The Noise Colour Index, 2026 Edition: Play Shares by Noise Colour from 96.5 Million Platform-Reported Streams, 2018–2026

Aggregate play counts from a large single-artist noise catalogue (500+ tracks) on Apple Music and Spotify: 96,513,948 platform-reported plays, January 2018 to July 2026. Headline findings, stated with their limits: pink noise accounts for 68.1% of all plays; white noise, the household name, for 8.3% (roughly 8:1). Queries reaching the publisher's own pages point the other way — white-noise searches outnumber pink at least 3:1 on Google and about 40:1 on Bing — a search-side proxy with a different base, since plays are dominated by high-frequency repeat listeners. The play distribution is heavily concentrated: the most-played track carries about 64% of pink plays and 44% of the catalogue total. Nine of the ten most-played tracks are low-pass-filtered (darkened) variants; whether laboratory findings on broadband noise generalise to such filtered stimuli is an open question this dataset cannot answer. The catalogue contains more white-noise tracks than pink, so the pink preference cannot be explained by track availability alone, though release age, promotion and platform placement are not controlled. Methodology: platform-reported play counts from Apple Music for Artists and Spotify for Artists track-level exports. Figures cover the publisher's own recordings only: five recordings that appear under the artist's Spotify profile but are not the publisher's recordings (a third-party release misattributed to the profile by the platform) are excluded. A play is a platform-defined event and definitions differ between services; play counts are not duration-weighted, so play share is not listening-time share. Colour classification by track title. Counting is complete within these two platforms; other services are not included. Figures describe plays within one commercial catalogue and should be treated as directional for the wider category, not as a survey or evidence of efficacy. The deposit contains the aggregate tables (CSV), a full codebook with variable definitions and limitations, a machine-readable citation pack (JSON), citation metadata (CITATION.cff) and a CC BY 4.0 licence. Per-track, city-level and longitudinal detail are not public; academic research access is available by arrangement via https://noise-foundation.com/contact/. Canonical statistics page: https://noise-foundation.com/noise-colour-index/