Sub-bass dominance without extreme dynamics: signal-level signatures of East Javanese 'sound horeg' YouTube streams
Year: 2026
Type: preprint
Abstract: East Java's community-built, high-SPL 'sound horeg' systems shake buildings and face broadband noise limits, yet the YouTube programme material operators reportedly play remains uncharacterised at signal level. Analysing three 60 s segments from 50 horeg tracks (20 uploaders) and 50 comparison tracks (46 uploaders) randomly sampled from YouTube (≥ 50,000 views), this study evaluated gain-invariant descriptors (20–60 Hz sub-bass energy fraction, sub-bass stereo correlation, crest factor, and a displacement-weighted index) alongside loudness under four Holm-corrected registered hypotheses. Contrary to expectations of extreme dynamic compression, horeg streams are set apart by sub-bass elevation rather than dynamics or phase. Horeg tracks carry higher sub-bass energy than comparisons (median fraction 0.23 against 0.09, δ = 0.69) below the registered 0.5 threshold, overlapping electronic tracks. Sub-bass stereo correlation is near unity (median 0.9995) and full-band crest factor (13.1 dB) matches comparisons, far above the 6 dB AES2 test signal. Crucially, the 7.5 dB sub-bass gap dwarfs the 2.0 dB LUFS gap, exposing how loudness weighting understates low-frequency energy, while AAC re-encoding shifts crest factor without altering sub-bass metrics. Findings describe decoded streams only, without claiming radiated SPL, acoustic fields, or transducer stress.
Institutions: Lambung Mangkurat University, Universitas Negeri Surabaya
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236285769
