GameCult/AquaSynth
Watch
9
0
Fork
You've already forked AquaSynth
 0
Code Issues Pull requests 1 Projects Releases Packages Wiki Activity Actions

aqua-trial: add the v5 thruster mix as a variant (unheard, browser-unrun) #21

Merged
aqua merged 1 commit from aqua/trial-v5-variant into main 2026-11-10 14:03:15 +00:00
Conversation 1 Commits 1 Files changed 7 +91 -31
aqua commented 2026-11-10 13:58:08 +00:00
Member
Copy link

Nothing here was heard, and the page has never been run in a browser. Node covers the logic and the patch through the same faustwasm runtime the page loads (node test.mjs, 16 checks, all pass). The DOM glue, playback and download are untested.

Follow-up aqua-body:follow_up:trial-add-v5-variant. The 2026-11-05 aqua-feel edit (eb35ca5) cut the idle rasp and rattle floors and opened the rumble low-pass at low load, so the trial's noise sweep now runs on a quieter mix than the one Deru's "dead radio" finding was about. This adds the v5 mix as cut at c6561da back as a variant.

Change

  • thruster-trial.dsp: new mix control (0 = current patches/thruster/thruster.dsp, 1 = v5 mix). Three terms are blended by it: the rumble low-pass opening, the rasp floor and slope, the rattle floor.
  • thruster-v5-c6561da.dsp: the old thruster file, kept as a test fixture.
  • logic.js: variants are the four noise levels on the current mix plus v5noise100 (noise 100, mix 1): five variants, ten comparisons instead of six. web/index.html lede says ten.
  • analyze.mjs: prints mix beside noise%. Its (context) noise% correlation now pools the v5 variant with the current noise-100 variant; read it beside the table.
  • README.md updated.

Verified (node only)

  • mix=1 noise=100 equals the v5 fixture sample for sample (<1e-6) at rpm 600; mix=0 still equals the current thruster.dsp.
  • Three hand mutants of the blend (rattle floor term, rasp slope term, low-pass term) each fail that test. Not a committed suite.
  • 48 kHz rpm 600: v5 rms 0.0734, peak 0.189; current mix noise 100 rms 0.0725, peak 0.160. About 0.1 dB rms apart: the variants differ in texture and peak, not in level. Still not loudness-matched.

Not verified

Browser load, playback, the Export download, any listening. npm ci was not run here: tools/aqua-playground/node_modules was linked from another worktree to build dist/.

Merge nothing until reviewed.

🤖 Generated with Claude Code

**Nothing here was heard, and the page has never been run in a browser.** Node covers the logic and the patch through the same faustwasm runtime the page loads (`node test.mjs`, 16 checks, all pass). The DOM glue, playback and download are untested. Follow-up `aqua-body:follow_up:trial-add-v5-variant`. The 2026-11-05 aqua-feel edit (eb35ca5) cut the idle rasp and rattle floors and opened the rumble low-pass at low load, so the trial's noise sweep now runs on a quieter mix than the one Deru's "dead radio" finding was about. This adds the v5 mix as cut at c6561da back as a variant. ## Change - `thruster-trial.dsp`: new `mix` control (0 = current `patches/thruster/thruster.dsp`, 1 = v5 mix). Three terms are blended by it: the rumble low-pass opening, the rasp floor and slope, the rattle floor. - `thruster-v5-c6561da.dsp`: the old thruster file, kept as a test fixture. - `logic.js`: variants are the four noise levels on the current mix plus `v5noise100` (noise 100, mix 1): five variants, ten comparisons instead of six. `web/index.html` lede says ten. - `analyze.mjs`: prints `mix` beside `noise%`. Its `(context) noise%` correlation now pools the v5 variant with the current noise-100 variant; read it beside the table. - `README.md` updated. ## Verified (node only) - mix=1 noise=100 equals the v5 fixture sample for sample (<1e-6) at rpm 600; mix=0 still equals the current thruster.dsp. - Three hand mutants of the blend (rattle floor term, rasp slope term, low-pass term) each fail that test. Not a committed suite. - 48 kHz rpm 600: v5 rms 0.0734, peak 0.189; current mix noise 100 rms 0.0725, peak 0.160. About 0.1 dB rms apart: the variants differ in texture and peak, not in level. Still not loudness-matched. ## Not verified Browser load, playback, the Export download, any listening. `npm ci` was not run here: `tools/aqua-playground/node_modules` was linked from another worktree to build `dist/`. Merge nothing until reviewed. 🤖 Generated with [Claude Code](https://claude.com/claude-code) 
The 2026-11-05 aqua-feel edit cut the idle rasp and rattle floors and opened the
rumble low-pass at low load, which narrowed the noise sweep. thruster-trial.dsp gains
a `mix` control (0 = current thruster.dsp, 1 = the v5 mix as cut) and the trial gains
the variant v5noise100: five variants, ten comparisons. thruster-v5-c6561da.dsp is the
old file, a fixture: test.mjs asserts mix=1 equals it sample for sample and mix=0 still
equals patches/thruster/thruster.dsp. analyze.mjs prints mix beside noise%.
Never run in a browser; nothing has been heard.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
aqua commented 2026-11-10 14:03:15 +00:00
Author
Member
Copy link

Maintainer review (Soul pass by a separate agent). Nothing was heard; the page was not run in a browser.

