Sabre is an open-source simulation environment for ABR algorithms. It is a Python tool that facilitates initial development and quick evaluation of algorithms in an environment similar to real production players, without requiring the algorithm researchers to learn the often complex implementation details of a production player. Sabre takes a video description, network trace, and an ABR algorithm as inputs and gives a collection of QoE metrics as output.

For documentation about Sabre, read our wiki.

You can checkout the project repository using the command

git clone https://github.com/UMass-LIDS/sabre.git

Otherwise, you can download a zipfile here.

If you want to cite Sabre, please refer to the following paper:

Kevin Spiteri, Ramesh Sitaraman, and Daniel Sparacio. 2018. From Theory to Practice: Improving Bitrate Adaptation in the DASH Reference Player. In MMSys '18: 9th ACM Multimedia Systems Conference, June 12-15, 2018, Amsterdam, Netherlands. https://doi.org/10.1145/3204949.3204953