Denktank 2010

414 views

Published on

Presentatie onderzoek Olmo Cornelis - 17 dec 2010 - IPEM Gent

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
414
On SlideShare
0
From Embeds
0
Number of Embeds
3
Actions
Shares
0
Downloads
1
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Denktank 2010

  1. 1. Theoretic and artistic research studying opportunities of symbiosis of Western and non-Western musical idiom (2008-2014) Olmo Cornelis
  2. 2. Outline <ul><li>Introduction </li></ul><ul><li>Background </li></ul><ul><li>Problem </li></ul><ul><li>Definition </li></ul><ul><li>Methodology </li></ul><ul><li>Case studies </li></ul><ul><li>Conclusion </li></ul>
  3. 3. Background <ul><li>Olmo Cornelis </li></ul><ul><li>- Musicology @ Ghent University </li></ul><ul><li>- Composition @ University College Ghent </li></ul><ul><li>- PhD @ University College Ghent / Ghent University </li></ul><ul><li>Previous work: DEKKMMA </li></ul><ul><li> “ Digitization of the Ethnomusicological Sound Archive of the Royal Museum for Central Africa” </li></ul>
  4. 4. DEKKMMA <ul><li>Digitization of the sound archive: </li></ul><ul><ul><ul><li>Audio : 50.000 sound recordings – 3.000 hours of music - 33.000 items digitized. </li></ul></ul></ul><ul><ul><ul><li>Meta-data : 35.000 items digitized </li></ul></ul></ul><ul><ul><ul><li>Provide contextual data </li></ul></ul></ul><ul><li>Development of database and website </li></ul><ul><li>Exploration of MIR-techniques. </li></ul><ul><li>http://music.africamuseum.be/ </li></ul>
  5. 6. <ul><li>Deliver detailed descriptions of music from another culture </li></ul><ul><ul><li>Different from Western musical idiom </li></ul></ul><ul><ul><li>(till now) Western concepts were used </li></ul></ul><ul><ul><li>Correct descriptions are needed for: </li></ul></ul><ul><ul><ul><li>Online collections </li></ul></ul></ul><ul><ul><ul><li>Research </li></ul></ul></ul><ul><ul><ul><li>Respect cultural heritage </li></ul></ul></ul>Actual problem I
  6. 7. <ul><li>How to transfer this knowledge into an artistic production? </li></ul><ul><li>How to transform musical elements into another culture? </li></ul><ul><li>How to enrich a compositional language with specific ethnic musical characteristics? </li></ul><ul><li>How to create a personal compositional idiom with implicit influences? </li></ul>Actual problem II
  7. 8. Work Path <ul><li>Study of problem: research question </li></ul><ul><li>Definition (definiëring en afbakening) </li></ul><ul><li>Theoretical analysis / set up framework (methodology) </li></ul><ul><li>Cooperation musicologist and computer engineer </li></ul><ul><li>Testing existing applications </li></ul><ul><li>Build platform for ‘objective’ analysis </li></ul><ul><li>Case studies </li></ul><ul><li>Result (diverse output) </li></ul>Listen and encounter Define research question Set up frame work (methodology) Experimental phase Composition Experiment Evaluation Composition Concert
  8. 9. Method <ul><ul><li>Ethnic music? </li></ul></ul><ul><ul><li>Three categories : </li></ul></ul><ul><ul><li>Ethnic music (in strict sense) </li></ul></ul><ul><ul><li>From cultures without written tradition </li></ul></ul><ul><ul><li>Non-Western classical music </li></ul></ul><ul><ul><ul><ul><li>Religious and court music from cultures with a written </li></ul></ul></ul></ul><ul><ul><ul><ul><li>tradition outside Europe </li></ul></ul></ul></ul><ul><ul><li>Folk music </li></ul></ul><ul><ul><ul><ul><li>From written culture, but not part of a classical tradition </li></ul></ul></ul></ul>
  9. 10. Historical tendencies <ul><li>Initially </li></ul><ul><ul><li>Aversion towards music </li></ul></ul><ul><ul><ul><li>Music connected with dance and rituals </li></ul></ul></ul><ul><li>After WO2 </li></ul><ul><ul><li>Collecting systematically </li></ul></ul><ul><ul><li>Describing semantically </li></ul></ul><ul><li>21th Century </li></ul><ul><ul><li>Digitization (conservation, accessibility) </li></ul></ul><ul><ul><li>Computational ethnomusicology </li></ul></ul>
  10. 11. <ul><li>Objective measurement: </li></ul><ul><li>Not using Western musicological assumptions </li></ul><ul><li>Approach actual characteristics </li></ul><ul><li>Music Information Retrieval </li></ul><ul><ul><li>Signal based research </li></ul></ul><ul><ul><ul><li>Computer assisted analysis </li></ul></ul></ul>Method
  11. 12. Music Information Retrieval <ul><li>Nowadays: </li></ul><ul><li>- audio is a digital object </li></ul><ul><li>- huge online collections </li></ul><ul><li>- personal consumption </li></ul><ul><li>=> Music discovery! </li></ul>
  12. 13. Music Information Retrieval <ul><li>- MIR: describes content of music </li></ul><ul><li>- Many applications </li></ul><ul><li>- Search/retrieve/organize </li></ul><ul><li>Three levels </li></ul><ul><ul><ul><li>Signal => physical </li></ul></ul></ul><ul><ul><ul><li>Symbolic => musical symbols </li></ul></ul></ul><ul><ul><ul><li>Semantic => metadata </li></ul></ul></ul>
