The SPID Corpus

The "SPID" (SPontaneous In-car Dialogues) corpus allows investigating - for the first time ever - the communication between speakers inside a driving car and the Lombard effect that emerges at different driving speeds. Currently, speech recordings are made in order to analyze the effects of an in-car-communication (ICC) system on speech production at different driving noise levels. ICC systems are meant to improve the communication of passengers inside a car and thus help increase driving safety.

At the heart of the SPID corpus is an acoustic ambiance simulation. That is, the speakers sit inside a stationary car and hear realistic driving noises of exactly this car. The acoustic simulation is further complemented by a visual (screen-based) projection of real driving situations that match with the driving noises. On this basis, the Lombard effect can be investigated under highly sophisticated and at the same time highly controlled laboratory conditions; and in addition, the noise can be entirely removed again from the signals after the recordings by means of adaptive cancellation and suppression approaches. Thereby, Lombardaffected speech features like intonation, stress, and formants can be analyzed in full detail, without the corresponding measurements being distorted by background noise.

The recordings themselves were based on the Map-Task paradigm, which elicits spontaneous speech with a number of selected target words included (names of streets, places, persons etc.). One speaker sat in the front passenger seat and the other one behind him/her on the backseat. The two speakers had their own microphone. Recordings were made in driving simulations at 50 km/h (city) and 130 km/h (highway), as well as in a silent reference condition.

 

Key Features

The corpus consists of approximately 25 hours of spontaneous dialogues whose individual durations vary depending on how long it took the dialogue partners to solve their Map Task. The speaker sample included 8 male and 8 female native speakers of Standard German. They lived for a long time in Northern Germany and were between 22 and 31 years old (average age: 26.5 years) at the time of the recording. Pairs of speakers were always of the same gender.

 

State of the Corpus

You can find the current state of the corpus here.

 

Download

The corpus can be used for non-commercial research purposes. Details can be found here.

 

Examples

 
Without an ICC system at 130 km/h.   With an ICC system at 130 km/h.

 

Creators of the Data Base

The data base was created as a joint work between Kiel University (CAU) and the University of Southern Denmark (SDU, Mads Clausen Institute). Involved researchers are:

  • Rabea Landgraf (CAU)
  • Christian Lüke (CAU)
  • Oliver Niebuhr (SDU)
  • Gerhard Schmidt (CAU)
  • Anne Theiss (CAU)

 

Corresponding Publications

R. Landgraf, G. Schmidt, J. Köhler-Kaeß, O. Niebuhr, and T. John: More Noise, Less Talk - The Impact of Driving Noise and ICC Systems on Acoustic-prosodic Parameters in Dialogue, Proc. DAGA, Kiel, Germany, open access, 2017

R. Landgraf, J. Köhler-Kaeß, C. Lüke, O. Niebuhr, G. Schmidt: Can You Hear Me Now? Reducing the Lombard Effect in a Driving Car Using an In-Car Communication System, Proc. Speech Prosody, pp. 479 - 483, 2016

A. Warhadpande, C. Lüke, A. Theiß, G. Schmidt: Improvement by Adding Video Feature in an Acoustic Ambiance Simulation for Automobiles, 7th Biennial Workshop on DSP for In-Vehicle Systems and Safety 2015, Berkeley, CA, USA

O. Niebuhr, B. Peters, R. Landgraf, G. Schmidt: The Kiel Corpora of "Speech & Emotion" - A Summary, Proc. DAGA 2015, March 16-19, 2015, Nürnberg, Germany

R. Landgraf, O. Niebuhr, G. Schmidt, T. John, C. Lüke, A. Theiß: Von der Straße ins Labor: Die Modifikation der Sprachproduktion bei lauten Fahrgeräuschen, DAGA 2015, Nürnberg, Germany

C. Lüke, A. Theiß, G. Schmidt, O. Niebuhr, T. John: Creation of a Lombard Speech Database using an Acoustic Ambiance Simulation with Loudspeakers, 6th Biennial DSP Workshop for In-Vehicle Systems 2013, Seoul, Korea

Recent Publications

P. Durdaut, J. Reermann, S. Zabel, Ch. Kirchhof, E. Quandt, F. Faupel, G. Schmidt, R. Knöchel, and M. Höft: Modeling and Analysis of Noise Sources for Thin-Film Magnetoelectric Sensors Based on the Delta-E Effect, IEEE Transactions on Instrumentation and Measurement, published online, 2017

P. Durdaut, S. Salzer, J. Reermann, V. Röbisch, J. McCord, D. Meyners, E. Quandt, G. Schmidt, R. Knöchel, and M. Höft: Improved Magnetic Frequency Conversion Approach for Magnetoelectric Sensors, IEEE Sensors Letters, published online, 2017

 

Website News

18.06.2017: Page about KiRAT news added (also visible in KiRAT).

31.05.2017: Some pictures added.

23.04.2017: Time line for the lecture "Adaptive Filters" added.

13.04.2017: List of PhD theses added.

Contact

Prof. Dr.-Ing. Gerhard Schmidt

E-Mail: gus@tf.uni-kiel.de

Christian-Albrechts-Universität zu Kiel
Faculty of Engineering
Institute for Electrical Engineering and Information Engineering
Digital Signal Processing and System Theory

Kaiserstr. 2
24143 Kiel, Germany

Recent News

Alexej Namenas - A New Guy in the Team

In June Alexej Namenas started in the DSS Team. He will work on real-time tracking algorithms for SONAR applications. Alexej has done both theses (Bachelor and Master) with us. The Bachelor thesis in audio processing (beamforming) and the Master thesis in the medical field (real-time electro- and magnetocardiography). In addition, he has intership erperience in SONAR processing.

We are pretty ...


Read more ...