2.2 Setup and stimuli
All tests were conducted in a double-walled soundproof laboratory. Participants sat in a chair placed 1 m in front of seven loudspeakers. Sound field hearing thresholds were obtained by warble tones for octave frequencies across 0.25–4 kHz in dB hearing level (HL). Speech perception under quiet was measured by the word recognition score (WRS [%]) of the Mandarin speech test materials (MSTMs) 12 at 65 dB speech presentation level (SPL). Speech perception in noise was measured by the speech reception threshold (SRT) of the MSTMs. The spectrum-shaped noise (SSN) was set at 65 dB SPL, and the speech signal started at 0 dB speech-to-noise ratio (SNR), with the following disyllables changing adaptively in 2 dB SPL steps as the participants responded. The SRT was defined as the speech signal level presented when a participant identified 50% of the words correctly. The SNR was calculated as the difference between the SRT and SSN. The sound localisation test (Supplementary Fig. 1a) consisted of seven audiometric loudspeakers placed at 30° intervals in a semicircle within a horizontal plane (± 90°, azimuth). Broadband noise (0.5–20 kHz) was randomly played at three different sound levels (65-, 70-, and 75-dB SPL). Each time the loudspeakers finished a presentation, participants indicated which loudspeaker they believed to be the source of the burst.