-
Unmixer: Loop Extraction with Repetition, with Dr. Jordan Smith and Tim de Reuse
Podcast: Play in new window | Download (Duration: 1:00:10 — 55.7MB)
Subscribe: Spotify | RSS | More
Music technology PhD Candidate Tim de Reuse recommends “Unmixer: An Interface for Extracting and Remixing Loops” by Jordan Smith,Yuta Kawasaki, and Masataka Goto, published in the proceedings of ISMIR 2019. Tim and Finn interview Jordan about the origins of this project, the algorithm behind the loop extraction, the importance of repetition in music, and the creative and playful applications of Unmixer.
Note: This conversation was recorded in December 2019. Techically issues with some tracks contributed to delays. Apologies for the choppy audio quality.
Time Stamps
- [0:01:40] Project Summary
- [0:05:05] Demonstration of Unmixer
- [0:14:27] Origins of the UnMixer project
- [0:19:44] Factorisation algorithm
- [0:28:37] Computational and musical objectives for factorisation
- [0:36:15] The Unmixer web interface
- [0:41:30] 2nd Demonstration, parameters and track selection
- [0:49:13] What Unmixer tells us about music
Show notes
- Recommended article:
- Smith, J, Kawasaki, Y, & Goto, M. (2019) Unmixer: An Interface for Extracting and Remixing Loops. Proceedings of 20th ISMIR meeting, Delft Netherlands.
- UnMixer website: https://unmixer.ongaaccel.jp/
- Project webpage
- Interviewee: Dr. Jordan BL Smith, Research Scientist at Tik Tok.Website, twitter
- Co-host: PhD Candidate Tim de Reuse, website, twitter
- Papers cited in the discussion:
- Smith, J. B., & Goto, M. (2018, April). Nonnegative tensor factorization for source separation of loops in audio. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 171-175). IEEE.
- Schmidhuber, J. (2009). Simple algorithmic theory of subjective beauty, novelty, surprise, interestingness, attention, curiosity, creativity, art, science, music, jokes. Journal of SICE, 48(1).
- Rafii, Z., & Pardo, B. (2012). Repeating pattern extraction technique (REPET): A simple method for music/voice separation. IEEE transactions on audio, speech, and language processing, 21(1), 73-84.
- Music sampled:
- Daft Punk, Random Access Memories (2013): Doing it Right (ft. Panda Bear)
- Martin Solveig & Dragonette, Smash (2011): Hello – Single Edit
- Mura Masa, Soundtrack To a Death (2014): I’ve Never Felt So Good
- Other references:
- Madeon’s Adventure Machine
- Chocolate Rain by Tay Zonday
Credits
The So Strangely Podcast is produced by Finn Upham, 2020. The closing music includes a sample of Diana Deutsch’s Speech-Song Illusion sound demo 1.
-
ISMIR 2019 Conference sampler
Podcast: Play in new window | Download (Duration: 40:37 — 37.7MB)
Subscribe: Spotify | RSS | More
This episode brings recommendations from the 2019 ISMIR conference at TUDelft in the Netherlands. A number of contributors, old and new, highlighted papers that had caught their attention.
Note: At ISMIR, all accepted papers were presented via a short 4 minute talk and a poster. This arrangement made it possible to keep all presentations in a single track. All papers and permited talks are posted on the ISMIR site.
