DOCX

Dissertation

By Lee Stewart,2014-12-06 15:49
11 views 0
They must be robust to compression and noise in the transmission channel.

    Machine Annotation of

    Traditional Irish Dance Music

    PhD Thesis

    Bryan Duggan MSc, BSc.

    School of Computing

    Dublin Institute of Technology

    Kevin St. Dublin 8, Ireland.

    Supervisors:

    Prof. Brendan O’ Shea

    Dr. Mikel Gainza

    Prof. Pádraig Cunningham

    Dublin Institute of Technology, School of Computing

    March 2009

    Declaration

    I certify that this thesis which I now submit for examination for the award of Doctor of Philosophy, is entirely my own work and has not been taken from the work of others save and to the extent that such work has been cited and acknowledged within the text of my work.

    This thesis was prepared according to the regulations for postgraduate study by research of the Dublin Institute of Technology and has not been submitted in whole or in part for an award in any other Institute or University.

    The Institute has permission to keep, to lend or to copy this thesis in whole or in part, on condition that any such use of the material of the thesis be duly acknowledged.

Signed

______________________

    Bryan Duggan

Date

______________________

    i

    Abstract

    Estimates put the canon of traditional Irish dance tunes at at least seven thousand compositions. The literature attributes this to the geographic isolation of rural communities which developed their own repertoire of tunes. Musicians playing traditional music have a personal repertoire of up to one thousand tunes. Given this diversity, a common problem faced by musicians and ethnomusicologists is identifying tunes from recordings. This is evident even in the number of commercial recordings whose title is gan ainm (without name).

     The work presented in this PhD thesis attempts to solve this problem by developing a Content Based Music Information Retrieval (CBMIR) system adapted to the characteristics of traditional Irish music. The thesis includes a comprehensive review of the domain of traditional Irish music and presents three chapters of related work in the fields of feature extraction, melodic similarity and music information retrieval. A system is presented called MATT2 (Machine Annotation of Traditional Tunes) whose primary goal is to annotate recordings of traditional Irish dance music with useful metadata including tune names. MATT2 incorporates a number of novel algorithms for transcription of traditional music and for adapting melodic similarity measures to expressiveness in the playing of traditional music. It makes use of an onset detection function developed for the playing of traditional music on woodwind instruments such as the concert flute and tin-whistle. It uses a novel transcription algorithm based on Brendan Breathneach’s observations about the transcription of

    traditional Irish music which provides transposition invariance for the keys and modes used to play traditional music. It incorporates a new algorithm for dealing with ornamentation notes and accommodating "the long note" in traditional music called Ornamentation Filtering. It makes use of publicly available collections of traditional music available in the ABC notation. It uses a matching algorithm tolerant to errors which aligns short queries with longer strings from a corpus of known tunes, meaning that the algorithm can match entire tunes, incipits and phrases from any part of tune with equal success. The matching algorithm has also been adapted to take account of phrasing and reversing effects. A new algorithm is presented called TANSEY (Turn

    ANnotation from SEts using SimilaritY profiles) which annotates sets of tunes played segue as is the custom in traditional Irish dance music.

    ii

     The work presented in this thesis is validated in experiments using 130 real-world field recordings of traditional music from sessions, classes, concerts and commercial recordings. Test audio includes solo and ensemble playing on a variety of instruments recorded in real-world settings such as noisy public sessions. Results are reported using standard measures from the field of information retrieval (IR) including accuracy, error, precision and recall and the system is compared to alternative approaches for CBMIR common in the literature.

