2014:Audio Downbeat Estimation - Revision history

Matthew Davies at 16:14, 6 October 2014

2014-10-06T16:14:32Z

2014-07-22T18:33:44Z

2014-07-22T18:32:38Z

2014-07-22T18:30:08Z

2014-07-22T18:27:08Z

2014-07-22T18:26:33Z

2014-07-22T18:25:48Z

2014-07-22T18:24:57Z

2014-07-22T18:23:57Z

2014-07-22T18:20:59Z

@@ Line 25: / Line 25: @@
 '''Turkish Data'''
-The Turkish corpus is an extended version of the annotated data used in Srinivasamurthy et al. (2014). It includes '''85''' excerpts of one
+The Turkish corpus is an extended version of the annotated data used in Srinivasamurthy et al. (2014). It includes '''82''' excerpts of one
-minute length each, and each piece belongs to one of four
+minute length each, and each piece belongs to one of three
 rhythm classes that are referred to as usul in Turkish Art
-music. 30 pieces are in the 9/8-usul Aksak, 18 pieces
+music. 32 pieces are in the 9/8-usul Aksak, 20 pieces
-in the 10/8-usul Curcuna, 28 samples in the 8/8-usul
+in the 10/8-usul Curcuna, 30 samples in the 8/8-usul
-Düyek, and 9 samples in the 4/4 Sofyan.
+Düyek.
 '''Cretan Data'''
@@ Line 45: / Line 45: @@
 For further information see Hockman et al (2012).
-In total this makes '''1357''' excerpts (of which 259 are full length songs).
+In total this makes '''1354''' excerpts (of which 259 are full length songs).
 === Audio Formats ===

@@ Line 122: / Line 122: @@
 The different command lines to evaluate the performance of each parameter set over the whole database will be generated automatically from each line in the README file containing both '%input' and '%output' strings.
-== Evaluation Procedures ==
+== Evaluation Procedure ==
 For the evalution procedure we will use

@@ Line 150: / Line 150: @@
 '''Turkish''': 8 excerpts (2x Aksak, 2x Curcuna, 2x Düyek, and 2x Sofyan).
-Please note the audio files are in .flac format.
+Please note the audio files are in ''.flac'' format.
 For each audio file, e.g. ''hjdb1.fla''c there is a corresponding annotation file ''hjdb1.txt''
-This text file contains timestamps corresponding to beat annotations and a label to denote the position in the bar.
+Each ''.txt'' file contains timestamps corresponding to beat annotations and a label to denote the position in the bar.
 All beat times labelled '1' correspond to downbeats.

@@ Line 150: / Line 150: @@
 '''Turkish''': 8 excerpts (2x Aksak, 2x Curcuna, 2x Düyek, and 2x Sofyan).
-Please note the audio files are in .flac format
+Please note the audio files are in .flac format.
 == Time and hardware limits ==

@@ Line 132: / Line 132: @@
 A total of 20 beat and downbeat annotated 30s excerpts are available for participants.
-The data is available to download here:
+The data (47MB) is available to download here:
 https://www.music-ir.org/evaluation/MIREX/data/2014/downbeat/

@@ Line 136: / Line 136: @@
 https://www.music-ir.org/evaluation/MIREX/data/2014/downbeat/
-''User'': downbeat
+''User:'' downbeat
-''Password'': d0wn63at
+''Password:'' d0wn63at
 Due to the availability of the Ballroom and Beatles datasets, we only include examples of the remaining styles as follows:

@@ Line 9: / Line 9: @@
 Submitted algorithms will be evaluated in terms of their accuracy in finding downbeat locations (only) as annotated by musical experts across several diverse datasets.
-'''Update 22/07/14''' A small set of training data is now available. Please see
+'''Update 22/07/14''' A small set of training data is now available. Please see [[#Example_Data]]
 == Data ==