Becoming Latency-Native
summary
Performance claims that musicians can treat networked audio as a native platform, demonstrated by a live duet across 1,325 km with ~4.4 ms one-way delay using a custom WebRTC/Chromium app (built from Source‑Elements tools) and Opus streams. Playback relies on a shared JavaScript score plus a convolutional neural network that classifies five phrases and detects simultaneity, adding modest extra latency but keeping interaction under the perceptual 30 milliseconds threshold for musical feel. The system shows workable multi‑location harmony, mediated presence limits, and a concrete workflow for remote ensemble performance.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 1:28Latency in digital networks
- 4:17Physics and distance
- 7:25The drive for connection
- 9:01Live performance experiment
- 27:22Aesthetics of network music
- 35:47Technological solutions
- 40:13Closing and initiatives
related talks
47mHow live music is evolving in a post pandemic world
17mYour realtime AI is ngmi
19mLow Latency Live from a Different Vantage Point
19mLatency and Synchronization in Web Audio
10mRTC @Scale, Future RTC Experiences Session - Live Q&A
06mWorld's Simplest Latency Simulat
35mLive Low Latency Video Streaming
16mCDN Challenges of HTTP-based Low Latency Live Streaming Delivery