Introducing DUST: A Dataset of Real-Time UDP Sound Packet Traces
This paper introduces DUST, a dataset composed of UDP audio packet traces generated by a real-time Networked Music Performance system. The dataset includes a variety of files collected under different network and device configurations. Each file combines traces of incoming audio packets with local playback status information. The dataset aims to support of-fline simulations of real-time audio streaming systems, aiding the development of algorithms to address challenges such as packet loss concealment, network jitter resilience, and clock drift compensation.
Authors
- Andrea Bianco (ORCID: https://orcid.org/0000-0002-5903-3558)
- Cristina Rottondi (ORCID: https://orcid.org/0000-0002-9867-1093)
- Matteo Sacchetto (ORCID: https://orcid.org/0000-0001-8795-9027)
- Leonardo Severi
Institutions
- Politecnico di Torino (IT)
Publication Details
- Published
- 2024-09-30
- DOI
- https://doi.org/10.1109/is262782.2024.10704190
- Primary Topic
- Speech Recognition and Synthesis
- Type
- article
- Field-Weighted Citation Impact
- 0.00