# HPPNet: Modeling the Harmonic Structure and Pitch Invariance in Piano Transcription

arXiv: https://arxiv.org/abs/2208.14339 · HTML: https://mlsanity.com/paper/2208.14339/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Aug 2022.

## Built on by (most cited first, 7 of 7)

- [Automatic Piano Transcription with Hierarchical Frequency-Time Transformer](https://mlsanity.com/paper/2307.04305/) (2023)
- [DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability](https://mlsanity.com/paper/2210.05148/) (2022)
- [Toward Fully Self-Supervised Multi-Pitch Estimation](https://mlsanity.com/paper/2402.15569/) (2024)
- [PianoVAM: A Multimodal Piano Performance Dataset](https://mlsanity.com/paper/2509.08800/) (2025)
- [D3RM: A Discrete Denoising Diffusion Refinement Model for Piano Transcription](https://mlsanity.com/paper/2501.05068/) (2025)
- [Towards Robust Transcription: Exploring Noise Injection Strategies for Training Data Augmentation](https://mlsanity.com/paper/2410.14122/) (2024)
- [Piano Transcription by Hierarchical Language Modeling with Pretrained Roll-based Encoders](https://mlsanity.com/paper/2501.03038/) (2025)

## Abstract (arXiv)

While neural network models are making significant progress in piano transcription, they are becoming more resource-consuming due to requiring larger model size and more computing power. In this paper, we attempt to apply more prior about piano to reduce model size and improve the transcription performance. The sound of a piano note contains various overtones, and the pitch of a key does not change over time. To make full use of such latent information, we propose HPPNet that using the Harmonic Dilated Convolution to capture the harmonic structures and the Frequency Grouped Recurrent Neural Network to model the pitch-invariance over time. Experimental results on the MAESTRO dataset show that our piano transcription system achieves state-of-the-art performance both in frame and note scores (frame F1 93.15%, note F1 97.18%). Moreover, the model size is much smaller than the previous state-of-the-art deep learning models.