Improving Perceptual Quality of Drum Transcription with the Expanded Groove MIDI Dataset (2020-04-01T00:00:00.000000Z)

TL;DR

This work optimize classifiers for downstream generation by predicting expressive dynamics (velocity) and shows with listening tests that they produce outputs with improved perceptual quality, despite achieving similar results on classification metrics.

Abstract

Classifier metrics, such as accuracy and F-measure score, often serve as proxies for performance in downstream tasks. For the case of generative systems that use predicted labels as inputs, accuracy is a good proxy only if it aligns with the perceptual quality of generated outputs. Here, we demonstrate this effect using the example of automatic drum transcription (ADT). We optimize classifiers for downstream generation by predicting expressive dynamics (velocity) and show with listening tests that they produce outputs with improved perceptual quality, despite achieving similar results on classification metrics. To train expressive ADT models, we introduce the Expanded Groove MIDI dataset (E-GMD), a large dataset of human drum performances, with audio recordings annotated in MIDI. E-GMD contains 444 hours of audio from 43 drum kits and is an order of magnitude larger than similar datasets. It is also the first human-performed drum dataset with annotations of velocity. We make this new dataset available under a Creative Commons license along with open source code for training and a pre-trained model for inference.

Authors

Jesse Engel

10 papers

Curtis Hawthorne

6 papers

Lee F. Callender

1 papers

TL;DR

Abstract

Authors

References35 items

GauGAN: semantic image synthesis with spatially adaptive normalization

Analyzing the role of model uncertainty for electronic health records

Deep Unsupervised Drum Transcription

Learning to Groove with Inverse Sequence Transformations

Bayesian Drum Transcription Based on Nonnegative Matrix Factor Decomposition with a Deep Score Prior

Semantic Image Synthesis With Spatially-Adaptive Normalization

Improving Peak-picking Using Multiple Time-step Loss Functions

Towards multi-instrument drum transcription

A Review of Automatic Drum Transcription

Efficient Neural Audio Synthesis

Natural TTS Synthesis by Conditioning Wavenet on MEL Spectrogram Predictions

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Onsets and Frames: Dual-Objective Piano Transcription

mixup: Beyond Empirical Risk Minimization

MDB Drums: An annotated subset of MedleyDB for automatic drum transcription

Regularizing and Optimizing LSTM Language Models

Drum transcription from polyphonic music with recurrent neural networks

Drum transcription using partially fixed non-negative matrix factorization

Deep Speech 2 : End-to-End Speech Recognition in English and Mandarin

Conditional Generative Adversarial Nets

An open-source drum transcription system for Pure Data and Max MSP

Decision Theory: Principles and Approaches

Further Steps towards Drum Transcription of Polyphonic Music

Theory of games and economic behavior

Increasing drum transcription vocabulary using data synthesis

Player Vs Transcriber: A Game Approach To Data Manipulation For Automatic Drum Transcription

From Labeled to Unlabeled Data - On the Data Challenge in Automatic Drum Transcription

Automatic Drum Transcription for Polyphonic Recordings Using Soft Attention Mechanisms and Convolutional Neural Networks

Drum Transcription via Joint Beat and Drum Modeling Using Convolutional Recurrent Neural Networks

Recurrent Neural Networks for Drum Transcription

Automatic Drum Transcription Using Bi-Directional Recurrent Neural Networks

Real-Time Transcription and Separation of Drum Recordings Based on NMF Decomposition

MIR_EVAL: A Transparent Implementation of Common MIR Metrics

DRUMKIT TRANSCRIPTION VIA CONVOLUTIVE NMF

ENST-Drums: an extensive audio-visual database for drum signals processing

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names