Sascha Grollmisch, Jakob Abeßer, Judith Liebetrau, Hanna Lukashevich

Sounding Industry: Challenges and Datasets for Industrial Sound Analysis, Proceedings of the 27th European Signal Processing Conference (EUSIPCO), A Coruña, Spain, 2019.

Dataset Overview

The IDMT-ISA-ELECTRIC-ENGINE dataset contains sound files of three similar units of an electrical engine (2ACT Motor Brushless DC 42BLF01, 4000 RPM, 24VDC), which simulate different acoustic conditions. The operational states “good”, “heavy load” and “broken” were provoked by a change of supply voltage and loading weight leading to a change of the operating sound. In March 2017 the IDMT_ISA_ELECTRIC_ENGINE dataset was recorded at Fraunhofer Institute for Digital Media Technology (IDMT). In each file, only one of the engines is active at the same time assuming an engine can only have one of the three operational states. The dataset consists of recordings of the electric engine plus the following background noise types:

  • File duration:  42.32 minutes
  • # Operational State “good”: 774
  • # Operational State “broken”: 789
  • # Operational State “heavyload”: 815
  • # Total WAV Files: 2378
  • Sampling rate: 44.1KHz
  • Resolution: 32-bit
  • Mono audio

More audio datasets

Fraunhofer IDMT has compiled further audio data sets for various research areas such as instrument recognition, fingerings or game analysis.

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content

Research topic

Environmental Sound Analysis

AI-based analysis of complex acoustic scenes and sounds

Research topic

Media Forensics

Trustworthy media content

Research topic

Industrial Sound Analysis

AI-based acoustic monitoring in production