  13. 15. Music Information Retrieval <ul><li>Many applications: </li></ul><ul><li>- Identification, plagiarism detection, copyright monitoring, version detection, melody finder, alignment, recommendation, sounds like, genre detection, segmentation, instrumentation, organization by metadata, mood, artist… </li></ul>
  14. 16. <ul><ul><li>domain specific techniques are required to approach the particular characteristics and constraints of different ethnic music styles. </li></ul></ul><ul><ul><li>Need for an interdisciplinary research field. </li></ul></ul><ul><ul><li>Automated approaches are needed because of large archives. </li></ul></ul><ul><ul><li>large variability of music, users, search intentions and expectations. </li></ul></ul><ul><ul><li>Metadata, content-based description, musical concepts are fundamentally different from Western standards . </li></ul></ul>(MIR) considerations
  15. 17. Work Path <ul><li>Study of problem: research question </li></ul><ul><li>Definition (definiëring en afbakening) </li></ul><ul><li>Theoretical analysis / set up framework (methodology) </li></ul><ul><li>Cooperation musicologist and computer engineer </li></ul><ul><li>Testing existing applications </li></ul><ul><li>Build platform for ‘objective’ analysis </li></ul><ul><li>Case studies </li></ul><ul><li>Result </li></ul>Listen and encounter Define research question Set up frame work (methodology) Experimental phase Composition Experiment Evaluation Composition Concert
  16. 18. Artistic research <ul><li>Composition (imho): </li></ul><ul><ul><li>Innovative element </li></ul></ul><ul><ul><li>Content </li></ul></ul><ul><ul><li>Coherence </li></ul></ul><ul><ul><li>Aesthetic component </li></ul></ul><ul><li>Ethnic music: kaleidoscope of sounds </li></ul><ul><li>Symbioses: added value of two components </li></ul><ul><li>=> Develop a personal idiom with implicit influence </li></ul>
  17. 19. Musical elements
  18. 20. Musical elements Text Timbre Pitch Time Dynamics
  19. 21. Musical elements
  20. 22. Musical elements
  21. 23. Musical elements <ul><li>Axes : </li></ul><ul><li>- Explicit <---------------------------------> Implicit </li></ul><ul><li>- Effect <---------------------------------> Concept </li></ul><ul><li>- Form <-----------------------------------> Musical </li></ul><ul><li>- Music an Sich <-------------------> Functional </li></ul><ul><li>- High level <---------------------> Low level MP </li></ul><ul><li>- Physical <--------------------------------> Virtual </li></ul><ul><li>- … <-------------------> … </li></ul>
  22. 24. Musical elements
  23. 25. Musical elements
  24. 26. Approach Artistic Research <ul><li>1. Acquire knowledge and specificities of ethnic music </li></ul><ul><li>=> scientific study </li></ul><ul><li>2. Set up compositional experiments </li></ul><ul><li>=> in exhaustive range of possibilities </li></ul><ul><li>3. Create framework to work in </li></ul><ul><li>=> see previous slides </li></ul><ul><li>4. Create (with implicit ethnic influence) </li></ul><ul><li>=> in a personal style </li></ul>
  25. 27. Results
  26. 28. Framework TARSOS <ul><li>Pre-processing: </li></ul><ul><ul><li>Audio signal improvement: </li></ul></ul><ul><ul><ul><li>Noise reduction </li></ul></ul></ul><ul><ul><ul><li>Signal optimization </li></ul></ul></ul><ul><ul><li>Source separation (ONO2010) </li></ul></ul><ul><ul><li>Band pass filters </li></ul></ul><ul><li>Processing: </li></ul><ul><ul><li>(existing) algorithms </li></ul></ul><ul><ul><li>Annotation optimization </li></ul></ul><ul><li>Post-processing: </li></ul><ul><ul><li>Interpretation </li></ul></ul><ul><ul><li>Visualization </li></ul></ul><ul><ul><li>Export </li></ul></ul><ul><ul><ul><li>Data </li></ul></ul></ul><ul><ul><ul><li>Graphical </li></ul></ul></ul>
  27. 29. <ul><li>SIGNAL LEVEL </li></ul><ul><li>Test of available pitch trackers </li></ul><ul><li>Representation of output </li></ul><ul><li>Real-time analysis </li></ul><ul><li>Optimization results </li></ul><ul><li>SYMBOLIC LEVEL </li></ul><ul><li>Peak extraction </li></ul><ul><li>Peak analysis </li></ul><ul><ul><li>Correlation </li></ul></ul><ul><li>OUTPUT </li></ul><ul><li>Sonification </li></ul><ul><li>Export </li></ul>Framework TARSOS
  28. 30. Demo
  29. 31. Exploring African Tone Scales <ul><li>Audio set : </li></ul><ul><li>901 audio files from the RMCA-archive </li></ul><ul><li>Solo pieces played on: </li></ul><ul><ul><li>Zither </li></ul></ul><ul><ul><li>Flute </li></ul></ul><ul><ul><li>Thumb piano </li></ul></ul><ul><ul><li>Musical bow </li></ul></ul>
  30. 32. Exploring African Tone Scales Remarkable finding 1: the evolution of the increasing number of pitches in the scale through time.
