SynIR: Synthetic Impulse Response Dataset using Geometric Sound Propagation

Zhenyu Tang, Dinesh Manocha

University of North Carolina at Chapel Hill, University of Maryland at College Park


SynIR dataset is a large set of synthetic impulse responses (IR), generated using geometric sound propagation in various room environment. We use 3D house models from the SUNCG dataset. As in the following image, several collision-free far-field sound source-listener pairs (shown as floating colored spheres) are sampled in each one-story house, and each pair's corresponding IR up to 8kHz is computed using geometric propagation.

This dataset is under heavy development, so please stay tuned.

Agreement of SynIR dataset

Before we are able to offer you access to SynIR dataset (the Dataset), please read the following aggrement and indicate your response.

  1. University of North Carolina at Chapel Hill, University of Maryland at College Park, and all developers of the Dataset make no representations or warranties regarding the Database, including but not limited to warranties of non-infringement or fitness for a particular purpose.
  2. Researcher may provide research associates and colleagues with access to the Database provided that they first agree to be bound by these terms and conditions.
  3. University of North Carolina at Chapel Hill and University of Maryland at College Park reserve the right to terminate Researcher's access to the Database at any time.

This dataset has several versions synthesized for different microphone specifications. Detailed channel layouts and the naming scheme of wavfiles will be given soon, or upon request. Each dataset contains about 10K multi-channel wavfiles. The corresponding statistic files contain the RT60 and Speech Transmission Index (STI) of each file averaged over channels.

Download SynIR 6-channel version (8GB) md5: 37aabca7ee22b8ebe4b4fd3bb55cfdf2
statistics csv (6-channel)
Download SynIR 8-channel version (10GB) md5: 7a713b950c011c6cac78cd8e0280a92e
statistics csv (8-channel)
Last updated: July 4, 2018