    iii

    Buíochas

    Buíochas le mo feitheoirí, an tOllamh Brendan O'Shea, an Dochtúir Mikel Gainza ón DIT agus an tOllamh Pádraig Cunningham ó UCD as ucht an tacaíocht, treoir agus spreagadh a thug siad le ceithre bliana anuas. Ba mhaith liom buíochas speisialta a ghabháil le mo chara agus feitheoir, an tOllamh O' Shea, a raibh an radharc agus an fís aige an clár PhD a chur le chéile roinnt blianta ó shin. Buíochas freisin le hInstitiúid Teicneolaíochta Átha Cliath a thug an t-airgead le go bhféadfainn freastal ar chomhdhálacha agus a chuir an t-am ar fáil chun an obair seo a chríochnú. Buíochas le mo chomhpháirtithe i Scoil na Ríomhaireachta, an grúpa intleachta saorga, an grúpa cluichí ríomhairí agus an grúpa clostrealamh dhigiteach, go speisialta le Dr John Kelleher, Dr Brian McNamee, Dr Sarah Jane Delaney, Hugh McAtamney, Damian Gordon, Ronan Fitzpatrick and Dan Barry as ucht an t-aiseolas dearfach a chur ar fáil. Buíochas le Michael Porter agus Dave Carroll a léith an leabhar seo dom.

    Buíochas leis na ceoltóirí agus na múinteoirí Maria Murphy, Rob O'Connor, Aideen Downs, Emily Sakier Donoho, Colm Ó Laoghóg, Markus Asunta, Frank Slocket, Donal Regan, Eamon Cotter, Harry Bradley, Catherine McEvoy, Dave Sheridan Treasa Harkin agus le Mick Mulvey a thug dom téip den cheoltóir Packie Duignan.

    Buíochas le mo thuismitheoirí agus mo mhuintir don cheol thar na mblianta. Ar deireadh buíochas le mo pháirtí ghrámhar Derek.

    iv

    Acknowledgements

    I would like to thank my supervisors Professor Brendan O' Shea and Dr Mikel Gainza of the DIT and Professor Pádraig Cunningham of UCD for their unwavering support, guidance and encouragement over the past four years. I would also like to pay tribute to my friend and supervisor, Professor O' Shea, who had the vision to initiate and develop the PhD program in the School of Computing several years ago. Thanks also to the Dublin Institute of Technology which kindly funded my attendance at various conferences and provided me with teaching relief to complete this work. Thank you to my colleagues at the DIT School of Computing, the AI Group, the Experimental Gaming Group and the Audio Research Group particularly Dr John Kelleher, Dr Brian McNamee, Dr Sarah Jane Delaney, Hugh McAtamney, Damian Gordon, Ronan Fitzpatrick and Dan Barry for their positive feedback at various stages of this work. Thank you to Michael Porter and Dave Carroll for proof reading.

    Thanks to the many musicians and teachers who consented to being recorded as part of this research: Maria Murphy, Rob O'Connor, Aideen Downs, Emily Sakier Donoho, Colm Logue, Markus Asunta, Frank Slocket, Donal Regan, Eamon Cotter, Harry Bradley, Catherine McEvoy, Treasa Harkin, Dave Sheridan and to Mick Mulvey for providing archive recordings of the flute player Packie Duignan.

     Thanks to my parents and family for all the music over the years. Finally thanks to my loving partner Derek.

