Overview

ACE XML is a set of XML file formats that are designed to meet the special representational needs of music information retrieval (MIR) research in general, and automatic music classification research in particular. The ACE XML formats are designed to represent a wide range of musical information clearly and simply using formally structured frameworks that are flexible and extensible. A special focus is placed on representing information related to musical machine learning, such as features, instance labels and class ontologies. It is hoped that ACE XML will eventually be adopted by the MIR community as a standardized format for use in data representation associated with automatic music classification.

ACE XML 2.0, the subject of this site, is a fundamental redesign of the earlier ACE XML 1.1 standard, which will soon be deprecated. ACE XML 1.1 is currently fully supported by jMIR, a general-purpose music information retrieval software suite. jMIR is in the process of being updated to support ACE XML 2.0.

This page contains updates on ongoing developments relating to the ACE XML 2.0 standard as well as documentation, sample ACE XML 2.0 files and associated softare tools. The information presented on this site is provided for the purpose of interaction and consultation with the NEMA community, and is to be considered preliminary until officially published.

Current Status of ACE XML 2.0 Development

The ACE XML 2.0 standard is now finalized. Suggestions for further modifications are still very welcome, but will be incorporated into a future ACE XML 2.1 update to the standard instead of ACE XML 2.0. Efforts will of course be made to ensure backwards compatibility between the two. The NEMA Wiki is a venue for discussion on ongoing developments.

Current work is focusing on:

1. Developing Java software to support ACE XML 2.0, including libraries and utilities for parsing, writing, merging and otherwise processing ACE XML 2.0, as well as data structures with a simple API that can be used to facilitate the incorporation of ACE XML 2.0 compatibility into other software.
2. Upgrading jMIR with this software so that it can be used to read and write ACE XML 2.0. jMIR currently only supports the older ACE XML 1.1 formats.
3. Writing Python, C++, Matlab, etc. implementations of these libraries and utilities.
4. Developing code for translating between ACE XML 2.0 and other languages used in machine learning, such as Weka ARFF.

The ACE XML File Types

• Feature Value: These files express feature values extracted from specific instances.
• Feature Description: These files express abstract information about features themselves (but not extracted feature values).
• Instance Label: These files associate class labels with specific instances and their subsections. They can be used to express either ground truth labels or predicted labels output by a classification system.
• Class Ontology: These files express relationships between classes. They can be used to simply catalogue candidate class labels, or for more sophisticated purposes such as expressing weighted ontological relationships between classes.
• Project: These files indicate associations between related groups of ACE XML files and other resources.

ACE XML 2.0 Documentation

• ACE XML 2.0 Specification Document
  • Contains the details of the the ACE XML 2.0 file standards as well as basic explanations and examples
  • The code examples presented in this document may be downloaded below from the ACE XML 2.0 Sample Files section
• ACE XML 2.0 draft chapter from Cory McKay's dissertation
  • Contains information about both ACE XML 2.0 and the older ACE XML 1.1
  • Contains more background information than the Specification Document
• ACE XML 2.0 data model
  • Graphically illustrates the ACE XML 2.0 data model as well as some information about the serialisation

ACE XML 2.0 Sample Files

• ACE XML 2.0 Feature Value sample file
• ACE XML 2.0 second Feature Value sample file demonstrating use of URI linking
• ACE XML 2.0 Feature Description sample file
• ACE XML 2.0 Instance Label sample file
• ACE XML 2.0 Class Ontology sample file

• ACE XML 2.0 Feature Value examples from the ACE XML 2.0 Specification Document
• ACE XML 2.0 Feature Description example from the ACE XML 2.0 Specification Document
• ACE XML 2.0 Instance Label example from the ACE XML 2.0 Specification Document
• ACE XML 2.0 Class Ontology example from the ACE XML 2.0 Specification Document

Associated Software

An updated version of the ACE and jMIRUtilities software is available that now includes functionality for loading, viewing, editing, processing and saving ACE XML 1.1 files (ACE XML 2.0 functionality is NOT yet implemented in this release), as well as functionality for dealing with ACE XML Project Files and ACE XML Zip Files. These versions of the software are very much still under development, and as such have a number of unresolved bugs and features remaining to be implemented. The earlier stable version of ACE is still available on the ACE SourceForge page, but does not include the GUI that is of particular use with respect to ACE XML files. The version posted below is still in a pre-publication state, and is made available only in order to demonstrate the general approaches being used and the API that is being made available:

• ACE XML Software Dev Version 0.1 (PRE-RELEASE VERSION)

There are explanatory web pages available providing an ACE manual and details on the ACE GUI.

Updated software associated with ACE XML 2.0 will be posted here as it is released.