  31. 33. Exploring African Tone Scales Remarkable finding 2: Increase of ‘regular’ intervals in recent years. Remarkable finding 3: Specific ‘irregular’ intervals (e.g. 160 or 370 cents) Interval < 1960 1960-75 > 1975 min.2nd 1,46 1,87 2,20 maj.2nd 1,57 1,71 5,20 min.3rd 2,26 3,39 2,84 maj.3rd 1,25 1,28 2,78 4th/5th 2,55 2,56 5,31 sum 9,10 10,81 18,33
  32. 34. Pitch (melody) <ul><li>Language steers melos: </li></ul><ul><li>Ex.: Lisaka: </li></ul><ul><li>march </li></ul><ul><li>promise </li></ul><ul><li>poison </li></ul><ul><li>=> mono syllabic </li></ul><ul><li>=> anti-melismatic </li></ul><ul><li>=> verse: no identical melody </li></ul><ul><li>=> polyphony: parallel (no counterpoint) </li></ul>Ngomi ngomi:
  33. 35. Rhythm and tempo <ul><li>Preliminary research: </li></ul><ul><li>Experiment: Tempo extraction </li></ul><ul><li>beat / pulse annotation </li></ul><ul><li>by human and by pc </li></ul><ul><li>Goal: </li></ul><ul><li>- human tempo perception </li></ul><ul><li>- set of audio with ground truth </li></ul><ul><li>- test quality beat algorithms </li></ul>
  34. 38. Tempo <ul><li>Set up: </li></ul><ul><li>- 70 audio fragments, 20 sec each </li></ul><ul><li>10 beat trackers </li></ul><ul><li>10 persons </li></ul><ul><li>Task: tap the tempo </li></ul>
  35. 39. Tempo 1 2 5 3 3 2 2 meter 1 1 4 0 0 0 0 other relation 0 3 0 3 1 0 0 binary/ternary 2 2 2 2 2 4 0 tempo octave 6 3 3 4 6 4 9 identic tempo consistency 107-121 92 92 144 63 65 129 median tempo 65 92 44 96 125 130 129 subject 9 94 69 122 73 63 64 129 subject 8 221 93 92 144 63 132 129 subject 7 117 92 92 47 63 124 129 subject 6 121 238 178 144 69 x 129 subject 5 115 140 130 142 129 127 129 subject 4 107 121 46 95 63 65 129 subject 3 122 186 111 144 63 65 129 subject 2 53 46 92 72 190 65 129 subject 1 no.1 no.69 no.49 no.31 no.30 no.20 no.6 song nr
  36. 40. Microtiming
  37. 41. Rhythm <ul><li>Western conceptually different: </li></ul><ul><li>Divisive, Z/L, symmetry </li></ul><ul><li>In 8: 3/3/2, 3/2/3, 2/3/3, 1/3/2/2, 1/2/1/2/2 </li></ul><ul><li>Ewe drumming </li></ul>
  38. 42. EME <ul><li>- Electronic sound and music controller , based on human movement and using custom made wireless motion sensors </li></ul><ul><li>- Hardware technology, software applications, music theory and concepts of Embodied Music Cognition </li></ul>
  39. 43. EME <ul><li>Applications: </li></ul><ul><li>demo , for musicians and non-musicians, a real-time sound control </li></ul><ul><li>Artistic performance , i.c. Con Golese, a piece for tape and voice, where an opera singer could control the tape composition through her gestures while performing. Bodily expressions - often gestures - ensued on the vocal part, feel natural and cause a closer blending of voice and tape. The possible real-time interaction of a performer with a tape composition creates new opportunities for electronic compositions. </li></ul>
  40. 46. <ul><li>Audio might not be sufficient to describe and/or retrieve ethnic music: </li></ul><ul><ul><li>-> The idea of separating sound from the rest of its physical environment (movement, smell, taste, colour) may well be a weird “invention” of the West. </li></ul></ul><ul><ul><li>=> We cannot understand ethnic music correct without its social function and context ! </li></ul></ul>
