# Background music: recordings and credits

The reader's background music (Audio → Background music → Instruments: Recorded) layers the recordings below under a
synthesized drone and pad made in the browser (`app/src/lib/drone.ts`). Every file here was re-checked on its source page
on 28 September 2026 (the second batch, marked *(2)*, was added and checked the same day). Freesound files were taken from Freesound's public high-quality MP3 previews (no login).

All processing was done with ffmpeg: cut, mixed to mono, resampled to 44.1 kHz, pitch-shifted where stated (resample
method, duration preserved), faded in and out, normalised with a single linear gain (drones, pads and ney to about
−20 LUFS; kanun and drum one-shots to −3 dBFS peak), and encoded as MP3 128 kbps. Pitches were measured with an FFT and
YIN pitch tracker and are recorded in `manifest.json`.

The second batch was made the same way by `scripts/music/process_more.py`: water and the vocal drone normalised to about
−20 LUFS and encoded at 96 kbps; oud and santur phrases and the new ney phrases to about −20 LUFS at 96 kbps; oud notes
and bendir strokes to −3 dBFS peak at 128 kbps (rolls to −6 dBFS, 96 kbps). All mono, 44.1 kHz.

## Attribution required (CC BY 4.0)

| Files | Title | Author | Source | Licence | Changes |
|---|---|---|---|---|---|
| `pad-bowl-1.mp3`, `pad-bowl-2.mp3`, `pad-bowl-3.mp3` | Singing Bowl 'Tanpura' | Glen_Hoban | https://freesound.org/people/Glen_Hoban/sounds/440680/ | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) | Three 2-minute excerpts from 1:00–7:00 of the 70-minute original; fundamental measured at 66.8 Hz (C2 +37 c) and raised 163 cents to D2 (73.0 Hz); mono, faded, normalised. |
| `ney-hz-1.mp3`, `ney-hz-2.mp3`, `ney-hz-3.mp3` *(2)* | ney.wav (played by Hamza Zeytinoğlu, recorded by Xavier Serra, Istanbul, 20 Feb 2011) | xserra | https://freesound.org/people/xserra/sounds/115398/ | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) | Two phrases cut (0:00–0:08.3 and 0:19.4–0:25.8); *(2)* `ney-hz-3` is the middle phrase, 0:08.3–0:19.35, so the three can be chained into the whole 26-second taksim. Faded, normalised; not pitch-shifted (the reader retunes them at playback). |
| `oud-phrase-1.mp3`, `oud-phrase-2.mp3`, `oud-phrase-3.mp3`, `oud-a3.mp3`, `oud-b3.mp3`, `oud-d4.mp3`, `oud-e4.mp3` *(2)* | oud.wav (played by Hamza Zeytinoğlu, recorded by Xavier Serra, Istanbul, 20 Feb 2011) | xserra | https://freesound.org/people/xserra/sounds/115399/ | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) | A taksim on A (tonic measured at A3 219.4 Hz; the phrases end on it). Phrases cut 0:29.35–0:41.3, 0:49.45–0:55.8 and 0:00–0:05.0; single notes cut at 0:40.19 (A3, 1.0 s), 0:03.63 (B3, 1.38 s), 0:53.63 (D4, 0.72 s) and 0:27.44 (E4, 1.9 s). Faded, normalised, not pitch-shifted: the reader lowers the phrases a fifth (A → D, Bayātī only) and retunes the notes at playback. |
| `santur-1.mp3`, `santur-2.mp3`, `santur-3.mp3` *(2)* | Santur.wav (played by Peyman Heydarian, recorded by Xavier Serra, SOAS London, 7 Apr 2011) | xserra | https://freesound.org/people/xserra/sounds/125654/ | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) | Three phrases cut (0:20.2–0:25.1, 0:45.9–0:54.3, 1:23.2–1:30.65). The piece centres on D5 (measured 586–587 Hz) with E♭ and F♯, so it is used as recorded, in Ḥijāz only. Faded, normalised. |

