Computer Music! with Stephen Pope (and other deep neural networks)

Monday 23rd November 2015, 6.30pm
Performance lab inside G2, Engineering Building - full performance: 60 mins

Program:

Prelude to the Night (2008)

Oded Ben-Tal

04:15

We Need Us (2014)

Julie Freeman

 08:00

15 excerpts with a video collage of tools and video art (1978-2006)

Stephen Travis Pope

15:00

You Slut! - Plural Sex (MCLD Remix) (2013)

Dan Stowell

02:51

Piquetitos (2007)

Jennifer Logan

05:18

Eight short outputs generated by a long short-term memory network with three fully connected hidden layers of 512 units each trained on over 23,000 ABC transcrip- tions of session music (Irish, English, etc.), and arranged by my own “personal” neural network trained on who knows what for who knows how long (I can’t remember any of the settings) (2015)

Bob L. Sturm

09:20

Secrets, Dreams, Faith, and Wonder – A Mass for the New Millennium in Five Parts (2000-14) short-subject

Stephen Travis Pope

21:00 


PROGRAM

Prelude to the Night (Oded Ben-Tal, 2008) This is the first movement of Music for Strings (a multichannel audio composition). The piece combines recordings of a string section (tuning at the beginning of a rehearsal courtesy of the Royal College of Music student orchestra) and simulated string sounds using a bowed string physical model. As any string player will attest, learning to bow properly takes time, patience, skill, and a good ear. My aim using this model is a similarly long and delicate process: finding settings of the different parameters that produce 'bad' bowing; locating the points were the model breaks but still produces sounds; and exploring the fascinating area on the edges of the simulated instrument. I am using Common Lisp Music (many thanks to Juan Reyes) for the model developed by Stefania Serafin and Julius Smith.

We Need Us (Julie Freeman, 2014) This animation responds to real-time metadata generated by Zooniverse participants. Zooniverse engages millions of volunteers to help scientists deal with the flood of data that confronts them. The audio and visual compositions in this work are affected each time a volunteer submits results.

15 excerpts with a video collage of tools and video art (Steven Travis Pope, 1978-2006) This piece consists of 15 1-minute musical excerpts from 9 of my pieces accompanied by a visual tour of the notations and software tools used to produce them; several clips are included from the video collaborations, as well as a discography. It was assembled 10 years ago for my 50th birthday.

You Slut! - Plural Sex (MCLD Remix) (Dan Stowell, 2013)
Remix of a track by the amazing mathrock band You Slut! It uses two stems from the original audio, and everything around them is programmed from scratch in SuperCollider (no samples or instrument packs). Some of the sounds are driven by audio analysis and some by the score. "Weirdest remix ever" according to the band.

Piquetitos (Jennifer Logan, 2007) In five short movements, connected with strings of chimes, a virtual instrument constructed from a blend of various keyboards and pitched percussion instruments, all modified.

Eight short outputs … (see above) (Bob L. Sturm, 2015)
  1. Hole's Mill (1:30)
  2. Segue: The Birthday (0:09)
  3. A Fhsoilah Kilnie (1:14)
  4. Segue: The Humours Of Time Pigeon (0:33)   
  5. The Arian (3:02)
  6. Segue: Larkin's With A Coma Pile Phana (0:28)
  7. A Bump Of Howled Sho The fetch (2:38)
  8. Segue: Barch Beach (0:07)
These eight short pieces come from my recent explorations of using deep learning to assist the process of music composition. The training of the text-based network aims to make it produce the “correct” output character following a given input character from the data. The end result is a generative system from which we can sample any amount of output. The network also produces titles in its output. While the system output often exemplifies the conventions in its training data (e.g., Irish and English folk music, see The Endless Traditional Music Session), it sometimes produces surprises. These pieces come from such surprises. Acknowledgments: Andrej Karpathy for his open source char-rnn code; João Felipe Santos for help in training and sampling the network; thesession.org contributors.