Time Stamps
- [0:01:51] Matan’s rec
- [0:07:27] Rachel’s rec
- [0:10:51] Andrew’s rec
- [0:15:20] Ashley and Felicia’s rec
- [0:19:59] Néstor’s rec
- [0:26:55] Tejaswinee’s rec
- [0:31:13] Brian’s rec
- [0:36:06] Finn’s recs
Show notes
- Matan Gover recommends [A13] Conditioned-U-Net: Introducing a Control Mechanism in the U-Net for Multiple Source Separations by Gabriel Meseguer Brocal and Geoffroy Peeters (paper, presentation)
- Andrew Demetriou recommends [F10] Tunes Together: Perception and Experience of Collaborative Playlists by So Yeon Park; Audrey Laplante; Jin Ha Lee; Blair Kaneshiro (paper, presentation)
- Tejaswinee Kelkar recommends [B03] Estimating Unobserved Audio Features for Target-Based Orchestration by Jon Gillick; Carmine-Emanuele Cella; David Bamman (paper, presentation)
- Ashley Burgoyne and Felicia Villalobos recommend [E13] SAMBASET: A Dataset of Historical Samba de Enredo Recordings for Computational Music Analysis by Lucas Maia; Magdalena Fuentes; Luiz Biscainho; Martín Rocamora; Slim Essid (paper, presentation)
- Néstor Nápoles López recommends the anniversary paper [E-00] 20 Years of Automatic Chord Recognition from Audio by Johan Pauwels; Ken O’Hanlon; Emilia Gomez; Mark B. Sandler (paper, presentation)
- Rachel Bittner recommends [A06] Cover Detection with Dominant Melody Embeddings by Guillaume Doras; Geoffroy Peeters (paper, presentation)
- Brian McFee recommends [E-06] FMP Notebooks: Educational Material for Teaching and Learning Fundamentals of Music Processing by Meinard Müller; Frank Zalkow (paper, presentation, webpage)
- And Finn’s rec:
- [D-12] AIST Dance Video Database: Multi-Genre, Multi-Dancer, and Multi-Camera Database for Dance Information Processing By Shuhei Tsuchida; Satoru Fukayama; Masahiro Hamasaki; Masataka Goto. (Paper, presentation)
- Keynotes: Henkjan Honing’s What makes us musical animals and Georgina Born’s MIR redux: Knowledge and realworld challenges, and new interdisciplinary futures
- [F-14] The ISMIR Explorer – A Visual Interface for Exploring 20 Years of ISMIR Publications by Thomas Low; Christian Hentschel; Sayantan Polley; Anustup Das; Harald Sack; Andreas Nurnberger; Sebastian Stober (paper, presentation, website)
Credits
The So Strangely Podcast is produced by Finn Upham, 2019. Algorithmic music samples from the blog post Music Transformer: Generating Music with Long-Term Structure, and included under the principles of fair dealing. The closing music includes a sample of Diana Deutsch’s Speech-Song Illusion sound demo 1.
-
Music Transformer and Machine Learning for Composition with Dr. Anna Huang
Podcast: Play in new window | Download (Duration: 54:21 — 50.4MB)
Subscribe: Spotify | RSS | More
Finn interviews Composer and Machine Learning specialist Dr. Cheng-Zhi Anna Huang about the Music Transformer project at Google’s Magenta Labs. They discuss representations of music for machine learning, algorithmic music generation as a compositional aid, the JS Bach Google Doodle, how self-reference defines structure in music, and compare the musicality of different systems with example outputs.
Time Stamps
- [0:01:05] Introducing Dr. Anna Huang
- [0:03:43] JS Bach Google Doodle
- [0:12:52] Representations of musical information for machine learning
- [0:16:26] Music Transformer project
- [0:25:15] RNN algorithm music sample
- [0:25:45] ABA structure challenge for generative systems
- [0:30:30] Vanilla Transformer algorithm music sample
- [0:32:07] Music Transformer algorithm music sample
- [0:36:30] Self Reference Visualisation (see blog post)
- [0:43:27] Everyday music implications
- [0:48:10] What this work says about music
- [0:50:01] Music Transformer trained on Jazz Piano
Show notes
- Recommended project:
- Blog post: Huang, C.Z.A., Simon, I., & Dinculescu, M. (2018, Dec 12). Music Transformer: Generating Music with Long-Term Structure [Blog Post]
- Paper: Huang, C.Z.A., Vaswani, A., Uszkoreit, J., Shazeer, N., Simon, I., Hawthorne, C., Dai, A.M., Hoffman, M.D., Dinculescu, M., & Eck, D. (2018) MUSIC TRANSFORMER: GENERATING MUSIC WITH LONG-TERM STRUCTURE on arXiv.org
- Interviewee: Dr. Cheng-Zhi Huang at Google AI, on twitter @huangcza
- Google Doodle Celebrating JS Bach with AI harmonising melodies
- Related papers:
- Huang, C.Z.A., Cooijmans, T., Roberts, A., Courville, A., Eck, D. (2017). Coconet: Counterpoint by Convolution. ISMIR.
- Huang, C.Z.A., Cooijmans, T., Dinculescu, M., Roberts, A., & Hawthorne, C. (2019, Mar 20). Coconet: the ML model behind today’s Bach Doodle.
- Huang, C.Z.A., Hawthorne, C., Roberts, A., Dinculescu, M., Wexler, J., Hong, L., Howcroft, J. (2019). The Bach Doodle: Approachable music composition with machine learning at scale. ISMIR.
Credits
The So Strangely Podcast is produced by Finn Upham, 2018. Algorithmic music samples from the blog post Music Transformer: Generating Music with Long-Term Structure, and included under the principles of fair dealing. The closing music includes a sample of Diana Deutsch’s Speech-Song Illusion sound demo 1.
