Puffer (Stanford Live Streaming Research)

Description

A free live TV streaming research study by Stanford University that uses volunteer viewers to train and evaluate new adaptive streaming algorithms in real-world conditions.

URL

• Visit resource

Related resources

• TiyuntsongSelf-play reinforcement learning + GAN-based adaptive bitrate streaming algorithm implementation (TensorFlow/tflearn), companion to arXiv:1…
• NeSt-VR: Network-aware Step-wise Adaptive Bitrate for VR Streaming over Wi-FiAcademic paper proposing a network-aware step-wise ABR algorithm for VR video streaming over Wi-Fi networks.
• FOSDEM 2024: Streaming Live Using RIST On Demand to ThousandsFOSDEM Open Media devroom talk on a real-world RIST architecture achieving 500ms glass-to-glass latency at scale, comparing RIST vs QUIC.
• Point Cloud Streaming with Latency-Driven Implicit Adaptation using MoQ2025 paper applying Media-over-QUIC (MoQ) transport to low-latency adaptive point-cloud streaming.
• Harnessing WebRTC for Large-Scale Live StreamingACM SIGCOMM 2025 paper on scaling WebRTC for large-audience low-latency live streaming.
• Performance of Low-Latency HTTP-based Streaming PlayersMMSys'21 paper comparing AVPlayer, Shaka, HLS.js, and Dash.js performance under LL-HLS and LL-DASH.
• low-latency-dash-ffmpegDockerized ffmpeg reference configuration for LL-DASH per-frame moof chunk streaming.
• KTH Thesis: Interoperable Retransmission Protocols for Low-Latency Video (RIST/SRT)Master's thesis academic evaluation comparing RIST and SRT retransmission protocols for low-latency video contribution.
• The Analysis of DASH Manifest OptimizationsAmazon-authored ACM Green Multimedia Systems workshop paper analyzing MPD manifest size optimization techniques.
• PantomimeSwift M3U8 parser supporting both master and media playlists.