deletions | additions       diff --git a/chordmir.tex b/chordmir.tex  index b8ed80d..95ba34c 100644  --- a/chordmir.tex  +++ b/chordmir.tex 
...
Chord identification in audio is a challenging task that has been the subject of much MIR research. A common difficulty in the task of identifying chords is the difficulty in establishing ground truth data upon which automatically identified chords can be compared (ref). Another difficulty is that a progression of chords may involve chords that are overlapping, have notes that linger or are anticipated, may involve notes outside of the tuning of the chromatic scale, or may involve unknown relies on extracting information about underlying notes, their pitches, interactions, and timbres, and classifying associated  chord qualities. (List study here % A common difficulty  in the task of identifying chords is the difficulty in establishing ground truth data upon  which people were paid) automatically identified chords can be compared (ref). Another difficulty is that a progression of chords may involve chords that are overlapping, have notes that linger or are anticipated, may involve notes outside of the tuning of the chromatic scale, or may involve unknown chord qualities.  % (List study here in which people were paid)  %  This paper is not focused on obtaining exact chord progression analyses from audio files; rather, how well can common music information retrieval tasks be performed on musical harmony with inexact data, and how can these differences be reconciled?