    v

    Table of Contents 1 INTRODUCTION ................................................................................................................... 1 1.1 RESEARCH AIMS .................................................................................................................... 3 1.2 USE CASES ............................................................................................................................ 4 1.3 ORIGINAL CONTRIBUTION ...................................................................................................... 6 1.4 ORGANISATION ..................................................................................................................... 7 2 TRADITIONAL IRISH MUSIC ........................................................................................... 10 2.1 TUNE TYPES ........................................................................................................................ 12 2.1.1 Reel ........................................................................................................................... 12 2.1.2 Jig ............................................................................................................................. 13 2.1.3 Hornpipe ................................................................................................................... 14 2.1.4 Polka ......................................................................................................................... 14 2.1.5 Mazurka .................................................................................................................... 14 2.1.6 Slow Air ..................................................................................................................... 15 2.2 MODES & TEMPO ................................................................................................................. 16 2.3 TUNE TITLES........................................................................................................................ 17 2.4 INSTRUMENTS ..................................................................................................................... 19 2.4.1 Tin-whistle ................................................................................................................. 19 2.4.2 Flute .......................................................................................................................... 20 2.4.3 Fiddle (Violin) ........................................................................................................... 23 2.4.4 Uilleannn Pipes ......................................................................................................... 24 2.4.5 Harp .......................................................................................................................... 26 2.4.6 Free-reed instruments ................................................................................................ 26 2.4.7 Percussion ................................................................................................................. 27 2.4.8 Lilting ........................................................................................................................ 28 2.5 SOLO VERSUS ENSEMBLE PLAYING ....................................................................................... 28 2.6 COLLECTIONS ...................................................................................................................... 30 2.7 COLLECTIONS IN ELECTRONIC FORMAT ................................................................................. 32 2.8 MUSICAL CREATIVITY .......................................................................................................... 33 2.9 STYLE & EXPRESSIVENESS IN TRADITIONAL IRISH DANCE MUSIC ........................................... 35 2.9.1 Ornamentation ........................................................................................................... 36 2.9.2 Phrasing .................................................................................................................... 40 2.9.3 Regional Styles .......................................................................................................... 41 2.9.4 Expressive examples .................................................................................................. 44 2.10 CONCLUSIONS ................................................................................................................. 45 3 FEATURES OF MUSIC........................................................................................................ 47 3.1 ONSET-DETECTION .............................................................................................................. 47 3.2 PITCH .................................................................................................................................. 54 3.3 TIMBRE ............................................................................................................................... 59 3.4 LOUDNESS ........................................................................................................................... 60 3.5 RHYTHM ............................................................................................................................. 60 3.6 STRUCTURE ......................................................................................................................... 61 3.7 CONCLUSIONS ..................................................................................................................... 62 4 MELODIC SIMILARITY ..................................................................................................... 64 4.1 MELODIC CONTOUR (PARSONS CODE) ................................................................................... 64 4.2 IMPLICATION-REALISATION .................................................................................................. 65 4.3 TRANSPORTATION DISTANCE ................................................................................................ 66 4.4 EDIT (LEVENSHTEIN) DISTANCE ........................................................................................... 68 4.5 HIDDEN MARKOV MODELS .................................................................................................. 73 4.6 CONCLUSIONS ..................................................................................................................... 75 5 CONTENT BASED MUSIC INFORMATION RETRIEVAL ............................................. 78 5.1 SEARCHING SYMBOLIC REPRESENTATIONS ............................................................................ 78 5.2 SEARCHING AUDIO DATA ..................................................................................................... 83 5.3 HYBRID APPROACHES .......................................................................................................... 86

    vi

    5.4 CONCLUSIONS ..................................................................................................................... 92 6 MACHINE ANNOTATION OF TRADITIONAL TUNES (MATT2) ................................. 94 6.1 SYSTEM DESIGN ................................................................................................................... 95 6.2 ONSET DETECTION ............................................................................................................... 96 6.3 PITCH DETECTION ................................................................................................................ 99 6.4 COMPENSATING FOR EXPRESSIVENESS IN QUERIES .............................................................. 101

    6.4.1 Ornamentation Filtering .......................................................................................... 102 6.5 BREATH DETECTION........................................................................................................... 107 6.6 PITCH SPELLING ................................................................................................................. 108 6.7 CORPUS NORMALISATION ................................................................................................... 110 6.8 MATCHING ........................................................................................................................ 112 6.9 INTERFACE ........................................................................................................................ 112 6.10 CONCLUSIONS ............................................................................................................... 114 7 EVALUATION .................................................................................................................... 115 7.1 EXPERIMENT ..................................................................................................................... 115 7.2 RESULTS ........................................................................................................................... 119 7.3 SIGNIFICANCE ................................................................................................................... 120 7.4 CONCLUSIONS ................................................................................................................... 124 8 ANNOTATING SETS OF TUNES PLAYED SEGUE ....................................................... 126 8.1 SETS OF TRADITIONAL IRISH DANCE TUNES ......................................................................... 127 8.2 TANSEY (TURN ANNOTATION FROM SETS USING SIMILARITY PROFILES) ALGORITHM ...... 128 8.3 EXPERIMENT ..................................................................................................................... 132 8.4 RESULTS ........................................................................................................................... 134 8.5 CONCLUSIONS ................................................................................................................... 136 9 CONCLUSIONS & FUTURE WORK ................................................................................ 138 9.1 CONCLUSIONS ................................................................................................................... 138 9.2 FUTURE WORK ................................................................................................................... 142 APPENDIX A TEST AUDIO LISTING.................................................................................... 145 APPENDIX B ABC NOTATION .............................................................................................. 149 APPENDIX C EXAMPLE TUNES IN ABC FORMAT (NORBECK 2007) ............................ 155 APPENDIX D EXAMPLE TUNES AFTER NORMALISATION ........................................... 158 APPENDIX E EXTRACT FROM A DISCUSSION ON THE TUNE "DOWN THE BROOM"

    FROM THESESSION.ORG (ACCESSED 22 AUGUST, 2008) .................................................. 160 APPENDIX F RESULTS OF MC-ED, TI-ED AND MATT2 (SECTIONS 7, 7.2 AND 7.3) .... 163 APPENDIX G RESULTS OF TANSEY EVALUATION DESCRIBED IN SECTIONS 8.3 AND

    8.4 .................................................................................................................................................. 166 APPENDIX H EXAMPLE TRANSCRIPTIONS IN ABC NOTATION DISCUSSED IN

    SECTION 2.9.4 ............................................................................................................................. 171 REFERENCES ............................................................................................................................. 174

    vii

    Table of Figures

    FIGURE 1: THE M-AUDIO MICRO TRACK II DIGITAL AUDIO FIELD RECORDER (M-AUDIO 2008) ............ 1 FIGURE 2: WAX CYLINDER RECORDINGS OF PIPER PATSY TOUHEY (COURTESY OF THE IRISH

    TRADITIONAL MUSIC ARCHIVE) ................................................................................................. 2 FIGURE 3: THE REEL "COME WEST ALONG THE ROAD" (O'NEILL 1907) (SEE ALSO FIGURE 14 AND FIGURE

    41 AND FIGURE 43)................................................................................................................... 13 FIGURE 4: THE HORNPIPE "THE PLAINS OF BOYLE"(O'NEILL 1907) ..................................................... 14 FIGURE 5: THE SLOW AIR "TÁIMSE IM' CHODLADH" (THESESSION.ORG 2007) ....................................... 15 FIGURE 6: FINGERING CHART FOR THE WOODEN FLUTE/TIN-WHISTLE................................................... 21 FIGURE 7: KEYED AND UNKEYED CONCERT FLUTES ............................................................................ 22 FIGURE 8: FIDDLE PLAYER SIOBHAN PEOPLES .................................................................................... 23 FIGURE 9: THE MAIN COMPONENTS OF THE UILLEANN PIPES (VALLELY 1999) ...................................... 24 FIGURE 10: FINGERING CHART FOR THE UILLEANN PIPES CHANTER ...................................................... 25 FIGURE 11: BODHRÁN PLAYER PETER BLANEY ................................................................................... 27 FIGURE 12: FIDDLE PLAYER COLM LOGUE, THE AUTHOR, AND FLUTE PLAYER PATSY HANLEY AT AN

    INFORMAL SESSION AT FLEADH CHEOIL NA HÉIREANN 2008 ...................................................... 29 FIGURE 13: POLICE CHIEF FRANCIS O' NEILL AND THE COVER OF O' NEILL'S "THE DANCE MUSIC OF

    IRELAND" ................................................................................................................................ 31 FIGURE 14: THE TUNE "COME WEST ALONG THE ROAD" IN THE ABC FORMAT (NORBECK 2007) (SEE

    ALSO FIGURE 3, FIGURE 41 AND FIGURE 43) .............................................................................. 32 FIGURE 15: AN EXAMPLE OF A RUN IN ABC NOTATION (SEE ALSO APPENDIX B AND

    HTTP://WWW.COMP.DIT.IE/BDUGGAN/MUSIC) ......................................................................... 39 FIGURE 16: EXAMPLES OF ORNAMENTATION IN PIANO ROLL FORMAT (SEE ALSO

