Go to CSIRO.AU Home Page
Go to Mathematical & Information Sciences home page
HOME RESEARCH CONTACT DOING BUSINESS POSITIONS VACANT MEDIA EVENTS PUBLICATIONS PRODUCTS

 

Australian audio analysis tools

Maaate!


Documentation

 

Related Information

 

The Australian audio analysis toolkit

Maaate (pronounce: ma:a:it) is a C++ toolkit to parse and analyse audio data in the compressed/frequency domain. It is designed to be able to support any kind of sound file such as MPEG-1/2 Layers 1-3 as well as raw files. It makes the frequency samples, and other useful pre-processed features as well as the file format specific fields (such as the encoded fields of an MPEG audio stream) accessible. A module interface allows content analysis functions (modules) such as silence detection to work on the extracted information.
Source code is available under the GNU General Public License.

MPEG Maaate is designed in tiers: at the bottom sits the file parsing tier, which currently contains an MPEG file parser ("mpeg" directory). The next tier implements a generic interface to the frequency domain samples and access to extracted and pre-processed information regardless of the audio file format in libMaaateP ("tier1" directory). The tier thereafter implements an interface to the content analysis on the extracted information libMaaateA ("tier2" directory). The analysis functions themselves are stored in dynamically loadable plugin modules. A standard set is included in the "plugins" (libMaaateM) and "plugins_brief" (libMaaateM_brief) directory. Applications sit on top of these tiers using any modules required.

For more indepth information on Maaate read the Maaate FAQ. It contains an overview of the design and an introduction on how to add code. If you would like to add new analysis functions (modules) to Maaate, read the module information.


Package content

The distribution of Maaate contains the following libraries:

  • a C++ library for the MPEG parser (libMaaateMPEG).
  • a C++ library for the generic field/frequency sample access (libMaaateP) including a C API.
  • a C++ library for the generic module access (libMaaateA) including a C API.
  • two C++ libraries implementing some fundamental modules and an easy API to access them (libMaaateM, libMaaateM_brief). libMaaateM_brief provides a backward compatible and shorter module API.
and the following command-line programs:
  • a C++ command-line application to access MPEG fields via tier 1 and demonstrate the usage of the tier1 API (inspectMPaudio).
  • a C++ command-line application to apply analysis functions and demonstrate the usage of the tier2 API (analyseSDaudio) and a C version of it (canalyseSDaudio).
  • a C++ command-line MPEG player using Maaate (MaaatePlay) and a C version of it (cMaaatePlay).
In a separate distribution you may find a sample GUI interface to Maaate called bewdy. It demonstrates a few of the abilities of Maaate visually.

Credits

The system was implemented by Silvia Pfeiffer, Thomas Vincent, Stefan Kudras and Conrad Parker (GUI) from the Analytic Audio Systems research group, CSIRO (Commonwealth Scientific and Industrial Research Organisation) Mathematical and Information Sciences Division (CMIS) in Australia. CMIS has worked on analysing MPEG audio since 1997. Other group members that have layed grounds for Maaate by performing MPEG audio analysis include: Stephen Barrass , Jordi Robert-Ribes, Brian Kim.

The code for the MPEG parser is based on the MPEG-1 audio reference C-implementation of the ISO MPEG working group. It contains some of their functions but has been redesigned from scratch with audio analysis in mind. A lot of code has been written to make the API as simple as possible.

Special thanks to Craig Lindley for drawing the Australian Mate.


Mailing-List/Contact

To be updated on new releases, bugfixes and the development status, join maaate@cmis.csiro.au by sending an email to majordomo@cmis.csiro.au containing "subscribe maaate" in the body and respond to the automatic mailer that majordomo sends back. In case of problems contact maaate-approval@nsw.cmis.csiro.au.

If you require more information, have found a bug, would like to help in developing etc. you may also contact:

Silvia Pfeiffer
CSIRO, Mathematical and Information Sciences
Locked Bag 17, North Ryde
NSW 2113, Australia
Phone: +61 2 9325 3146
Fax:     +61 2 9325 3200

 

Applications Download Code FAQ Sceen Snapshots Project Team

last updated 10/01/2002
Silvia.Pfeiffer@csiro.au

© Copyright 1997-2002, CSIRO Australia
Use of this web site and information available from
it is subject to our
Legal Notice and Disclaimer and Privacy Statement