  41. 47. Next <ul><li>Composition </li></ul><ul><ul><li>Dag van het Onderzoek (March 2011) </li></ul></ul><ul><li>New proposals (COST, Tetra, FET Open) </li></ul><ul><li>Pitch </li></ul><ul><ul><li>Finalize application </li></ul></ul><ul><ul><li>Article methodology TARSOS (nearly finished) </li></ul></ul><ul><ul><li>Article application TARSOS (several options) </li></ul></ul><ul><li>Rhythm </li></ul><ul><ul><li>Put documentation and experiment into article (advanced) </li></ul></ul><ul><ul><li>Set up framework for TARSOS </li></ul></ul>
  42. 48. Dissemination <ul><li>Publications </li></ul><ul><li>Compositions </li></ul><ul><li>Activities </li></ul><ul><li>Website </li></ul><ul><li>Doctoral Schools </li></ul>
  43. 49. Publications <ul><li>2010 - Cornelis O., Lesaffre M., Moelants D., Leman M., 'Access to ethnic music: advances and perspectives in content-based music information retrieval' , Signal Processing, Elsevier, 2010 - Lidy T., Silla C.N., Cornelis O., Gouyon F., Rauber A. ,Kaestner C.A.A., Koerich A.L., 'Western vs. Ethnic Music: On the Suitability of State-of-the-art Music Retrieval Methods for Analyzing, Structuring and Accessing Ethnic Music Collections' , Signal Processing, Elsevier, 2010 - Cornelis O., 'Het etnomusicologisch klankarchief van het KMMA, digitaliseren en beyond…' , in Handboek Muzikaal Erfgoed (Ed. Schreurs E.), Resonant, 2010, in press </li></ul><ul><li>2009 - Cornelis O., Moelants D., Leman M., 'Global access to ethnic Music: the next big challenge?' Proceedings ISMIR 2009, Kobe, Japan. http://www.columbia.edu/~tb2332/fmir/Papers/Moelants-fmir.pdf - Moelants D., Cornelis O., Leman M., 'Exploring African tone scales' , Proceedings ISMIR, (Kobe, Japan), 2009 - Lesaffre, M., Cornelis, O., Moelants, D., Leman, M., 'Integration of Music Information Retrieval techniques into the practice of ethnic music collections' , Proceedings Unlocking Audio 2 (London), 2009 </li></ul><ul><li>2008 - Antonopoulos I., Cornelis O., Moelants D., Leman M., Pikrakis A., 'Music Retrieval by Rhythmic Similarity applied on Greek and African Traditional Music' , Proceedings, ISMIR 2007 - De Mey M., Cornelis O., Leman M., 'The IPEM_EME: a Wireless Music Controller for Real-time Music Interaction' , Proceedings ICAD 2008, Paris - Cornelis O., De Mey M., Leman M., 'EME: a wireless music controller for Real-Time Music Interaction' , Proceedings ARTECH 2008, Porto </li></ul>
  44. 50. Compositions <ul><li>Shaken </li></ul><ul><li>Con Golese </li></ul><ul><li>Nelumbo </li></ul><ul><li>EME experiment: </li></ul><ul><ul><li>ICAD </li></ul></ul><ul><ul><li>ARTECH </li></ul></ul><ul><ul><li>RESONANT </li></ul></ul><ul><ul><li>E-Dinges </li></ul></ul>
  45. 51. Doctoral Schools
  46. 52. Website <ul><li>http://webs.hogent.be/olmocornelis/index.html </li></ul><ul><li>http://tarsos.0110.be </li></ul>
  47. 53. 30% <ul><li>Coaching students in Master </li></ul><ul><li>Follow up English Master students </li></ul><ul><li>Guide Potential PhD students </li></ul><ul><li>New means: </li></ul><ul><ul><li>Hercules </li></ul></ul><ul><ul><li>TETRA </li></ul></ul><ul><ul><li>COST </li></ul></ul><ul><ul><li>FET Open </li></ul></ul>
  48. 54. <ul><li>[email_address] </li></ul><ul><li>http://tarsos.0110.be </li></ul><ul><li>http://music.africamuseum.be </li></ul><ul><li>http://webs.hogent.be/olmocornelis </li></ul>

×