Something went wrong. Try again.
Monorepo for Aesthetic.Computer aesthetic.computer
Something went wrong. Try again.
README.md
# pop / xpld — "Explody Head"
The start of a pop lane for **`xpld` — "Explody Head"**, written and
performed by **Jeffrey Alan Scudder** (Whistlegraph, 2023, 191K views
across five tagged posts). A solo sung whistlegraph — one voice, one pen,
no accompaniment — an emo self-portrait: hit a hump, gave up, got a new
name, and a face full of internet, safe in a bubble with the emo-gee
that signifies an identity.
What's here: the **measured foundation** (tempo, key, the melody per
word, a word-level transcript, the per-phrase take comparison, a
vocalized click track) and the **aesthetivox pass** — the note chart
(`bin/notes.py`), the sample bank (`vox/`), and the WORLD sung-note
render (`bin/aesthetivox.py` → `out/xpld-aesthetivox.wav/.mp3` and the
`-halo` variant).
## The takes
All five tagged posts are downloaded (`bin/fetch.sh` re-pulls them from
the assets mirror; `source/` stays out of git):
| post | date | length | what it is |
| ---- | ---- | ------ | ---------- |
| **7275499036398865706** | 2023-09-06 | 100 s | **the spine** — 95.4K views: spoken intro (0–4.5 s, "Hey come here, I want to show you a whistlegraph called Explodey Head"), then the whole song clean (6.0–99.2 s) |
| 7257275807209458986 | 2023-07-18 | 82 s | "a real adventure" — the premiere; sung straight through, quiet room |
| 7257695616694881582 | 2023-07-20 | 93 s | "thing like a doggy #mood" take — the **quietest room of the five** (noise floor 0.0028, 20× below the primary) |
| 7258670360357276970 | 2023-07-22 | 85 s | "no match for my graphic tee ✒️💀" — real dog barks and coughs mid-take; carries the **cleanest "do you see"** |
| 7278943795482283307 | 2023-09-15 | 81 s | "whistlegraphing is not lame" — the last take, nine days after the spine; the most-voiced delivery of several phrases |
`bin/takes.py` → `analysis/takes.json` is the per-phrase comparison
(loner's take-rotation move): every canonical phrase of the primary
matched into every other take by fuzzy transcript, measured for voiced
coverage, cents to the derived grid (in each…