Toward Part-Aware Choral Transcription with singing voice assignment

Hugging Face Daily Papers Papers
automatic-music-transcription choral-amt singing-voice audio-processing music-information-retrieval deep-learning

Summary

The paper introduces PawCT, the first end-to-end neural framework for part-aware choral transcription that identifies active SATB vocal parts and transcribes each into a separate note-level track, achieving substantial F1 improvements over prior choral AMT baselines on the YouChorale dataset.

Single-instrument automatic music transcription (AMT) has advanced substantially, yet choral applications require soprano, alto, tenor, and bass (SATB) to be transcribed as separate parts. Recent note-level choral AMT instead produces a single merged note track, limiting rehearsal, education, and score reconstruction. To address this limitation, we introduce Part-aware Choral Transcription (PawCT), to our knowledge the first end-to-end neural framework that identifies active SATB parts from choral audio and transcribes each into a separate note-level track. PawCT combines part-specific onset, offset, and frame prediction with part-presence estimation, union-level supervision, and structured training targets using a range prior (RP) based on SATB pitch ranges and its ordered-continuity (OC) extension, which adds within-part melodic continuity and cross-part pitch ordering. On YouChorale, PawCT-RP-OC achieves a macro part-aware note F1 of 0.225 at a 50-ms onset tolerance, outperforming an adapted choral baseline (0.165) by 36.4% relative and a two-stage post-hoc assignment pipeline (0.175). Its part-agnostic variant, PagCT, achieves a 50-ms onset F1 of 0.382, compared with 0.237 for the previous state-of-the-art choral AMT model. Cross-dataset evaluations on CSD and Cantoria further assess performance under dataset shift. These results demonstrate the benefit of jointly modeling note transcription and vocal-part assignment. Code and demos are available at https://hanyu-meng.github.io/Paw_Choral_AMT_Demo/.
Original Article
View Cached Full Text

Cached at: 10/22/26, 03:46 PM

Paper page - Toward Part-Aware Choral Transcription with singing voice assignment

Source: https://huggingface.co/papers/2610.09861

Abstract

PawCT is an end-to-end neural framework that jointly transcribes choral audio into separate SATB note tracks using part-specific predictions, range priors, and ordered-continuity constraints.

Single-instrumentautomatic music transcription(AMT) has advanced substantially, yet choral applications require soprano, alto, tenor, and bass (SATB) to be transcribed as separate parts. Recent note-levelchoral AMTinstead produces a single merged note track, limiting rehearsal, education, and score reconstruction. To address this limitation, we introduce Part-aware Choral Transcription (PawCT), to our knowledge the first end-to-end neural framework that identifies activeSATBparts from choral audio and transcribes each into a separate note-level track. PawCT combines part-specific onset, offset, andframe predictionwithpart-presence estimation,union-level supervision, and structured training targets using arange prior(RP) based onSATBpitch ranges and itsordered-continuity(OC) extension, which adds within-part melodic continuity and cross-part pitch ordering. On YouChorale, PawCT-RP-OC achieves a macro part-aware note F1 of 0.225 at a 50-ms onset tolerance, outperforming an adapted choral baseline (0.165) by 36.4% relative and a two-stage post-hoc assignment pipeline (0.175). Its part-agnostic variant, PagCT, achieves a 50-ms onset F1 of 0.382, compared with 0.237 for the previous state-of-the-artchoral AMTmodel.Cross-dataset evaluations on CSD and Cantoria further assess performance under dataset shift. These results demonstrate the benefit of jointly modeling note transcription and vocal-part assignment. Code and demos are available at https://hanyu-meng.github.io/Paw_Choral_AMT_Demo/.

View arXiv pageView PDFProject pageGitHub0Add to collection

Get this paper in your agent:

hf papers read 2610\.09861

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.09861 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.09861 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.09861 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Multi-Task Multi-Frame Visual Piano Transcription

Hugging Face Daily Papers

This paper introduces V2N, the first complete visual piano transcription system that jointly predicts onset, offset, keyhold, and velocity from video, achieving state-of-the-art results on PianoVAM and R3 benchmarks.

Under the Hood: Building a Real-Time Chord Recognizer

Lobsters Hottest

This article explains the technical architecture of a real-time chord recognizer, detailing a four-stage pipeline using pitch-class bitmasks, candidate generation, score normalization, and musical heuristics.

SheetSage2: Coherent Lead-Sheet Transcription with Synthetic Supervision

Hugging Face Daily Papers

SheetSage2 is a unified music transcription framework combining synthetic MIDI supervision, task-specific structured decoding, and autoregressive distillation, outperforming prior systems on 12 of 15 benchmark-metric pairs with publicly available model weights and code.