NNSVS
What it is
NNSVS is a free, MIT-licensed open-source library for building neural-network singing voice synthesis systems, primarily for research. Users can create voicebanks from their own datasets and adapt reproducible Kaldi- and ESPnet-style recipes covering preprocessing, training, and synthesis. Packed models accept MusicXML or UST input and can produce WAV output. The documented pipeline includes time-lag, duration, acoustic, post-filter, and vocoder modules. Custom PyTorch models can be added through the NNSVS BaseModel interface. Optuna supports hyperparameter optimization, and MLflow tracks and visualizes search results. Installation supports Linux, Mac OS X, and Windows; Linux is recommended for development. Python 3.7 or later and a C/C++ compiler are required, with GPU/CUDA strongly recommended. Google Colab is available without a GPU, though not recommended for hyperparameter optimization. Acoustic-model training may take several hours or a full day. The post-filter is not currently recommended, and neural-vocoder support is limited.
For and against
For
- Open source under the MIT license
- Custom voicebanks from personal datasets
- Recipes cover preprocessing through synthesis
- Supports MusicXML and UST inputs
- Runs on Linux, macOS, and Windows
Against
- Designed primarily for research
- GPU/CUDA recommended for performance
- Training may take hours or a day
- Post-filter and neural-vocoder support limited
The verdict
NNSVS offers a configurable research workflow for training and synthesizing singing voices from user data. It is most appropriate for developers comfortable with Python, compilers, and model training.
Who it suits
NNSVS suits researchers and developers building custom singing voice systems and voicebanks. Non-researchers may prefer the more user-friendly tools the project points to.
On the record
- Compute
- GPU/CUDA is strongly recommended for performance, while Google Colab is listed as an option for users without GPUs.nnsvs.github.io · read 8 Oct 2026
- Customization
- Users can implement custom PyTorch models by inheriting NNSVS's BaseModel interface.nnsvs.github.io · read 8 Oct 2026
- Input formats
- Packed models can generate singing voice from MusicXML or UST files.nnsvs.github.io · read 7 Oct 2026
- Inputs and outputs
- With a packed model, users can generate singing voice by inputting MusicXML or UST files.nnsvs.github.io · read 8 Oct 2026
- Integrations
- NNSVS integrates Optuna for hyperparameter optimization and uses MLflow to visualize and track search results.nnsvs.github.io · read 7 Oct 2026
- Intended users
- The project says it was originally designed for research and recommends other tools for users who are neither researchers nor software developers.nnsvs.github.io · read 8 Oct 2026
- Languages
- NNSVS has been used to create singing voice synthesis systems for multiple languages, with the documentation saying 8 or more as far as its author knows.nnsvs.github.io · read 8 Oct 2026
- Licensing
- The project’s GitHub repository lists the MIT license.github.com · read 7 Oct 2026
- Notable limitation
- The documentation says the post-filter is currently not recommended and neural-vocoder support is limited.nnsvs.github.io · read 8 Oct 2026
- Open source
- NNSVS is fully open-source, and users can create their own voicebanks with their dataset.nnsvs.github.io · read 8 Oct 2026
- Optional compute
- Users without a GPU can run NNSVS on Google Colab, though the documentation does not recommend Colab for hyperparameter optimization.nnsvs.github.io · read 7 Oct 2026
- Purpose
- NNSVS is a neural network based singing voice synthesis library designed for research.nnsvs.github.io · read 8 Oct 2026
- Requirements
- The guide specifies Python 3.7 or later and says a C/C++ compiler must be installed in advance.nnsvs.github.io · read 8 Oct 2026
- Research recipes
- It includes reproducible Kaldi and ESPnet style recipes for creating baseline systems in research.nnsvs.github.io · read 7 Oct 2026
- Research workflows
- NNSVS includes reproducible recipes modeled on Kaldi and ESPnet approaches for building baseline systems in research.nnsvs.github.io · read 8 Oct 2026
- Supported operating systems
- The installation guide lists Linux, Mac OS X, and Windows as supported platforms and says Linux is recommended for development.nnsvs.github.io · read 7 Oct 2026
- Supported platforms
- The installation guide lists Linux, Mac OS X, and Windows as supported and says NNSVS is tested on them with GitHub Actions.nnsvs.github.io · read 8 Oct 2026
- Support information
- The opened project pages provide installation guides, tutorials, and API reference documentation; they do not state a dedicated support service.nnsvs.github.io · read 7 Oct 2026
