CLASSIFICATION
Swallow ID:
10040
Partner Institution:
Concordia University
Source Collection Title:
SpokenWeb AV
Source Collection ID:
ArchiveOfThePresent
Source Collection Description:
SpokenWeb Audio Visual Collection
Source Collection Contributing Unit:
SpokenWeb
Source Collection URI:
Source Collection Image URL:
https://archiveofthepresent.spokenweb.ca/_nuxt/img/header-img_1000.fd7675f.png
Series Title:
SpokenWeb Events
Sub Series Title:
SpokenWeb AV
ITEM DESCRIPTION
Title:
SpokenWeb Events AV, Day 3 Digital Audio Tools: Sound Searching and Visualization Keynotes, The Literary Audio Symposium, 4 December 2016
Title Source:
https://spokenweb.ca/symposia/#/literary-audio-symposium
Language:
English
Production Context:
Classroom recording
Identifiers:
[]
Rights
CREATORS
Name:
Jason Camlot
Dates:
1967-
Name:
Steve McLaughlin
Name:
Tanya Clement
CONTRIBUTORS
MATERIAL DESCRIPTION
DIGITAL FILE DESCRIPTION
Sample Rate:
44.1 kHz
Duration:
01:57:27
Size:
281,871,672 bytes
Notes:
MP3 audio
Title:
LS102519
Content Type:
Sound Recording
Dates
Date:
2016-12-04
Type:
Production Date
LOCATION
Address:
1400 Boulevard de Maisonneuve O, Montreal, QC, H3G 1M8
Venue:
Concordia University McConnell Building
Latitude:
45.4968036
Longitude:
-73.57792785757887
CONTENT
Contents:
TANYA CLEMENT (U Texas at Austin)
“High Performance Sound Technologies for Access and Scholarship”
Co-Presented with Steve McLaughlin. Humanists have few opportunities to use advanced technologies for analyzing large, messy sound archives. In response to this lack, the HiPSTAS (High Performance Sound Technologies for Access and Scholarship) Project is developing a research environment that uses machine learning and visualization to automate processes for describing unprocessed spoken-word collections of keen interest to humanists. This paper describes how we have developed, as a result of HiPSTAS, a machine learning system called ARLO (Adaptive Recognition with Layered Optimization). I describe a use case for finding moments of applause in the PennSound collection, which includes approximately 36,000 files comprising 6,200 hours of poetry performances and related materials. We conclude with a brief discussion about our preliminary results and some observations on the efficacy of using machine learning to facilitate generating data about unprocessed spoken-word sound collections in the humanities.
Clement’s research centers on infrastructure information impacting academic research, research libraries, and the creation of research tools/resources in the digital humanities. Projects include High Performance Sound Technologies for Access and Scholarship, “Improving Access to Time-Based Media through Crowdsourcing and Machine Learning” project. Important articles include: “Measured Applause: Toward a Cultural Analysis of Audio Collections.” Cultural Analytics 1; “A Rationale of Audio Text.” Digital Humanities Quarterly 10; “The Ear and the Shunting Yard: Meaning Making as Resonance in Early Information Theory.” Information & Culture 49; and, “Distant Listening: On Data Visualisations and Noise in the Digital Humanities.” Text Tools for the Arts. Digital Studies 3.
STEVE McLAUGHLIN (U Texas at Austin)
“High Performance Sound Technologies for Access and Scholarship”
Co-Presented with Tanya Clement. Humanists have few opportunities to use advanced technologies for analyzing large, messy sound archives. In response to this lack, the HiPSTAS (High Performance Sound Technologies for Access and Scholarship) Project is developing a research environment that uses machine learning and visualization to automate processes for describing unprocessed spoken-word collections of keen interest to humanists. This paper describes how we have developed, as a result of HiPSTAS, a machine learning system called ARLO (Adaptive Recognition with Layered Optimization). I describe a use case for finding moments of applause in the PennSound collection, which includes approximately 36,000 files comprising 6,200 hours of poetry performances and related materials. We conclude with a brief discussion about our preliminary results and some observations on the efficacy of using machine learning to facilitate generating data about unprocessed spoken-word sound collections in the humanities.
The Literary Audio Symposium
Digitized spoken-audio archives have proliferated over the past two decades, making a wide range of historically significant analog spoken recordings originally captured in different media formats accessible to listeners and scholars for the first time. Online repositories like PennSound and the Cylinder Archive Project, have begun to transform previously multi-format collections into a massive resource, the potential of which is just beginning to be realized. Still, many local audio archives with recordings that document literary events remain either inaccessible or, if digitized, largely disconnected from each other. Given the potential usefulness of online audio archives for scholars, teachers and the general public, The Literary Audio Symposium aims to explore possibilities around a coordinated and collaborative approach to literary historical study, digital development and critical and pedagogical engagement with diverse collections of spoken recordings.
The Symposium emerges from a joint venture of the AMP Lab and TAG Centre, COHDS: Centre for Oral History and Digital Storytelling, and the Concordia University Libraries, all based at Concordia, in collaboration with literary scholars, digital humanists and librarian/archivists from the University of Alberta, University of Calgary, and Simon Fraser University, and local community partners with unique analogue holdings. Invited participants include colleagues from McGill U, U Victoria, U Texas, Austin, UCSB, and The Canadian Centre for Architecture.
This symposium will offer a productive scene of discussion and collaboration between academic researchers, librarians and archivists and emerging scholars and students, as well as community-based cultural and literary practitioners. The primary aims of The Literary Audio Symposium are to share knowledge and provide discussion and debate about
1) new forms of historical and critical scholarly engagement with coherent collections of spoken recordings;
2) digital preservation, aggregation techniques, asset management and infrastructure to support sustainable access to diverse collections of archival spoken audio recordings;
3) techniques and tools for searching and visualizing corpora of spoken audio (for features relevant to humanities research and pedagogy); and
4) innovative ways of mobilizing digitized spoken and literary recordings within pedagogical and public contexts.
These objectives will be met through a structured set of keynote topic-organizing panels, tool demonstrations, case-study presentations, and collaborative workshop discussions, led by experts from a variety of relevant backgrounds including Literature, Library, Archives and Information Science, Oral History and Digital Storytelling, Computer Science, and Communications and Media History. Each day of the Symposium will be initiated by a plenary presentation that frames key questions concerning one of the four key symposium themes, to be followed by hands-on presentations of relevant digital platforms and tools, case-study presentations that elaborate upon the day’s theme, followed by collaborative workshop discussion that will debate, reflect upon, and formulate new approaches to engaging with the implications of the day’s materials.
From a range of relevant perspectives, The Literary Audio Symposium will enable the collaborative formulation of answers to core questions surrounding the preservation, digital presentation and critical use of humanities-oriented spoken audio materials, and temporal media holdings of cultural significance, in general. Our work will benefit scholars, students and society by establishing processes for making a generally dispersed corpus of cultural heritage widely available in useful and meaningful ways.
NOTES
RELATED WORKS