Secrets, Dreams, Faith, and Wonder – A Mass for the New Millennium in Five Parts (short subject Stephen Travis Pope, 2000-14), is assembled from my feature-length abstract music/video ritual of thanksgiving, Secrets, Dreams, Faith and Wonder. It is a in five parts: (1) a lament of surrender (Jerusalem's Secrets), (2) the reading of the lesson (Leur Songe de la Paix), (3) the celebration of the ritual (Evigt Dröm), (4) the recitation of the creed (Credo), and (5) a hymn of benediction (Ora penso...). When looked at this way, it follows the structure of rituals of gratitude celebrated throughout the ages and across cultures and religions (and especially the Catholic Mass). The five pieces of music incorporate voices in Latin, English and Arabic (texts from the Bible, by M. L. King and M. K. Gandhi) as well as bird and whale songs. Each of the five parts has its own tonal and timbral language, and yet they fuse into a whole when viewed as a single large-scale work. The two inner parts (Leur Songe de la Paix and Credo) have text subtitles incorporated into the videos (texts by Martin Luther King, Jr. and Mohandas K. Gandhi, respectively), while the other parts each has a related text of some sort. Each of the videos was made to fit the music of the respective movement. The motivation for Secrets, Dreams, Faith and Wonder, for making a new mass for the new millennium, is summed up in the following paraphrased quote from the late Joseph Campbell: “Those who have heard the rhythms and hymns of the angels, who have understood any of the words of the angels, will try to recite those hymns in such a way that the angels will be attracted." More specific details of each of the five parts are as follows:

1: Jerusalem's Secrets - Kyrie or opening lament
Music in 5 movements based on samples from Ernst Krenek's Lamentatio Jeremiae Prophetae (text taken from the Bible) and Brian Eno and David Byrne’s My Life in the Bush of Ghosts; video by R. Lane Clark assisted by Tyler Beckert

2: Leur Songe de la Paix (Their Dream of Peace) - Teaching from scripture
Music in three parts for voices, bells, analog synthesizer, orchestral samples, and Morse-code machine; text and voice of Rev. Dr. Martin Luther King, Jr’s speech, A Time to Break Silence; video by R. Lane Clark

3: Evigt Dröm (Eternal Dream): A Ritual - Mystical ritual, transfiguration
Affirmative symphonic pandemonium for voices and drums in 6 chapters; video by Lance J. Putnam using his “Wrapture pattern generator” software

4: Credo - Recitation of the creed
Music in three parts for whale and human voices; text and voice of Mohandas K. Gandhi, whale songs from various sources; video by S. T. Pope using models of the Calabi-Yau manifold

5: Ora penso invece che il mondo... (Today, however, I think that the world...) - Hymn of benediction
Three quick snapshots of a really beautiful enigma for string quartet and two player pianos, performed by the Echo Quartet; video by S. T. Pope captured from the "Artificial Nature" software developed by Haru Ji and Graham Wakefield.

BIOGRAPHIES

Stephen Travis Pope (1955) grew up just outside New York City, and studied at Cornell University, Vienna Music Academy and the Mozarteum in Salzburg. He has realized his musical works in the North America (Toronto, Stanford, Berkeley, Santa Barbara, Havana) and Europe (Paris, Amsterdam, Stockholm, Salzburg, Vienna, Berlin). His music is available from Centaur Records, Perspectives of New Music, Touch Music, SBC Records and Absinthe Records. In 2007, The Electronic Music Foundation in New York released a triple-disc retrospective of his works called "Ritual and Memory"; his latest release is the award-winning feature-length visual/music film "Secrets, Dreams, Faith and Wonder: A Mass for the New Millennium in Five Parts" (with videos by Clark, Putnam and others). Stephen also has over 100 technical publications on music theory and composition, computer music, software engineering and artificial intelligence. He has lived in Santa Barbara, California for the last 20 years.

