从任意 .wav 音频文件生成特斯拉自定义灯光秀（.fseq）。

这是一个Claude代码技能，并且是 独立的 Python CLI。指向一首歌，告诉它你的特斯拉模型，它生成一个 验证器清理.fseq与音频配对——把两者放进U盘，你就 准备好通过Toybox和Light Show播放→→Schedule Show。

支持的型号：

• Model 3（2024年前）
• Model 3 Highland（2024+改款——解锁内饰装饰RGB条）
• Model S（2021+）
• Model X（2021+）
• Y型机
• 赛博卡车

Highland 节目是向前兼容的：它们是包含 200 个频道的文件，包括 既有正常光照，也有室内RGB色相数据。高地.fseq 在更新前的Model 3上播放完全正常——固件会忽略这些通道 车子没有。

有什么不同

其他节目生成器则选择节拍和手电筒。这个节目把节目当作 三幕剧：

• 叙事弧线——每部剧都有开场（安静）、铺位（密度提升）， 高潮（全力+身体动作）和尾声（平静渐弱）。高潮 总是落在歌曲中最响亮的持续时刻，所以戏剧性 揭示其实和音乐很搭。

tesla-light-show

Generate Tesla custom light shows (.fseq) from any .wav audio file.

This is a Claude Code skill and a standalone Python CLI. Point it at a song, tell it your Tesla model, and it produces a validator-clean .fseq paired with the audio — drop both on a USB stick and you’re ready to play the show via Toybox → Light Show → Schedule Show.

Supported models:

• Model 3 (pre-2024)
• Model 3 Highland (2024+ refresh — unlocks the interior accent RGB strip)
• Model S (2021+)
• Model X (2021+)
• Model Y
• Cybertruck

Highland shows are forward-compatible: they’re 200-channel files that include both the normal lighting and the interior RGB hue data. A Highland .fseq plays on a pre-refresh Model 3 just fine — the firmware ignores the channels the car doesn’t have.

What makes it different

Other show generators pick beats and flash lights. This one treats the show as a three-act piece:

• Narrative arc — every show has an intro (quiet), build (ramping density), climax (full intensity + physical movement), and outro (calm fade). The climax always lands on the loudest sustained moment of the song, so the dramatic reveal actually lines up with the music.
• Physical choreography at the climax — liftgate pre-opens 14 s early, mirrors flap in a continuous wiper pattern, charge port does the rainbow Dance, and model-specific tricks fire (Model S door handles; Model X falcon
  • front doors).
• Model-aware lighting — uses ramping on cars that support it (Model 3/Y/CT), crisp boolean hits on cars that don’t (S/X). Generates 200-channel output for Cybertruck so the light bars and interior RGB actually do something.
• Stereo-aware audio analysis — panning information drives which side of the car lights up (hi-hats panned right → right-side markers). Stereo width detection catches “the mix opens up” moments and fires a matching light wash.
• HPSS percussive separation — kick/snare beats are detected on a clean drums-only signal rather than inferred from EQ ratios, so non-percussive music (anthems, ballads) still gets good beat synchronisation.

Requirements

• Python 3.10+
• numpy (required)
• librosa (optional — chroma, CQT, structural segmentation, fallback beat tracker)
• madmom (optional — state-of-the-art beat + downbeat tracker, cross-checks librosa)
• scipy (optional — speeds up HPSS; pure-numpy fallback included)

pip install numpy
pip install librosa    # optional, ~80 MB
pip install madmom     # optional, adds ~60 MB; cross-checks librosa beats
pip install scipy      # optional, speeds up HPSS

When both librosa and madmom are installed the analyzer runs them as a dual-tracker cross-check: beats confirmed by both get a higher confidence score, madmom’s downbeats give us real meter detection (3/4 vs 4/4) and proper downbeats, and the composer gates its biggest hits on high-confidence