Pack: SALAMI Ground Truth stats

Description:

This very simple workflow was used in the analysis of the metadata produced by the human annotation phase of the SALAMI (Strucural Analysis of Large Amounts of Music Inforamtion) project.  The input data is the metadata available from

http://ddmal.music.mcgill.ca/salami/annotations

It produces a histogram of song durations and scatterplot of combined annotation time verus song duration.

The fields are

SONG_ID                   Unique identifier for piece of music

SOURCE                    Either Codaich, IA (Internet Archive), or RWC

ANNOTATOR1                ID number for first annotator

ANNOTATOR2                ID number for second annotator

SONG_DURATION             Duration of the piece, in seconds

SONG_TITLE                Title

ARTIST                    Artist

ANNOTATION_TIME1          Self-reported time to complete annotation for first annotator

ANNOTATION_TIME2          Self-reported time to complete annotation for second annotator

TEXTFILE1                 File path for first annotator's file

TEXTFILE2                 File path for second annotator's file

CLASS                     Broad genre (classical, jazz, popular, world, Live_Music_Archive, or unknown)

GENRE                     Narrow genre

SUBMISSION_DATE1          Date of submission of first annotation

SUBMISSION_DATE2          Date of submission of second annotation

XEQS1                     Was the first annotation converted automatically from X/= notation? X indicates yes; 0 indicates no.

XEQS2                     Was the second annotation converted automatically from X/= notation? X indicates yes; 0 indicates no.