Development Team

ACE XML 2.0 is designed, implemented and maintained by the Distributed Digital Music Archives and Libraries (DDMAL) group at McGill University as part of the NEMA project. The people involved are:

• Ichiro Fujinaga
  • NEMA Co-Principal Investigator
  • Associate Professor, Music Technology, McGill University
• Cory McKay
  • NEMA Developer
  • Creator of jMIR and ACE XML 1.1 and 2.0
  • Doctoral Candidate, Music Technology, McGill University
• J. Ashley Burgoyne
  • NEMA Developer
  • Doctoral Candidate, Music Technology, McGill University
• Jessica Thompson
  • NEMA Programmer
  • Undergraduate Student, Music Technology and Psychology, McGill University

ACE XML Development History

• October 2009
  • Paper on ACE XML 2.0 presented at ISMIR 2009
    • Using ACE XML 2.0 to store and share feature, instance and class data for musical classification
  • Paper on ACE 2.0 presented at ISMIR 2009
    • Additions and improvements to the ACE 2.0 music classifie
• August 2009
  • ACE XML 2.0 data model posted
  • jMIR demo presented at ICMC 2009
    • jMIR: Tools for automatic music classification
• June 2009
  • jMIR and ACE presented at the Joint Conference on Digital Libraries
    • jMIR and ACE XML: Tools for performing and sharing research in automatic music classification
• May 2009
  • ACE XML 2.0 Specification Document and more sample ACE XML 2.0 files posted
• April 2009
  • Final amendments and extensions incorporated into ACE XML 2.0
    • Dissertation chapter draft and sample ACE XML 2.0 files updated to reflect changes
  • All uri elements now have a predicate attribute in order to enable linking to external documents (e.g., Music Ontology ontologies) via RDF-like triples
  • Feature Value files now have an extractor element and associated fname attribute in order to make it possible to specify the feature extractor used to extract each feature
  • "Feature Definition" files were renamed "Feature Description" files in order to more accurately reflect their nature
• March 2009
  • Deadline for final proposed modifications to the ACE XML 2.0 standard, with modifications proposed after this date to be instead included in a future ACE XML 2.1 release
• February 2009
  • Completed implementation of improvements to ACE XML 2.0 requested at UIUC NEMA meeting
    • Dissertation chapter draft and sample ACE XML 2.0 files updated to reflect changes
  • Posting of the pre-release for the updated ACE XML Software Dev Version 0.1, which includes updates relating to ACE XML 1.1 functionality that will later be ported to ACE XML 2.0
• January 2009
  • Preliminary ACE XML 2.0 specification released to the NEMA community
  • ACE XML 2.0 draft chapter from Cory McKay's dissertation and sample ACE XML 2.0 files posted
  • NEMA meeting at UIUC, improvements suggested to ACE XML 2.0
• September 2008
  • Paper describing research that took advantage of the expressivity of ACE XML files presented at the ISMIR conference
    • Combining Features Extracted From Audio, Symbolic and Cultural Sources
  • ACE XML 1.1 discussed with NEMA members at ISMIR conference
  • Release of prototype ACE GUI for displaying and manually editing ACE XML files
  • Release of Project file extension to ACE XML 1.1
  • Release of ACE ZIP extension to ACE XML 1.1
• March 2008
  • Began work on incorporating ACE XML into the NEMA framework
• September 2005
  • Paper related to ACE XML 1.1 presented at the ISMIR conference
    • ACE: A Framework for Optimizing Music Classification
  • Public release of the ACE XML 1.1 standard
• March 2005
  • Internal release of ACE XML 1.0 at McGill University

Publications Relating to ACE XML

McKay, C., J. A. Burgoyne, and I. Fujinaga. 2009. jMIR and ACE XML: Tools for performing and sharing research in automatic music classification. Presented at the ACM/IEEE Joint Conference on Digital Libraries Workshop on Integrating Digital Library Content with Computational Tools and Services, University of Texas, Austin, USA. 19 June 2009.

McKay, C., J. A. Burgoyne, J. Thompson, and I. Fujinaga. 2009. Using ACE XML 2.0 to store and share feature, instance and class data for musical classification. Proceedings of the International Society for Music Information Retrieval Conference. 303–8.

McKay, C., and I. Fujinaga. 2009. jMIR: Tools for automatic music classification. Proceedings of the International Computer Music Conference.

Thompson, J., C. McKay, J. A. Burgoyne, and I. Fujinaga. 2009. Additions and improvements to the ACE 2.0 music classifier. Proceedings of the International Society for Music Information Retrieval Conference. 435–40.

McKay, C., and I. Fujinaga. 2008. Combining features extracted from audio, symbolic and cultural sources. Proceedings of the International Conference on Music Information Retrieval. 597–602.

McKay, C., R. Fiebrink, D. McEnnis, B. Li, and I. Fujinaga. 2005. ACE: A framework for optimizing music classification. Proceedings of the International Conference on Music Information Retrieval. 42–9.

Questions and Comments

cory.mckay@mail.mcgill.ca

-top of page-