# Moosher edition — sources and review status

This is a fan-made Duvera Arcade edition, not an official creator website or endorsement.

## Visual reference

- Channel inspected: https://www.youtube.com/@Moosherr on 2026-09-23.
- Public channel display name: Moosher; handle: @Moosherr.
- Visible branding: teal banner, cow-pattern lettering, “Always Masti”, curly hair and dark glasses.
- Profile photo was used as a reference for new stylized artwork, not as evidence of endorsement.
- Higgsfield GPT Image 2.5 portrait job: `4bec17eb-9319-4244-9aed-c03d73ff3bc9`.
- Higgsfield GPT Image 2.5 six-expression atlas job: `d0e9effb-ea23-4443-afc4-712e97187e5e`.
- Runtime files are optimized exports: 640px WebP portrait, six 256px PNG emotes, a 255×60 three-frame sprite, and a title PNG. Atlas cropping, sprite packing and title compositing were performed in the Higgsfield media sandbox.
- Expression artwork is fictional cartoon emotion; it does not document a real crying or shouting incident.

## Audio source

“Drawing My Members LIVE under 5 mins (GONE WRONG)” — Moosher:
https://www.youtube.com/watch?v=LnAsPIae94Q

This is the creator’s edited livestream-highlights upload, not an unedited VOD. Cuts were selected using its automatic Hindi captions. Four provisional dialogue extracts are included; speaker identity, boundaries, background sounds and suitability still need a human listening pass. The game and sound-review page expose this status. No voice cloning or synthesized dialogue was used.

| File | Source seconds | Default trigger |
| --- | --- | --- |
| hello.mp3 | 1.85–2.80 | Starting a run |
| lets-go.mp3 | 68.80–70.45 | Scoring milestone |
| badhiya.mp3 | 110.25–112.45 | Win / new best |
| theek-hai.mp3 | 277.55–280.60 | Loss / retry reaction |

Total selected source duration: 7.85 seconds. Processing: 90Hz high-pass, loudness target −20 LUFS, true-peak target −2dB, 15ms fade-in, 45ms fade-out, mono 44.1kHz MP3 at 96kbps. Normalization improves consistency; it does not remove other voices or background music. Check pronunciation, full words, speaker identity and volume by ear.

Laughter, crying and shouting have **not** been verified or included. Do not relabel the generated facial expressions or these dialogue clips as authentic emotional recordings. Add only speaker-verified clips and record their provenance here and in the registry. A full-video analysis request remained queued during this work and was not used as evidence.

## Rights and attribution

Original game licenses are in their game directories. The extracted creator audio and the source likeness are not relicensed under the game’s GPL or MIT licenses. Rights remain with their respective owners. These media assets should be evaluated separately before redistribution beyond this fan-made project. Attribution and a fan label are not a grant of media rights.

## Human playtest

1. Open `creators/moosher/sounds.html`; preview each clip and visit its timestamp link.
2. Disable a clip if it contains another voice, a meme insert, distracting music, a bad cut or unwanted language.
3. Play each game on a phone and desktop. Listen for delayed reactions, overlap, harsh loudness and repetition.
4. Test Voice on/off, volume zero, individual clip disable, background-tab silence, and M in Moosh Flap.
5. Replace an audio file under the same filename to keep its mapping, or edit `assets/creator-profile.js` for new clips and timestamps. Update this document whenever a source changes.