SimSceneTVB Perception

This is a corpus of 100 sound scenes of 45s each representing urban sound environments, including:

• 6 scenes recorded in Paris,
• 19 scenes simulated using simScene (https://bitbucket.org/mlagrange/simscene) to replicate recorded scenarios, including the 6 recordings in this corpus,
• 75 scenes simulated using simScene with diverse new scenarios, containing traffic, human voices and bird sources.

The base audio files used for simulation are obtained from Freesound (https://freesound.org) and LibriSpeech (http://www.openslr.org/12).

This corpus has been evaluated by a panel of participants in a listening experiment, with assessments on the following 0-10 Likert scales:

1. Pleasantness: Unpleasant - Pleasant,
2. Liveliness: Inert, amorphous - Lively, eventful,
3. Overall loudness: Quiet - Noisy,
4. Interest: Boring, uninteresting - Stimulating, interesting,
5. Calmness: Agitated, chaotic - Calm, peaceful,
6. Sound level of passing vehicles: Very low - Very high,
7. Time of presence of traffic: Never - Continuously,
8. Time of presence of voices: Never - Continuously,
9. Time of presence of birds: Never - Continuously.

Assessments from 23 subjects are available for the 6 recorded and 19 simulated scenes, and from 7 to 8 subjects for the 75 simulated scenes.

The contents of this dataset are as follow:

• assessments: contains evaluations by 23 subjects of perceptual scales on the corpus
  • sXX: Folder corresponding to participant XX
    • Pt_.txt: Contains assessments. Each line corresponds to one scene, columns correspond to (resp.) scene number (see audio_list.txt for correspondance), pleasantness, liveliness, overall loudness, interest, calmness, sound level of passing vehicles, time of presence of traffic, human voices, bird sources.
• audio
  • rec: contains the 6 recorded 45s scenes
  • rep: contains the 19 replicated 45s scenes, with separated tracks for source contributions
  • sim: contains the 75 simulated 45s scenes, with separated tracks for source contributions
• audio_list.txt: list of audio files in the corpus, line ordering corresponds to numbers in the first column of assessments
• rep_exp.mat, sim_exp.mat: Additional information about the corpus, with playback Leq and physical estimations of the perceptual time of presence of sources for each scene.