Week 1 Into the Swing of Things - Jan. 20, 2023
Our team is very excited to start the project and collaborate with the Paul Taylor Dance Company. During our the meeting, the team met Michael, Josh and Carolyn. Michael Novak is the artistic director of the dance company. Carolyn Adams is an alumni of the company from 1965 to 1982. Currently, She runs the education program there and focuses on the legacy of the company’s work. Josh Jeffery is connecting us together being an ETC alumni on the Board of Advisors at Paul Taylor. After the quick introduction, we went straight into discussing what the dance company hopes to achieve with this project.
The Paul Taylor Dance Company is celebrating its 70th anniversary in 2024. The company wants to use technology and computer intelligence to analyze its extensive archive of over 50 videos tracking the evolution of its dance from 1962 to 2022, as it would take an extensive amount of time to do so manually. The main objective is to see how the dance evolves over time and use it to support the education of dancers and dance analysts. The core questions are whether technology can be used to analyze the films and summarize the evolution, and whether there is a correlation between what a computer sees and what humans interpret. The company is interested in identifying the themes that intersect between the two, including shapes, mood, artistic intent, movement trajectory, and more. The tool should also provide additional metadata and information to help inform the instructor’s decisions. The main challenge is to determine if changes between dances are artistic or accidental and to have the machine identify and present these changes for interpretation.
After discussing what our client had in mind with this project, we presented two design directions we prepared based on the project description. The analysis dashboard focuses on what the project initially proposed while the interactive application is a new direction we are pitching. We received positive feedback for both ideas. Michael and Carolyn expressed needing some time to think whether they would prioritize in-company usage or outreach.
For next week, we will try to run actual tools to see what kind of data is extractable from the videos of Aureole that Michael is sending over. We have some concerns because the differences in formatting, camera movement might be way too noisy. We will start by trying to compare accuracy and synchronization within or between videos.