Saved in:
Bibliographic Details
Main Authors: Melucci, Pierfrancesco, Merialdo, Paolo, Akama, Taketo
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.09520
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908765244096512
author Melucci, Pierfrancesco
Merialdo, Paolo
Akama, Taketo
author_facet Melucci, Pierfrancesco
Merialdo, Paolo
Akama, Taketo
contents Deep learning models define the state-of-the-art in Automatic Drum Transcription (ADT), yet their performance is contingent upon large-scale, paired audio-MIDI datasets, which are scarce. Existing workarounds that use synthetic data often introduce a significant domain gap, as they typically rely on low-fidelity SoundFont libraries that lack acoustic diversity. While high-quality one-shot samples offer a better alternative, they are not available in a standardized, large-scale format suitable for training. This paper introduces a new paradigm for ADT that circumvents the need for paired audio-MIDI training data. Our primary contribution is a semi-supervised method to automatically curate a large and diverse corpus of one-shot drum samples from unlabeled audio sources. We then use this corpus to synthesize a high-quality dataset from MIDI files alone, which we use to train a sequence-to-sequence transcription model. We evaluate our model on the ENST and MDB test sets, where it achieves new state-of-the-art results, significantly outperforming both fully supervised methods and previous synthetic-data approaches. The code for reproducing our experiments is publicly available at https://github.com/pier-maker92/ADT_STR
format Preprint
id arxiv_https___arxiv_org_abs_2601_09520
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Towards Realistic Synthetic Data for Automatic Drum Transcription
Melucci, Pierfrancesco
Merialdo, Paolo
Akama, Taketo
Sound
Artificial Intelligence
Deep learning models define the state-of-the-art in Automatic Drum Transcription (ADT), yet their performance is contingent upon large-scale, paired audio-MIDI datasets, which are scarce. Existing workarounds that use synthetic data often introduce a significant domain gap, as they typically rely on low-fidelity SoundFont libraries that lack acoustic diversity. While high-quality one-shot samples offer a better alternative, they are not available in a standardized, large-scale format suitable for training. This paper introduces a new paradigm for ADT that circumvents the need for paired audio-MIDI training data. Our primary contribution is a semi-supervised method to automatically curate a large and diverse corpus of one-shot drum samples from unlabeled audio sources. We then use this corpus to synthesize a high-quality dataset from MIDI files alone, which we use to train a sequence-to-sequence transcription model. We evaluate our model on the ENST and MDB test sets, where it achieves new state-of-the-art results, significantly outperforming both fully supervised methods and previous synthetic-data approaches. The code for reproducing our experiments is publicly available at https://github.com/pier-maker92/ADT_STR
title Towards Realistic Synthetic Data for Automatic Drum Transcription
topic Sound
Artificial Intelligence
url https://arxiv.org/abs/2601.09520