Voice Synthesis is Slow
  • Check whether GPU is being recognized (confirm in log)
  • When operating on CPU only, processing time will be 5–10 times longer
!
Poor Voice Quality
  • Check voice sample quality (clear, low-noise recording)
  • Try a different voice sample
  • Adjust length of voice sample (3–10 seconds is optimal)
!
Numbers Not Being Read Correctly
  • Turn on the Auto Number Conversion function
  • Check conversion results in log
  • For specific numbers with issues, manually convert to hiragana
Release History

Version History

VersionRelease DateChanges
1.0.0 2025-03-12 Initial release — basic voice synthesis, auto number conversion, multiple voice sample selection, playback speed adjustment, log display
i
Copyright Notice

BR-TTS Neural Network Edition
Copyright (C) 2025 BR SYSTEMS. All rights reserved.
This software is provided under the GPL v3 license.
For copyright information of third-party open source software used, please refer to THIRD_PARTY_NOTICES.txt.

Support & Contact

Commercial Edition Support

📧 info@brsystems.jp
Hours: Weekdays 10:00–17:00
(excluding weekends and holidays)

Open Source Edition (Free)

GitHub Issues:
github.com/oyamat/br-tts-nnw
Community support only (no official support)

General Inquiries

BR SYSTEMS
📞 0467-40-4318
📧 info@brsystems.jp

Contact Form

Contact Form →
brsystems.jp