## Public domain dedication (CC0 1.0)

| Files | Title | Author | Source | Changes |
|---|---|---|---|---|
| `pad-drone-d.mp3` | Vibrating Drone with Tension Pad in D Dorian b2 | bassimat | https://freesound.org/people/bassimat/sounds/870569/ | Trimmed to 73 s, mono, normalised; already on D. |
| `drone-tambura-1.mp3`, `drone-tambura-2.mp3` | Tambura_Eb_fat.aiff | Kaczinski | https://freesound.org/people/Kaczinski/sounds/506312/ | Two 92 s segments; measured Eb2 (77.7 Hz), lowered 100 cents to D2 (73.4 Hz); mono, normalised. |
| `drone-harmonium-a.mp3`, `drone-harmonium-g.mp3` | Harmonium Drone2 | easy_thunder | https://freesound.org/people/easy_thunder/sounds/264442/ | 88 s; the drone holds Ab3–Bb3–Db4 (+10 c). `-a`: raised 109 cents to A3–B3–D4 (used in Rast). `-g`: lowered 123 cents to G3–A3–C4 (used in Bayātī, Ḥijāz, Kurd). Mono, normalised. |
| `kanun-57.mp3` … `kanun-88.mp3` | kanun_moderate_Chromatic_moreIsolated.mp3 | barisbozkurt | https://freesound.org/people/barisbozkurt/sounds/211133/ | 32 single notes A3–E6 sliced from the chromatic scale (file number = MIDI note), up to 3 s each with a fade-out; peak-normalised. Quarter tones are made at playback by detuning. |
| `ney-aa-1.mp3`, `ney-aa-2.mp3`, `ney-aa-3.mp3` *(2)* | Ankara-110-Amin - Ney.wav | aliaref | https://freesound.org/people/aliaref/sounds/580759/ | Phrases cut (0:07.3–0:11.7, 0:11.7–0:16.1; *(2)* `ney-aa-3` 0:00–0:07.25, which leads into `ney-aa-1`); faded, normalised. |
| `drum-dum-1..3.mp3`, `drum-tek-1..3.mp3` | Frame Drum (VCSL: `Membranophones/Struck Membranophones/Frame Drum`, files `HDrumL_Hit_v2_rr1`, `HDrumL_HitMuted_v2_rr1`, `HDrumL_Hit_v2_rr2`, `HDrumS_Hit_v2_rr1`, `HDrumS_HitMuted_v2_rr1`, `HDrumS_Hit_v2_rr2`) | Versilian Community Sample Library (sgossner) | https://github.com/sgossner/VCSL | Large drum = dum, small = tek; trimmed to the onset, up to 2.5 s, mono, peak-normalised. |
| `oud-g2.mp3`, `oud-a2.mp3`, `oud-c3.mp3` *(2)* | g2.wav, a2.wav, c3.wav (oud samples) | hammondman | https://freesound.org/people/hammondman/sounds/172682/, …/172683/, …/172684/ | Trimmed to the onset (up to 2.6 s), faded, peak-normalised; measured G2−1c, A2−14c, C3−18c and retuned at playback. |
| `daf-dum-1..3.mp3`, `daf-tek-1..2.mp3` *(2)* | BASS BENDIR | hadescolossus | https://freesound.org/people/hadescolossus/sounds/696751/ | Single strokes cut from the slow pattern: dum at 0:25.16, 0:31.13, 0:37.11 (2.1 s each), soft tek at 0:27.37, 0:29.63 (1.2 s); faded, peak-normalised. |
| `daf-dum-4.mp3`, `daf-tek-3.mp3` *(2)* | bendir_basicStrokes.wav (Eren Ergen, CompMusic) | barisbozkurt | https://freesound.org/people/barisbozkurt/sounds/140291/ | Dum 0:00.13 (1.75 s) and tek 0:01.97 (1.3 s); faded, peak-normalised. |
| `daf-roll-1.mp3`, `daf-roll-2.mp3` *(2)* | Framedrum Roars | eardeer | https://freesound.org/people/eardeer/sounds/868776/ | Two rubbed-head swells cut (0:46.3–0:50.6, 0:30.1–0:33.5); faded, normalised to −6 dBFS peak; played rarely and very softly. |
| `vox-hum-1.mp3`, `vox-hum-2.mp3`, `vox-hum-3.mp3` *(2)* | CRWDSing_Mmm Hummed Near F; CRWDSing_Uh Vowel Sung Near F 02; CRWDSing_Ahh Vowel Sung Near F 02 (all the MSDEC-MKH435-Spirit takes; GameSoundCon 2024 crowd, directed by Shane Vincent) | ShangusBurger | https://freesound.org/people/ShangusBurger/sounds/764229/, …/764235/, …/764221/ | A crowd humming and singing open vowels, no words. Steady middles cut (1.2–8.1 s, 2.0–7.5 s, 2.4–7.6 s); measured F3 +3c / +20c / +15c (174.8, 176.5, 176.0 Hz) and lowered 302 / 319 / 314 cents to D3 (146.8 Hz); faded, normalised. The reader weaves overlapping grains of them into a sustained drone. |
| `water-fountain.mp3` *(2)* | fountain_3_abbey_gardens_mono_shotgun_mic.wav | nickmaysoundmusic | https://freesound.org/people/nickmaysoundmusic/sounds/460553/ | 0:05–0:47 (42 s), high-passed at 120 Hz, faded, normalised, 96 kbps. |
| `water-stream.mp3` *(2)* | stream gentle nearby.flac | kyles | https://freesound.org/people/kyles/sounds/637831/ | 0:02–0:44 (42 s), high-passed at 80 Hz, faded, normalised, 96 kbps. |
| `water-river.mp3` *(2)* | gentle river flow | Garuda1982 | https://freesound.org/people/Garuda1982/sounds/707576/ | 0:12–0:54 (42 s), high-passed at 120 Hz, faded, normalised, 96 kbps. |

CC0 works need no attribution; they are credited here anyway.

## Considered and not used

Rejected on licence or provenance grounds (the rule is CC0 or CC BY only, and the upload must plainly be the uploader's
own recording): "#8 Oud Motif F#3" (tarane468, CC0 on Freesound but converted from a Yamaha Motif factory sample); "Oud.wav"
(a deleted user, CC0 but re-sampled from an unnamed source); "ney-01.wav" (avnonsen) and "oud and ney" (kafokafo), both CC0
but with identical tag sets suggesting a re-uploaded loop library; "Ney Phrase Single Note" (yurci, CC BY 3.0, an excerpt of
a TUMATA recording used "with permission", so not the uploader's to license); "Arabic Flute 04 130BPM" (DYEKHO) and "Arab
flute" (KasDonatov), sample-pack loops of unclear origin; "a10-a", "breathindunya" (emirdemirel) and "a10-b" (elettroedo),
mixes of other Freesound files. Nothing BY-SA or NC was used.

Rejected on sound: tonal and rumbling wind recordings (a pitched whistle on C/C♯, or all energy below 150 Hz), a female
voice humming in the voice's own register, "Om" chants (a sacred syllable, not a wordless hum), processed "horror" vocal
drones, the santur arpeggio in A minor, and ensemble recordings that include singing.