2010:Audio Key Detection - Revision history

Kriswest at 09:18, 5 June 2010

2010-06-05T09:18:35Z

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:58:17Z

‎Evaluation Procedures

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:57:02Z

‎Evaluation Procedures

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:55:22Z

‎Evaluation Procedures

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:54:35Z

‎Evaluation Procedures

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:53:50Z

‎Evaluation Procedures

AndreasEhmann: /* Evaluation Procedures */

2010-05-26T18:53:06Z

‎Evaluation Procedures

AndreasEhmann: /* Relevant Development Collections */

2010-05-26T18:48:43Z

‎Relevant Development Collections

AndreasEhmann at 18:48, 26 May 2010

2010-05-26T18:48:05Z

AndreasEhmann: /* Output File Format (Audio Key Detection) */

2010-05-26T18:42:43Z

‎Output File Format (Audio Key Detection)

@@ Line 11: / Line 11: @@
 * CD-quality (PCM, 16-bit, 44100 Hz)
 * single channel (mono)
 == Submission Format ==
@@ Line 59: / Line 80: @@
   foobar(.1,'%input','%output')
 === README File ===
 A README file accompanying each submission should contain explicit instructions on how to to run the program (as well as contact information, etc.). In particular, each command line to run should be specified, using %input for the input sound file and %output for the resulting text file.
@@ Line 71: / Line 112: @@
   matlab -r "foobar(.1,'%input','%output');quit;"
-== Evaluation Procedures ==
+== Time and hardware limits ==
-The error analysis will center on comparing the key identified by the algorithm to the actual key of the piece. The key of the piece is the one defined by the composer in the title of the piece. We will then determine how "close" each identified key is to the corresponding correct key. Keys will be considered as "close" if they have one of the following relationships: distance of perfect fifth, relative major and minor, and parallel major and minor. A correct key assignment will be given a full point, and incorrect assignments will be allocated fractions of a point according to the following table:
+Due to the potentially high number of participants in this and other audio tasks, hard limits on the runtime of submissions will be imposed.
-{|border="1"
+Friday 4th June 2010
-The points are counted over all files and averaged. The number of correctly identified keys as well as the distribution of the errors is also reported.
+== Submission closing date ==

← Older revision		Revision as of 18:58, 26 May 2010
Line 87:		Line 87:
	\|Other\|\|0.0		\|Other\|\|0.0
	\|}		\|}
		+
		+	The points are counted over all files and averaged. The number of correctly identified keys as well as the distribution of the errors is also reported.

@@ Line 74: / Line 74: @@
 The error analysis will center on comparing the key identified by the algorithm to the actual key of the piece. The key of the piece is the one defined by the composer in the title of the piece. We will then determine how "close" each identified key is to the corresponding correct key. Keys will be considered as "close" if they have one of the following relationships: distance of perfect fifth, relative major and minor, and parallel major and minor. A correct key assignment will be given a full point, and incorrect assignments will be allocated fractions of a point according to the following table:
-{|
+{|border="1"
-|'''Relation to Correct Key''' ||    '''Points'''
+|'''Relation to Correct Key''' ||'''Points'''
 |-
-|Same||    1.0
+|Same||1.0
 |-
-|Perfect fifth||    0.5
+|Perfect fifth||0.5
 |-
-|Relative major/minor||    0.3
+|Relative major/minor||0.3
 |-
-|Parallel major/minor||    0.2
+|Parallel major/minor||0.2
 |-
-|Other||    0.0
+|Other||0.0
 |}

@@ Line 75: / Line 75: @@
 {|
-|'''Relation to Correct Key''' |||| '''Points'''
+|'''Relation to Correct Key''' ||    '''Points'''
 |-
-|Same|||| 1.0
+|Same||    1.0
 |-
-|Perfect fifth|||| 0.5
+|Perfect fifth||    0.5
 |-
-|Relative major/minor|||| 0.3
+|Relative major/minor||    0.3
 |-
-|Parallel major/minor|||| 0.2
+|Parallel major/minor||    0.2
 |-
-|Other|||| 0.0
+|Other||    0.0
 |}

@@ Line 75: / Line 75: @@
 {|
-|'''Relation to Correct Key''' || '''Points'''
+|'''Relation to Correct Key''' |||| '''Points'''
 |-
-|Same|| 1.0
+|Same|||| 1.0
 |-
-|Perfect fifth|| 0.5
+|Perfect fifth|||| 0.5
 |-
-|Relative major/minor|| 0.3
+|Relative major/minor|||| 0.3
 |-
-|Parallel major/minor|| 0.2
+|Parallel major/minor|||| 0.2
 |-
-|Other|| 0.0
+|Other|||| 0.0
 |}

@@ Line 5: / Line 5: @@
 == Data ==
 === Collections ===
-The collection used for this year's evaluation is the same as the one used in 2005. It consists of 1252 classical music audio pieces rendered from MIDI using the timidity MIDI synthesizer.
+The collection used for this year's evaluation is the same as the one used in 2005. It consists of 1252 classical music audio pieces rendered from MIDI using the timidity MIDI synthesizer. The ground-truth key is drawn from the title of the piece. The entire piece is not used, but rather the first 30 seconds. This is done because usually the beginnings of pieces are in the labeled key before they possibly deviate due to key modulation.
 === Audio Formats ===
@@ Line 26: / Line 26: @@
 === Output Data ===
-The audio key detection algorithms will return the estimated key in an ASCII text file for each input .wav audio file. The specification of this output file is immediately below.
+The audio key detection algorithms will return the estimated key in an individual ASCII text file for each input .wav audio file. The specification of this output file is immediately below.
 === Output File Format (Audio Key Detection) ===

@@ Line 37: / Line 37: @@
   C    major
-−
- or
+or
-−
-  G# minor
+  G#   minor
 === Algorithm Calling Format ===