Piano melodies transformed into full pop instrumentals without aligned data
Piano2Pop: Expanding Piano Demos into Pop Productions Without Aligned Supervision
Machine LearningSound
Summary
Turning simple piano music into full pop songs usually needs matching examples of both to learn from. The authors developed a method that can do this transformation without having those exact matches. They create pairs from synthetic music and use clever training tricks to make the system better. Their approach works well on real piano recordings, making pop instrumentals that sound better than previous attempts that didn’t train for this task.
What this means in practice
- •For music producers: Generate full pop instrumentals automatically from simple piano demos to speed up music creation.$Commercial implications: Enables products that auto-expand piano ideas into pop tracks for producers seeking quick arrangements.
- •For digital audio workstation developers: Integrate piano-to-pop expansion tools into DAWs to provide users with creative accompaniment generation without aligned dataset needs.
Authors
Yun-Chen Cheng, Chih-Pin Tan, Tzu-Hung Huang, Yi-Hsuan Yang
Abstract
We introduce Piano2Pop, a music-expansion task that transforms piano MIDI into instrumental pop productions. Enabled by recent advances in music generation and polyphonic music transcription, our pseudo-supervised approach conditions a pretrained flow-matching model on piano rolls without aligned piano-pop supervision. We train on synthetic piano-instrumental pairs derived from pop recordings, comparing pair-free transcription, including function-selective (FST) and register-density (RDA) augmentations, with a pair-supervised piano-cover generator as pseudo-demo sources. On 476 testing real-world piano-cover/pop pairs, the pair-supervised approach achieves the best Frechet Audio Distance, Accompaniment Prompt Adherence, and spectral shape on real-piano inputs, while FST mitigates piano-heavy outputs in pair-free training and remains a promising alternative. All fine-tuned systems significantly outperform the zero-shot baselines, which suggests a role for task-specific modeling.