README.md

---
license: apache-2.0
base_model: openbmb/VoxCPM2
pipeline_tag: text-to-speech
language:
  - zh
datasets:
  - pymaster/CrawlSinger-OS
tags:
  - singing-voice-synthesis
  - score-conditioned
  - music
  - arxiv:2607.27768
---

# VocalRender

**VocalRender: Score-Native Singing Voice Synthesis for Real-World Composition**

- [Paper](https://arxiv.org/abs/2607.27768)
- [Code and documentation](https://github.com/pymaster17/VocalRender)
- [Open-source training dataset](https://huggingface.co/datasets/pymaster/CrawlSinger-OS)

VocalRender is a score-native singing voice synthesis (SVS) model designed for
real-world composition. It synthesizes 48 kHz singing directly from lyrics,
MIDI pitches, symbolic note values, and tempo through an original combination
of an interleaved lyric--note representation, continuous acoustic latents, and
autoregressive diffusion modeling. It does not require phoneme-level
durations, an explicit duration predictor, or a time-aligned acoustic
reference.

![VocalRender overview](https://raw.githubusercontent.com/pymaster17/VocalRender/main/assets/intro.png)

## How it works

VocalRender combines three components:

1. A score-native interleaved representation serializes BPM followed by each
   lyric syllable and its associated `(pitch, note-value)` pairs. This retains
   lyric-to-note alignment and supports melisma.
2. An Audio VAE represents singing as compact continuous acoustic latents.
3. An autoregressive diffusion model generates the latent sequence patch by
   patch and decides when to stop. The VAE decoder renders the result as a
   waveform.

![VocalRender architecture](https://raw.githubusercontent.com/pymaster17/VocalRender/main/assets/structure.png)

## Checkpoints

This repository contains two variants with the **same architecture, parameter
count, and speech-pretrained base-model initialization**. Their only
difference is the training recipe (training corpus and schedule):

| Variant | Repository path | Training data and strategy | Paper results |
| --- | --- | --- | --- |
| **VocalRender** | `VocalRender/` | CrawlSinger-OS (>2,300 h); 40k-step synthetic pretraining followed by 20k-step real-data finetuning | Stronger subjective score following (MS-MOS 2.96) |
| **VocalRender-Pro** | `VocalRender-Pro/` | CrawlSinger (>5,600 h of in-house real singing); 160k training steps | Better intelligibility, speaker similarity, naturalness, and OOD robustness; MS-MOS 2.71 |

VocalRender-Pro reduces WER from 4.44 to 3.88 and improves speaker similarity
from 0.922 to 0.929 on Opencpop. On CrawlSinger-Eval, WER changes from 4.52 to
4.45 and similarity from 0.919 to 0.926. The paper relates these improvements
to the larger amount of real singing and broader singer coverage. VocalRender's
higher score-following score may reflect more reliable score annotations in
its real-data finetuning subset.

Each directory contains the model weights, AudioVAE, model configuration, and
extended SVS tokenizer required by the inference code. Both checkpoints have
the same model size; each is approximately 9.5 GB to download.

## Installation

```bash
git clone --recurse-submodules https://github.com/pymaster17/VocalRender.git
cd VocalRender
uv sync
source .venv/bin/activate
```

## Download

Download either checkpoint into the repository's `pretrained_models/`
directory:

```bash
# VocalRender
hf download pymaster/VocalRender \
    --include "VocalRender/*" \
    --local-dir pretrained_models

# Or VocalRender-Pro
hf download pymaster/VocalRender \
    --include "VocalRender-Pro/*" \
    --local-dir pretrained_models
```

## Quick inference

The GitHub repository includes three ready-to-use prompt/score pairs selected
from OpenCpop. This command runs the first demo:

```bash
python scripts/infer_vocalrender_svs_single.py \
    --ckpt_dir pretrained_models/VocalRender \
    --json_file examples/opencpop_demo.json \
    --item_name 2003000087 \
    --prompt_audio examples/prompt_audio/2003000081.wav \
    --output demo_2003000087.wav
```

Additional bundled pairs are `2017000646` with prompt `2017000644.wav`, and
`2044001652` with prompt `2044001666.wav`. The prompt clips are 4.19-6.17
seconds long and do not overlap their paired target segment. Clone the
[GitHub repository](https://github.com/pymaster17/VocalRender) to obtain the
scripts and examples. The excerpts remain subject to the
[OpenCpop](https://wenet.org.cn/opencpop/) terms.

Use `pretrained_models/VocalRender-Pro` as `--ckpt_dir` to run the Pro model.
Full generation requires a CUDA-capable compute node. Prompt audio is required:
the released checkpoints were trained with prompt audio on every sample
(`prompt_audio_prob=1.0`), so prompt-free inference is unsupported and may
substantially degrade quality. Use a clean 2-8 second singing clip, which also
provides the target timbre. See the
[repository README](https://github.com/pymaster17/VocalRender#batch-inference) for
batch-inference options.

## Input format

```json
{
  "item_name": "demo",
  "word": ["我", "的", "孤", "独"],
  "pitch": [65, 64, 64, 65, 67, 65],
  "note": ["<NOTE_8>", "<NOTE_32>", "<NOTE_16>", "<NOTE_16>", "<NOTE_16>", "<NOTE_8>"],
  "pitch2word": [0, 1, 2, 2, 2, 3],
  "bpm": 64
}
```

`pitch2word[i]` identifies the lyric index associated with note `i`, allowing
one lyric syllable to span multiple notes. `word_dur` and `pitch_dur` are
optional fields used only for visualization and evaluation; they are not model
inputs.

## Limitations

- The released checkpoints primarily target Mandarin Chinese singing.
- Checkpoint loading and generation require substantial disk space, system
  memory, and GPU memory.
- Output quality depends on the musical validity and lyric-to-note alignment of
  the input score.

## Model foundation

The released implementation initializes its ARDM backbone from the
speech-pretrained [VoxCPM2](https://huggingface.co/openbmb/VoxCPM2) weights.
The score-native representation, SVS formulation, data pipeline, and singing
training recipe are introduced by VocalRender.

## Citation

```bibtex
@article{chen2026vocalrender,
  title   = {VocalRender: Score-Native Singing Voice Synthesis for Real-World Composition},
  author  = {Chen, Yukun and Wang, Tianrui and Mu, Zhaoxi and Yang, Xinyu and Chng, EngSiong},
  journal = {arXiv preprint arXiv:2607.27768},
  year    = {2026},
  url     = {https://arxiv.org/abs/2607.27768}
}
```

## License

Apache License 2.0. See the repository's `LICENSE` file for details.