[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[help-GIFT] Extensions to MRML, web page bugs and other comments

From: Reiss. Dr J
Subject: [help-GIFT] Extensions to MRML, web page bugs and other comments
Date: Wed, 20 Aug 2003 19:05:36 +0100

I work with music information retrieval (MIR) and there has been a lot of 
debate recently on how to benchmark MIR systems. See the Evaluation Project on for background info, and specifically [1] for how it relates 
to MRML. Most of the MIR community would like to see a TREC-like evaluation. It 
seems that the best way for us to do this (without TREC support) is following 
the example of Benchathlon. 
To do this, I would need to thoroughly understand MRML and implement it and 
extend it to support musical queries (query-by-humming systems, symbolic as 
well as audio data, monophonic and polyphonic similarity, ...).
So I'm mainly seeking advice from those involved with MRML. I've read the 
documentation but its not obvious what to do. I've also tried working with some 
of the online demonstrations and related software resources but I've had 
problems with all of them. 
For instance,

1. According to the MRML website, the main MRML resource, libMRML, is 
supposedly on the GIFT page, 
<>  . But on that page, under “Downloading the 
GIFT” it states “The latest version can be downloaded at” ! 
I then found the GIFT package on a different page, but the documentation on 
libMRML, and how to use it outside of GIFT, is very sparse.

2. The jMRML, 
<>  , page does not seem to be 

3. fer-de-lance has a software depository with no files and only a handful of 
two year old messages in the mailing list.

4. The php client has the image loading facility disabled and hence does 

5. Also, many of the viper links, such as 
<> , are dead.


So I guess my questions are as follows: whats the best way for me to get up to 
speed on MRML quickly? How should I augment it to support musical queries? Is 
there an MRML tutorial (this would be quite different from the specification 
document)? Is this the right approach? Would anyone be interested in 
collaborating on this ;) ?


[1] Reiss, J.D. and M. Sandler. MIR Benchmarking: Lessons Learned from the 
Multimedia Community. in The 26th Annual International ACM SIGIR Conference. 
2003. Toronto, Canada. 


Joshua D. Reiss


Office 0207-882-5528

Fax 0207-882-7997 

Dept of Electronic Engineering

Queen Mary, University of London

Mile End Road, London E14NS 

United Kingdom

Email address@hidden



reply via email to

[Prev in Thread] Current Thread [Next in Thread]