-
Episode 6: Relevance of vocals to music listener preferences, with Brian McFee and guest Andrew Demetriou
Podcast: Play in new window | Download (Duration: 1:14:48 — 70.9MB)
Subscribe: Spotify | RSS | More
Music tech and data science professor Brian McFee recommends Vocals in Music Matter: The Relevance of Vocals in the Minds of Listeners by Andrew Demetriou, Andreas Jansson, Aparna Kumar, and Rachel M. Bittner, published in the 2018 ISMIR proceedings. Brian and Finn interview Andrew Demetriou about this research combining descriptions of music on Spotify and survey responses on what users pay attention to, like, and dislike in music generally and vocals specifically.
Time Stamps
- [0:00:00] Introduction with Brian
- [0:10:05] Interview: Introduction: Origins of paper and Survey 1 analysis
- [0:20:15] Interview: Results of survey 1 and ethical research practices at Spotify
- [0:27:03] Interview: Second Survey construction, analysis, and results
- [0:34:37] Interview: Problems of terminology and labeling
- [0:43:27] Interview: Overall results and absence of vocals terms in music descriptions
- [0:53:30] Interview: Implications for everyday music listening
- [0:58:40] Closing with Brian (12/10 for efficient summary)
Show notes
- Recommended article:
- Demetriou, A., Jansson, A., Kumar, A., & Bittner, R. M. Vocals in Music Matter: The Relevance of Vocals in the Minds of Listeners. Proceedings of ISMIR 2018 (pp. 514-520).
- Slide deck from the corresponding ISMIR talk that caught Brian’s attention
- Interviewee: Andrew Demetriou
- Co-host: Prof. Brian McFee
And here is the action shot of the research team card sorting participants’ text responses to Survey 1.
Spotify researchers (left to right) Rachel Bittner, Andreas Jansson, Andrew Demetriou, and Aparna Kumar working through the text responses to Survey 1. Credits
The So Strangely Podcast is produced by Finn Upham, 2018. The closing music includes a sample of Diana Deutsch’s Speech-Song Illusion sound demo 1.
-
Episode 2: Aligned Hierarchies and Segmentation with Vincent Lostanlen and guest Katherine Kinnaird
Podcast: Play in new window | Download (Duration: 1:13:48 — 70.0MB)
Subscribe: Spotify | RSS | More
Data Scientist Vincent Lostanlen recommends Katherine Kinnaird’s “Aligned Hierarchies: A Multi-Scale Structure-Based Representation for Music-Based Data Streams”, published in the proceedings of ISMIR (2016). Vincent and Finn interview Dr. Kinnaird about this method for abstracting structure in music through repetition, how it has been implemented for fingerprinting on Chopin’s Mazurkas, and how Aligned Hierarchies could be used for other tasks and on other musics.
Show notes
- Recommended article:
- Kinnaird, K. M. (2016). Aligned Hierarchies: A Multi-Scale Structure-Based Representation for Music-Based Data Streams. In ISMIR (pp. 337-343). http://m.mr-pc.org/ismir16/website/articles/020_Paper.pdf
- Interviewee: Dr. Katie Kinnaird, Data Sciences Postdoctoral Fellow, Affiliated to the Division of Applied Mathematics at Brown University twitter @kmkinnaird
- Co-host: Dr. Vincent Lostanlen, Postdoctoral Researcher at the Cornell Lab of Ornithology, Visiting scholar at MARL at NYU, twitter: @lostanlen
- Papers cited in the discussion:
- M. Casey, C. Rhodes, and M. Slaney. Analysis of minimum distances in high-dimensional musical spaces. IEEE Transactions on Audio, Speech, and Language Processing, 16(5):1015 – 1028, 2008.
- J. Foote. Visualizing music and audio using self- similarity. Proc. ACM Multimedia 99, pages 77–80, 1999.
- M. Goto. A chorus-section detection method for musical audio signals and its application to a music listening station. IEEE Transactions on Audio, Speech, and Language Processing, 14(5):1783–1794, 2006.
- P. Grosche, J. Serrà, M. Müller, and J.Ll. Arcos. Structure-based audio fingerprinting for music retrieval. 13th International Society for Music Information Retrieval Conference, 2012.
Time Stamps
- [0:00:10] Intro with Vincent Lostanlen
- [0:17:22] Interview: Origins of the Aligned Hierarchies
- [0:30:22] Interview: Implementation & Fingerprinting on the Mazurkas
- [0:52:55] Interview: New applications and developments for Aligned Hierarchies
- [1:02:57] Closing with Vincent Lostanlen
Credits
The So Strangely Podcast is produced by Finn Upham, 2018.
The closing music includes a sample of Diana Deutsch’s Speech-Song Illusion Sound Demo 1.
- Recommended article:
