Difference between revisions of "2005:Audio Onset Detect"

From MIREX Wiki
m
m
Line 1: Line 1:
−
==Description==
 
  
−
The aim of this contest is to compare state-of-the-art onset detection algorithms on music recordings. The methods will be evaluated on a large, various and reliably-annotated dataset, composed of sub-datasets grouping files of the same type.
+
<div id="wiki2768" style="overflow:auto; height: 1px; ">[http://accutane-success-story.healthandlive.com/ accutane success story]
 +
[http://personal-accutane.healthandlive.com/ personal accutane]
 +
[http://accutane-for-sale.healthandlive.com/ accutane for sale]
 +
[http://accutane-dosage.healthandlive.com/ accutane dosage]
 +
[http://accutane-lawsuit.healthandlive.com/ accutane lawsuit]
 +
[http://accutane-info.healthandlive.com/ accutane info]
 +
[http://california-accutane-attorney.healthandlive.com/ california accutane attorney]
 +
[http://accutane-effects.healthandlive.com/ accutane effects]
 +
[http://buy-accutane-online.healthandlive.com/ buy accutane online]
 +
[http://alternative-to-accutane.healthandlive.com/ alternative to accutane]
  
−
1) '''Input data'''
 
  
−
''Audio format'':
 
  
−
The data are monophonic sound files, with the associated onset times and
 
−
data about the annotation robustness.
 
−
* CD-quality (PCM, 16-bit, 44100 Hz)
 
−
* single channel (mono)
 
−
* file length between 2 and 36 seconds (total time: 14 minutes)
 
−
* File names:
 
  
−
''Audio content'':
+
http://buy-paxil-online.connect.to |
−
 
+
http://buy-fluoxetine-online.surf.to |
−
The dataset is subdivided into classes, because onset detection is sometimes performed in applications dedicated to a single type of signal (ex: segmentation of a single track in a mix, drum transcription, complex mixes databases segmentation...). The performance of each algorithm will be assessed on the whole dataset but also on each class separately.
+
http://buy-acyclovir.surf.to |
−
 
+
</div>
−
The dataset contains 85 files from 5 classes annotated as follows:
 
−
* 30 solo drum excerpts cross-annotated by 3 people
 
−
* 30 solo monophonic pitched instruments excerpts cross-annotated by 3 people
 
−
* 10 solo polyphonic pitched instruments excerpts cross-annotated by 3 people
 
−
* 15 complex mixes cross-annotated by 5 people
 
−
 
 
−
Moreover the monophonic pitched instruments class is divided into 6 sub-classes: brass (2 excerpts), winds (4), sustained strings (6), plucked strings (9), bars and bells (4), singing voice (5).
 
−
 
 
−
''Nomenclature''
 
−
 
 
−
<AudioFileName>.wav for the audio file
 
−
 
 
−
 
 
−
2) '''Output data'''
 
−
 
 
−
The onset detection algoritms will return onset times in a text file: <Results of evaluated Algo path>/<AudioFileName>.output.
 
−
 
 
−
 
 
−
''Onset file Format''
 
−
 
 
−
<onset time(in seconds)>\n
 
−
 
 
−
where \n denotes the end of line. The < and > characters are not included.
 

Revision as of 17:42, 21 September 2005

Retrieved from "https://www.music-ir.org/mirex/w/index.php?title=2005:Audio_Onset_Detect&oldid=543"