What's actually in your headphones.
Every AmberRoom session is built from up to four sound layers, each chosen for a specific reason. Some have decades of research behind them; some are tradition we use transparently as tradition. Here's what plays, and why.
The instrument bed is what most listeners mean when they say "sound therapy." The complex inharmonic stack of a struck Tibetan bowl — many partials decaying at different rates — is what your nervous system reads as a held room. Crystal bowls add bright high partials; together they form a recognizable, settled timbre that has been used in contemplative practice for centuries.
Different recipes use different bowls. Sleep uses low crystal bowls (Solfeggio-tuned, slow strike rotation). Anxiety/Calm uses low Tibetan bowls. Energy uses bright crystals. Grief uses a single sustained 396 Hz crystal bowl. Each session selects bowls deterministically from a pool, so two sessions of the same intent audibly differ.
The pad is the harmonic field underneath the bowls — a slow evolving drone that gives the bowls something to resonate against. It's also the layer that solves a real problem: the brain develops tolerance to repeated audio (habituation researchers call this "the loop tune-out"). A subtly different pad each session keeps the recipe effective across multi-week use.
F-minor pentatonic for most recipes; bright major for energy; 528 Hz tonic for visualization. Generated fresh each session. The point isn't novelty — it's that you don't tune out a recipe you've used 30 times.
This is the layer with the strongest research backing — and the one every free-tier user gets. Two pure tones a few Hz apart, one in each ear, generate a phantom third tone (the binaural beat) at the difference frequency. The auditory cortex tracks that phantom; the surrounding brain regions are nudged toward the same rhythm.
Each intent targets a specific Hz band: 2.5 Hz delta for sleep, 6 Hz theta for calm and visualization, 6.5 Hz theta for meditation, 5 Hz theta for grief, 10 Hz alpha for focus, 13 Hz low-beta for energy. Underneath sits a noise floor — pink for sleep, brown for the others, notched white for tinnitus.
Spoken cues for users who want them. Breath pacing, body scan, intention prompts — the kind of guidance that helps a beginner settle into a practice. The layer is optional: focus and energy recipes run silent on purpose (cognitive work and morning lift don't benefit from a voice).
Layer 4 ships intermittently across recipes — sleep and meditation use it lightly; calm uses a short opening prompt; grief and visualization are silent by default to leave room for the listener's own inner voice.
Each intent — sleep, calm, focus, grief, energy, meditation, visualization, tinnitus, pain — maps to a specific recipe: frequency band, noise color, BPM, bowl selection, pacing curve. The mapping comes from the cited research and traditional practice; the per-session variation comes from deterministic seeding so the same session never repeats. The full bibliography lives in /learn/research.