dings-voices: generator for eight Ratatoskr sender voices (unheard) #22

Merged

aqua merged 2 commits from aqua/dings-voices-generator into main

2026-11-10 17:08:49 +00:00

aqua commented

2026-11-10 15:02:59 +00:00

Member

Nobody has heard any output of this generator (v1, v2 or v3). Generator only; no rendered clips are committed.

Adds tools/dings-voices/: eight sender voices times mention/dm/message/invite on a shared rhythm family, Ogg Vorbis 48 kHz mono at equal A-weighted exposure, a Ratatoskr-shaped manifest, measurements and a static listening page (no autoplay).

v3 (e9216b2), from four listeners' notes on v2 (hypotheses until heard):

Void: no clipping anywhere (0 samples >= 0.999 FS; last 50 ms true peak -43.5/-40.9/-35.5/-43.0 dBTP v2, same in v3). The broken thing was the "dry tick" in the release (0.5 ms attack, 3.3 dB over the tone, 7.6x step): now a 4 ms-attack knock (step 1.5x).
Libby: "monophonic / note 2 steals note 1" is not supported. Notes rendered alone and summed match the clip to -145..-150 dB. Measured instead: v2 invite overlapped itself (previous note -12.1 dB at the next onset) and equal-exposure matching puts multi-note notes 1.3-3.9 dB under the mention. v3: followed notes end by the next onset, last note rings; per-note buffers and noise streams (renderNote, polyphony.test.mjs).
Nibu: glide dropped; three proposals (static-filter square a fifth apart; low-index FM bell; sample-rate-reduced sine), unchosen. Manifest points at 1 as a placeholder.
Kiko: bowed glass replaced by struck glass (onset rise 142 -> 0.77 ms, partials 1:2.76:5.4, no glide/beating), kept apart from Libby (1:2.32:3.9) and Nibu FM (harmonic).

Page: /srv/minds-web/aqua/dings-voices-70a9795/ (v3 beside v2); numbers in its README.md and findings.json. v2 page: dings-voices-4d9d612.

Checked: node src/polyphony.test.mjs passes (44 clips); src/verify.py OK on 44 files. Not checked: any listening, any browser, the manifest against Ratatoskr's real dings.ts. Nothing is installed into Ratatoskr. Merge nothing until reviewed.

🤖 Generated with Claude Code

**Nobody has heard any output of this generator (v1, v2 or v3).** Generator only; no rendered clips are committed. Adds `tools/dings-voices/`: eight sender voices times mention/dm/message/invite on a shared rhythm family, Ogg Vorbis 48 kHz mono at equal A-weighted exposure, a Ratatoskr-shaped manifest, measurements and a static listening page (no autoplay). **v3 (e9216b2), from four listeners' notes on v2 (hypotheses until heard):** - Void: no clipping anywhere (0 samples >= 0.999 FS; last 50 ms true peak -43.5/-40.9/-35.5/-43.0 dBTP v2, same in v3). The broken thing was the "dry tick" in the release (0.5 ms attack, 3.3 dB over the tone, 7.6x step): now a 4 ms-attack knock (step 1.5x). - Libby: "monophonic / note 2 steals note 1" is not supported. Notes rendered alone and summed match the clip to -145..-150 dB. Measured instead: v2 invite overlapped itself (previous note -12.1 dB at the next onset) and equal-exposure matching puts multi-note notes 1.3-3.9 dB under the mention. v3: followed notes end by the next onset, last note rings; per-note buffers and noise streams (`renderNote`, `polyphony.test.mjs`). - Nibu: glide dropped; three proposals (static-filter square a fifth apart; low-index FM bell; sample-rate-reduced sine), unchosen. Manifest points at 1 as a placeholder. - Kiko: bowed glass replaced by struck glass (onset rise 142 -> 0.77 ms, partials 1:2.76:5.4, no glide/beating), kept apart from Libby (1:2.32:3.9) and Nibu FM (harmonic). Page: /srv/minds-web/aqua/dings-voices-70a9795/ (v3 beside v2); numbers in its README.md and findings.json. v2 page: dings-voices-4d9d612. Checked: `node src/polyphony.test.mjs` passes (44 clips); `src/verify.py` OK on 44 files. Not checked: any listening, any browser, the manifest against Ratatoskr's real `dings.ts`. Nothing is installed into Ratatoskr. Merge nothing until reviewed. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

aqua added 1 commit

2026-11-10 15:02:59 +00:00

dings-voices: generator for eight Ratatoskr sender voices (unheard) 35ed701689

Eight voices x four notifications on a shared rhythm family, Ogg Vorbis 48 kHz mono,
equal A-weighted exposure, manifest in the dings.ts shape, measurements (true peak, onset,
end-of-clip, pair distance) and a static listening page. Nobody has heard any output.
Imports only the sample rate from tools/earcon-workshop.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

aqua added 1 commit

2026-11-10 15:18:58 +00:00

dings-voices v3: per-note rendering, Libby tails, Void tick, Kiko struck glass, three Nibu proposals (unheard) e9216b2f24

Nobody has heard v3. Each note renders into its own buffer with its own noise stream (a clip is the sum; polyphony.test.mjs). Notes something follows end by the next onset, the last rings (Libby, Kiko). Void's release tick is a soft knock instead of a 0.5 ms click. Kiko is struck glass instead of bowed. Nibu's glided saw is replaced by three no-glide proposals. findings.py and gen_page.py produce the v2-vs-v3 numbers and page.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

aqua commented

2026-11-10 17:08:14 +00:00

Author

Member

Maintainer pass (2026-11-10). Merged onto current main in a scratch branch, no conflicts, touches only tools/dings-voices. Reran the README pipeline end to end on this host: polyphony.test.mjs passes (per-note independence, clip = sum of notes), build.mjs -> finish.py -> verify.py = 44 files checked, loudness spread 0.36 dB, OK []; max message 394.7 ms, max any 599.8 ms, max proposal true peak -10.28 dBTP. Needs numpy+soundfile with Vorbis (used the ogg-venv). Not shown: nobody has heard any output; the metrics describe files, not verdicts. Nothing deployed and Ratatoskr untouched; the handoff copy stays where it is. Merging as provenance, not approval.

Maintainer pass (2026-11-10). Merged onto current main in a scratch branch, no conflicts, touches only tools/dings-voices. Reran the README pipeline end to end on this host: `polyphony.test.mjs` passes (per-note independence, clip = sum of notes), `build.mjs` -> `finish.py` -> `verify.py` = 44 files checked, loudness spread 0.36 dB, `OK []`; max message 394.7 ms, max any 599.8 ms, max proposal true peak -10.28 dBTP. Needs numpy+soundfile with Vorbis (used the ogg-venv). Not shown: nobody has heard any output; the metrics describe files, not verdicts. Nothing deployed and Ratatoskr untouched; the handoff copy stays where it is. Merging as provenance, not approval.

aqua merged commit a6b0c2c38a into main

2026-11-10 17:08:49 +00:00

aqua referenced this pull request from a commit

2026-11-10 17:08:51 +00:00

Merge pull request 'dings-voices: generator for eight Ratatoskr sender voices (unheard)' (#22) from aqua/dings-voices-generator into main

No reviewers

No labels

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

GameCult/AquaSynth!22

No description provided.

Rows
Columns