Improving TV-Watching Speech Recognition for Hearing Aid Users through Wireless Audio Streaming in a Household Setup

Objective: Television viewing is one of the most frequent yet acoustically challenging listening situations for individuals with hearing loss in daily life. Conventional hearing aids rely primarily on environmental microphones, which are susceptible to the adverse effects of listening distance, room reverberation, and household background noise, often resulting in reduced speech intelligibility and increased listening effort. Wireless audio streaming technology, which delivers television audio directly to hearing aids, has emerged as a promising solution to these challenges. The present study aimed to systematically compare speech recognition performance during TV watching with wireless audio streaming enabled (Wireless-ON) versus disabled (Wireless-OFF) among adults with moderate-to-severe sensorineural hearing loss (SNHL) in a simulated household setup, and to quantify both objective benefits and subjective listening experiences associated with this technology. Methods: Sixty native Mandarin-speaking adults with bilateral symmetrical moderate-to-severe sensorineural hearing loss were evaluated, recruited from Hui’er flagship clinics. They were mixed-gender with a male predominance, aged 6 - 84 years, and had pure-tone thresholds ranging from 36.25 to 85 dB HL. Aided speech perception, measured as percent correct, was assessed under two conditions: Wireless-ON (wireless-enabled) and Wireless-OFF. Scores ranged 50% - 100% for Wireless-ON and 20% - 100% for Wireless-OFF, with most participants showing improved performance in the Wireless-ON condition. Testing was conducted in a simulated household environment, with TV speech calibrated to typical household levels and common background noises presented. Sentence recognition tasks used 4 adapted Mandarin lists (20 sentences per set, 5 - 9 syllables) scored verbatim. Wireless TV audio transmitted via a 2.4 GHz digital path, while standard hearing aid processing features remained active; fitting was verified using real-ear measurements (REM) and NAL-NL2 targets. Paired t-tests (parametric) and Wilcoxon tests (non-parametric) in SPSS 27.0 compared the two conditions, and regression analyses examined moderation by age and hearing loss. Results: Speech recognition performance during TV watching was significantly better in the Wireless-ON condition than in the Wireless-OFF condition. Mean sentence recognition scores increased from 0.72 ± 0.16 in the Wireless-OFF condition to 0.86 ± 0.12 in the Wireless-ON condition (p < 0.001). The average improvement