Oded Ben-Tal was born in Israel, and studied music and physics in Jerusalem. Winning a scholarship to study at Stanford University he finished his doctorate there in 2002 studying composition with Jonathan Harvey and Brian Ferneyhough, and working at Stanford's Center for Computer Research in Music and Acoustics in both research and composition activities. He participated in various international festivals and workshops such as the Dartington International Festival, Domaine Forget course with the Nouvel Ensemble Moderne, the Spark festival of electronic art, and Seoul's International Electroacoustic Music Festival. His compositions – including instrumental and vocal works, acousmatic music, interactive electronic pieces, and multimedia performances – have been played in numerous countries including: France, USA, Columbia, Israel, The Netherlands, and Denmark. Recent projects include commissions from Kate Ryder as part of her Residency at Kingston University, a piece for soprano trombone and electronics for Torbjorn Hultmark, and a commission by Musica Nova Israel. Oded is a senior lecturer at the music department, Kingston University.

Julie Freeman (www.translatingnature.org) translates data from natural sources into kinetic sculptures, physical objects, images, sound compositions and animations, exploring relationships between science, art, and the natural world; questioning the impact of technology in how we translate nature.


Dan Stowell is an EPSRC Research Fellow at C4DM applying machine learning to sound. He develops new techniques in structured "machine listening", using both machine learning and signal processing to analyse soundscapes with multiple birds. He has also worked on voice, music, birdsong and environmental soundscapes.

Jennifer Logan (www.jenniferlogan.com) is the Director of the Los Angeles Sonic Odyssey, and a composer of both electro-acoustic and instrumental works. She explores narratives and conceptual/visual analogies, composing works that are influenced by naturally occurring mathematical phenomena while infusing them with sensuality, spirituality, story-telling, and philosophy, expressing the evolving human condition and interconnectivity with the world we live in, imbued with multi-layered textural lyricism. Dr. Logan published a 70-minute sacred electronic work, Charon’s Pantheon (commissioned by artist Myron Dyal), which was premiered in February 2011 in Los Angeles, and since 2007, has been featured on Classical KUSC, KPFK, and KXLU in Los Angeles for her concerts and collaborations with an international pool of artists for the Los Angeles Sonic Odyssey. Her third and fourth albums of electro-acoustic compositions were released during the summer of 2012, while a fifth album was released online with the netlabel Stasisfield in 2014. Her research and writing on archetype, mysticism, and geometry as it pertains to Charon’s Pantheon has been published by the Athens Journal of Arts and Humanities. She received Bachelor and Master of Arts degrees in piano performance and composition from CSU, Fresno, a Ph.D. in composition from the University of California, Santa Barbara, and pursued further doctoral research in computer music aesthetics at the University of Paris VIII. Her primary mentors have been Curtis Roads, William Kraft, Alejandro Planchart, Jack Fortner, Andreas Werz, and Philip Lorenz, additional seminars and masterclasses with notable musicians such as Karlheinz Stockhausen, Steven Stucky, Philip Glass, Donald Crocket, Paul Badura-Skoda, Abbey Simon, Paul Berkowitz, Ena Bronstein, and Horacio Vaggione. Currently an adjunct assistant professor of music at Occidental College in Los Angeles, her music has been programmed in the US, Europe, and Asia, and is available through iTunes, CDBaby, Stasisfield and other resources. Notable reviews include those by Mark Swed in the Los Angeles Times as “enchanting,” Josef Woodard in the Los Angeles Times as “sonorous and sensuous,” Lisa Derrick in the Huffington Post as "controversial, uplifting, disturbing," and the Journal for the Society of Electro-acoustic Music in the US (SEAMUS) for “beauty of sound surfaces, structures, and lyricism.”

Bob L. Sturm (www.eecs.qmul.ac.uk/~sturm/) is currently a Lecturer in Digital Media at the School of Electronic Engineering and Computer Science at QMUL, teaching signal processing, and researching machine music listening and evaluation. He was persuaded to pursue a career in signal processing because of his interest in music, and electronic music in particular. He has studied electronic music and composition with Prof Curtis Roads (UCSB), Prof Jonathan Berger (Stanford), and Dr. John Drumheller (CU Boulder). His debut album, “Music From the Ocean” (2000), was favourably reviewed by Aquarius Records and swan fungus. It currently has 526 scrobbles and 9 listeners, at last.fm. From 2000-2002, he was the conductor of Fern Street Marching Band, a neighbourhood circus marching band in San Diego.