validating a technique for post hoc estimation of a
play

Validating a technique for post-hoc estimation of a listener's - PowerPoint PPT Presentation

Validating a technique for post-hoc estimation of a listener's focus in music structure analysis Jordan B. L. Smith National Institute of Advanced Industrial Science and Technology (AIST), Japan Elaine Chew Queen Mary University of London


  1. Validating a technique for post-hoc estimation of a listener's focus in music structure analysis Jordan B. L. Smith National Institute of Advanced Industrial Science and Technology (AIST), Japan Elaine Chew Queen Mary University of London

  2. What was the listener thinking? 
 (when they annotated musical structure) Melody 
 Harmony 
 Rhythm 
 Meter 
 Timbre 
 etc.

  3. Isaac Hayes, Run Fay Run

  4. Isaac Hayes, Run Fay Run

  5. RELATING GROUPING STRUCTURE TO MUSICAL FEATURES time A B C Audio-derived SSM

  6. RELATING GROUPING STRUCTURE TO MUSICAL FEATURES A Listener 1: AAB B C Audio-derived SSM Listener 2: ABB

  7. RELATING GROUPING STRUCTURE TO MUSICAL FEATURES A Listener 1: AAB Timbre SSM B C Audio-derived SSM Listener 2: ABB Harmony SSM

  8. RELATING GROUPING STRUCTURE TO MUSICAL FEATURES A Suggestion: We might explain the differences Listener 1: AAB Timbre SSM B between two annotations by matching them with the musical features. C Audio-derived SSM Listener 2: ABB Harmony SSM

  9. RELATING GROUPING STRUCTURE TO MUSICAL FEATURES Listener 1: AAB Timbre SSM Listener 2: ABB Harmony SSM

  10. Target SSM Feature-derived SSMs (original annotation) Segment-based Reconstruction masks components

  11. Target SSM Feature-derived SSMs (original N annotation) x 11 M 11 M 11 x 12 M 12 M 12 x 13 M 13 M 13 ... x 15 M 15 M 15 x 21 M 21 M 21 x 31 M 31 M 31 . . Segment-based Reconstruction . masks components . . . x 71 M 71 M 71 x 75 M 75 M 75

  12. Target SSM Feature-derived SSMs (original N annotation) x 11 M 11 M 11 x 12 M 12 M 12 x 13 M 13 M 13 ... x 15 M 15 M 15 x 21 M 21 M 21 find x 31 M 31 M 31 . that minimizes . Segment-based Reconstruction . masks components . . . x 71 M 71 M 71 x 75 M 75 M 75

  13. WHICH FEATURE BEST EXPLAINS THE ANALYSIS?

  14. CRITICISM ➤ Model not validated with real-world data: ➤ I.e., analyses paired with attention state of listener

  15. EYE TRACKING

  16. EAR TRACKING

  17. CRITICISM ➤ Model not validated with real-world data: ➤ I.e., analyses paired with attention state of listener ➤ One source of data: ➤ Ran experiment testing whether focus of listener could affect perception of grouping

  18. STIMULUS EXAMPLES � � � � � � � � � � � �� �� � � �� �� � � �� � � � � � � � � � � � � � � � � � � � � � � Melody ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � � � � � � Organ

  19. STIMULUS EXAMPLES � � � � � � � � � � � �� �� � � �� �� � � �� � � � � � � � � � � � � � � � � � � � � � � A B B Melody ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � A A B � � � � � Organ

  20. STIMULUS EXAMPLES � � � � � � � � � � � �� �� � � �� �� � � �� � � � � � � � � � � � � � � � � � � � � � � A B B Melody ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � A A B � � � � � � � Organ � � � � � � � � � � � � � � � � � � � � � �� �� � � � �� �� � � � �� � � � � � � � � Timbre ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � � � � Organ Harpsichord � � � � � � � � � � � � 3

  21. STIMULUS EXAMPLES � � � � � � � � � � � �� �� � � �� �� � � �� � � � � � � � � � � � � � � � � � � � � � � A B B Melody ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � A A B � � � � � � � Organ � � � � � � � � � � � � � � � � � � � � � �� �� � � � �� �� � � � �� � � � � � � � � Timbre ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� A A B � � � � � � � � � � � Harmony AAB � � � � A B B Organ Harpsichord � � � � � � � � � � � � 3

  22. STIMULUS EXAMPLES � � � � � � � � � � � �� �� � � �� �� � � �� � � � � � � � � � � � � � � � � � � � � � � A B B Melody ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� � � � � � � � � � � � Harmony AAB � A A B � � � � � � � Organ � � � � � � � � � � � � � � � � � � � � � �� �� � � � �� �� � � � �� � � � � � � � � Timbre ABB 
 � � � � � � � � � � � �� �� � � �� �� �� � � �� A A B � � � � � � � � � � � Harmony AAB � � � � A B B Organ Harpsichord � � � � � � � � � � � � 3 ➤ Experiment outcome: giving listeners a distractor task (to fix attention) affected AAB vs. ABB decisions ➤ “Does this chord progression occur in the stimulus?” → prefer AAB ➤ Note: sometimes, organizing features are convolved

  23. CRITICISM ➤ Model not validated with real-world data: ➤ I.e., analyses paired with attention state of listener ➤ One source of data: ➤ Ran experiment testing whether focus of listener could affect perception of grouping ➤ Is model needlessly complicated? ➤ Try simple correlation instead

  24. Target SSM Feature-derived SSMs (original N annotation) QP approach: - find sum of reconstruction components to generate target SSM - interpret coefficients as relevance Segment-based Reconstruction masks components Correlation approach: - compute correlation between feature SSMs and masks - interpret correlation as relevance

  25. EXPERIMENT ➤ Use algorithm to estimate relevance of features to analysis of experiment stimuli ➤ If the most relevant feature is the feature that changed, count correct

  26. RESULTS chance level What feature was found to be most relevant? (i.e., greatest QP weights or correlation)

  27. SUMMARY ➤ Past work: ➤ 1. Algorithm to estimate feature relevance from analyses (Smith and Chew, 2013) ➤ 2. Artificial stimuli from psych experiment studying attention ➤ Current work: ➤ Use stimuli (2) to validate algorithm (1) and new variations

  28. SUMMARY ➤ Algorithm performs above baseline… ➤ …but artificial context makes non-perfect performance a disappointment! ➤ Random effect of small set of stimuli ➤ Mismatch between features and musicological interpretations ➤ Future work: ➤ re-optimize (or train) model with more features, more artificial stimuli ➤ apply algorithm to investigate trends in large datasets THANKS!

  29. ADVERTORIAL SUPPLEMENT

  30. REFERENCES ➤ Jordan B. L. Smith and Elaine Chew. Using Quadratic Programming to estimate feature relevance in structural analyses of music. In Proceedings of the ACM International Conference on Multimedia , 113–122, Barcelona, Spain, 2013. ➤ Jordan B. L. Smith. Explaining listener differences in the perception of musical structure. Ph.D. thesis. Queen Mary University of London. See Chapter 6: “The effect of attention on grouping decisions.”

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend