This dataset accompanies the publication Müller, L. Ahrens, J. & Kropp, W.: Loudspeaker Array-Based Auralization of Electric Vehicle Noise in Living Environments. Submitted to Forum Acusticum 2025: 11th Convention of the European Acoustics Association, 23rd – 26th June 2025 The .wav Files contain 32-bit floating point binaural sound pressure data and were played back via a calibrated system. This means the numerical values in the audio files correspond to the pressure to which participants were exposed. Since we are simulating realistic indoor sound levels, the files are relatively quiet and best listened to via high-quality headphones. The dataset contains 10 s long binaural indoor auralizations for 25 km/h passages at 7.5 m street distance using a two-tone AVAS, a multi-tone AVAS, a noise AVAS, and a combustion engine noise. All stimuli were generated by simulating wave field synthesis via a linear, rectangular, and circular WFS array in the sending room and for a continuous free-field case to serve as ground truth. Additionally, a slower 10 km/h passage and a mono version of the continuous free-field reference were generated to serve as low-quality anchors in the perceptual comparison.