    HTTP://WWW.COMP.DIT.IE/BDUGGAN/MUSIC FOR EXAMPLE AUDIO RECORDINGS) .................... 39 FIGURE 17: GEOGRAPHIC ORIGIN OF REGIONAL STYLE (SOURCE: AUTHOR BASED ON (KEEGAN 1992)) . 42 FIGURE 18: JOHN MCKENNA (FLUTE) AND MICHAEL GAFFNEY (BANJO) .............................................. 43 FIGURE 19: WAVEFORM PLOT OF A CONCERT FLUTE PLAYING THE NOTES A TO G LEGATO.................... 48 FIGURE 20: WAVEFORM PLOT OF A PIANO PLAYING THE NOTES A TO G................................................ 49 FIGURE 21: ONSET DETECTION FUNCTION (ODF) FOR A MUSICAL PHRASE CALCULATED USING ODCF

    IMPLEMENTED BY THE AUTHOR IN JAVA (CHAPTER 6) ................................................................ 53 FIGURE 22: A CONCERT FLUTE PLAYING THE NOTE D4 ........................................................................ 55 FIGURE 23: THE NORMALISED ABSOLUTE FFT VALUES OF THE SIGNAL FROM FIGURE 22 ...................... 55 FIGURE 24: HANNING FUNCTION ........................................................................................................ 57 FIGURE 25: A FRAME OF AUDIO FROM FIGURE 22 WINDOWED BY THE HANNING FUNCTION FROM FIGURE

    24 ............................................................................................................................................ 58 FIGURE 26: THE FIRST 2 BARS FROM THE TUNE "BANISH MISFORTUNE" IN ABC FORMAT AND IN MUSIC

    NOTATION, WITH THE CORRESPONDING PARSONS' CODE ............................................................. 65 FIGURE 27: THE THEMEFINDER USER INTERFACE ................................................................................ 79 FIGURE 28: MUSICIANS IN A SESSION COMPARE TUNES USING TUNEPAL .............................................. 82 FIGURE 29: SCREENSHOTS OF TUNEPAL RUNNING ON A WINDOWS MOBILE SMARTPHONE ................... 82 FIGURE 30: THESESSION.ORG USER INTERFACE. (SEE ALSO APPENDIX E) ............................................. 83 FIGURE 31: SHAZAM AUDIO FINGERPRINTING RUNNING ON AN IPHONE (SHAZAM 2008) ....................... 85 FIGURE 32: THE "PICARD" MUSICBRAINZ CLIENT ............................................................................... 86 FIGURE 33: MELDEX INTERFACE. A USER CAN PLAY A PART OF MELODY OR RECORD A QUERY FOR

    TRANSCRIPTION ........................................................................................................................ 88 FIGURE 34: TUNEBOT USER INTERFACE .............................................................................................. 91 FIGURE 35: HIGH LEVEL DIAGRAM OF THE MATT2 TUNE ANNOTATION SYSTEM .................................. 95 FIGURE 36: EXTRACT FROM THE AUTHORS IMPLEMENTATION OF A TIME DOMAIN COMB FILTER

    (EQUATION 1) IN JAVA.............................................................................................................. 97 FIGURE 37: SIGNAL AND ODF PLOTS OF THE FIRST BAR OF THE TUNE "THE BOYNE HUNT" ................... 99 FIGURE 38: EXTRACT FROM THE AUTHOR'S FREQUENCY DOMAIN, HARMONICITY BASED PITCH DETECTOR

    CODE IN JAVA......................................................................................................................... 101 FIGURE 39: PSEUDOCODE FOR THE QUAVER DURATION CALCULATION ALGORITHM ............................ 103 FIGURE 40: HISTOGRAM OF CANDIDATE NOTE DURATIONS IN SECONDS, FROM A 25 SECOND PHRASE FROM

    THE TUNE "THE KILMOVEE JIG" .............................................................................................. 104 FIGURE 41: NORMALISATION STAGES FOR THE A PART OF THE TUNE "COME WEST ALONG THE ROAD".

    SEE ALSO FIGURE 3, FIGURE 14 AND FIGURE 43 ....................................................................... 111 FIGURE 42: SCREENSHOT OF MATT2 ............................................................................................... 113

    viii

null

Report this document

For any questions or suggestions please email
cust-service@docsford.com