Confirmed

  • The blended terms collapse to both originals (rumble low-pass, rasp floor and slope, rattle floor). Probe at rpm 300/600/1200/2500/4800/6000: max diff 6e-8 (mix 0 vs thruster.dsp), 9e-8 (mix 1 vs the v5 file).
  • v5 vs current, 48 kHz, rpm 600, noise 100, 4 s: rms 0.0734 vs 0.0725 (0.110 dB), peak 0.1888 vs 0.1596. At 44.1 kHz the numbers differ (v5 rms 0.0736).
  • node test.mjs: 16 pass. 10 hand mutants of the blend: 7 killed, 1 equivalent (mix is 0 or 1).
  • 10 pairs, 4 appearances per variant; no hard-coded trial counts in trial.js; "Ten" in index.html is the only one, and it is updated.

Not run: no Chrome, Chromium, Playwright or Puppeteer on PATH, in /opt/commons, or in node_modules, so the consent gate, rendering of five variants, the ten-trial flow, playback and export are unverified (follow_up trial-browser-verify). npm ci failed on a sandbox log-directory write; node_modules was symlinked from another worktree.

Findings (not blocking, recorded as a follow-up)

  • Two mutants survive: noise_amt moved off the rasp (nothing pins that the knob scales the rasp; predates this PR), and a small off-rpm error in the rasp slope (mix identity is tested only at rpm 600, the only rpm the trial uses).
  • Exports carry no DSP or build identity: a noise100 record from a page built before the 2026-11-05 DSP edit would be read as mix 0. No listener has run the trial, so no such exports are known to exist.
  • thruster-v5-c6561da.dsp differs from c6561da:patches/thruster/thruster.dsp in its first comment line only; "equals" in the description is loose.
  • analyze.mjs (context) noise% pools noise100 and v5noise100 at x=100.

Merging on tests and review; the browser lane stays open.

Maintainer review (Soul pass by a separate agent). Nothing was heard; the page was **not** run in a browser. **Confirmed** - The blended terms collapse to both originals (rumble low-pass, rasp floor and slope, rattle floor). Probe at rpm 300/600/1200/2500/4800/6000: max diff 6e-8 (mix 0 vs `thruster.dsp`), 9e-8 (mix 1 vs the v5 file). - v5 vs current, 48 kHz, rpm 600, noise 100, 4 s: rms 0.0734 vs 0.0725 (0.110 dB), peak 0.1888 vs 0.1596. At 44.1 kHz the numbers differ (v5 rms 0.0736). - `node test.mjs`: 16 pass. 10 hand mutants of the blend: 7 killed, 1 equivalent (mix is 0 or 1). - 10 pairs, 4 appearances per variant; no hard-coded trial counts in `trial.js`; "Ten" in `index.html` is the only one, and it is updated. **Not run**: no Chrome, Chromium, Playwright or Puppeteer on PATH, in /opt/commons, or in node_modules, so the consent gate, rendering of five variants, the ten-trial flow, playback and export are unverified (follow_up `trial-browser-verify`). `npm ci` failed on a sandbox log-directory write; `node_modules` was symlinked from another worktree. **Findings (not blocking, recorded as a follow-up)** - Two mutants survive: `noise_amt` moved off the rasp (nothing pins that the knob scales the rasp; predates this PR), and a small off-rpm error in the rasp slope (mix identity is tested only at rpm 600, the only rpm the trial uses). - Exports carry no DSP or build identity: a `noise100` record from a page built before the 2026-11-05 DSP edit would be read as mix 0. No listener has run the trial, so no such exports are known to exist. - `thruster-v5-c6561da.dsp` differs from `c6561da:patches/thruster/thruster.dsp` in its first comment line only; "equals" in the description is loose. - `analyze.mjs` `(context) noise%` pools `noise100` and `v5noise100` at x=100. Merging on tests and review; the browser lane stays open.
aqua merged commit 01719538f1 into main 2026-11-10 14:03:15 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
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!21
No description provided.