Difference between revisions of "2008:Query by Tapping"

From MIREX Wiki
(Task description)
(Task description)
Line 5: Line 5:
 
* '''Test database''': 129 ground-truth mono MIDI files.
 
* '''Test database''': 129 ground-truth mono MIDI files.
 
* '''WAV Quries''': 272 query files to retrieve 129 known target from collection. So far, 1~6 human assessors have listened and tapped a 15 seconds query rhythm from beginning for each target songs.
 
* '''WAV Quries''': 272 query files to retrieve 129 known target from collection. So far, 1~6 human assessors have listened and tapped a 15 seconds query rhythm from beginning for each target songs.
* '''CSV Quries''': 129 query files in CSV format which are drawn from original MIDI. It include time interval (in ms) between first 26 NoteOn event separate with comma. Rest note and time interval smller than 100ms will be discarded.
+
* '''CSV Quries''': 129 query files in CSV format which are drawn from original MIDI. It include time interval (in ms) between first 25 NoteOn event separate with comma. Rest note and time interval smller than 100ms will be discarded.
 
* '''Evaluation''': Mean Reciprocal rank. Return top 10 candidates for each query file.
 
* '''Evaluation''': Mean Reciprocal rank. Return top 10 candidates for each query file.
  

Revision as of 03:00, 15 August 2008

Overview

The main purpose of QBT(Query by Tapping) is to evaluate MIR system in retrieval ground-truth MIDI files by the tapping rhythm. This task provides query rhythm files both in WAV and CSV(symbolic) formate. Evaluation database and query files can be download from http://210.68.135.13/ki/QBT.rar

Task description

  • Test database: 129 ground-truth mono MIDI files.
  • WAV Quries: 272 query files to retrieve 129 known target from collection. So far, 1~6 human assessors have listened and tapped a 15 seconds query rhythm from beginning for each target songs.
  • CSV Quries: 129 query files in CSV format which are drawn from original MIDI. It include time interval (in ms) between first 25 NoteOn event separate with comma. Rest note and time interval smller than 100ms will be discarded.
  • Evaluation: Mean Reciprocal rank. Return top 10 candidates for each query file.

Data processing proposal for calling formats

Indexing the MIDIs collection

Indexing_exe <var1> <var2>

where

<var1>==<directory_of_MIDIs> 
<var2>==<indexing_files_output_and_working_directory>

Running for the query files

Running_exe <var3> <var4> <var5>

where

<var3>==<directory_of_indexed_file> 
<var4>==<directory_of_query_rhythm_files> 
<var5>==<answer_file_output_directory> 

Data processing for output answer file formats

The answer file for each run would look like:

Q001:0003,0567,0999,<insert X more responses>,XXXX
Q002:0103,0567,0998,<insert X more responses>,XXXX
Q00X:0002,0567,0999,<insert X more responses>,XXXX

Each line represents to each of the queries in a given task run.

Submission closing date

22th August 2008.

Interested Participants

  • Shu-Jen Show Hsiao(show.cs95g at nctu.edu.tw)
  • Rainer Typke: I would be interested if the query data can also be made available in symbolic form so we can see what part of the performance comes from good onset detection from audio, and what comes from a good matching algorithm.