尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證The open-source engines behind the spectrometry.mp4 videos: fluid sims, field lines, particle swarms, a simulated CRT and molecule spectra, made with Python + ffmpeg. Free to use (MIT).
翻譯準備中
取自儲存庫 README
正在載入完整 README。 以下為已儲存的摘錄。
Fluid sims, field lines, particle swarms, bouncing shapes and molecules.
Written in Python, turned into video with ffmpeg. Free to use.
Hey! If you got here from one of my videos, you're in the right place. Nothing on the channel is
animated by hand: a real simulation runs, and every frame it produces goes straight into the video.
This repo is where I share the engines that do that, so you can pull them apart and make your own
things with them.
Tap the one that looks like what you saw.檢視完整 README ↗Fluid sims, field lines, particle swarms, bouncing shapes and molecules. Written in Python, turned into video with ffmpeg. Free to use.
Hey! If you got here from one of my videos, you're in the right place. Nothing on the channel is animated by hand: a real simulation runs, and every frame it produces goes straight into the video. This repo is where I share the engines that do that, along with the scene scripts behind most of my videos, so you can render them yourself, pull them apart, and make your own things with them.
Tap the one that looks like what you saw.
These made some of my newer videos, but their code isn't cleaned up for sharing yet. They'll show up here when it is. Follow along on Instagram in the meantime.
Seen a video that isn't on either list? Ask which engine made it.
You need Python 3.10 or newer and ffmpeg. A GPU makes things faster, but nothing needs one.
git clone https://github.com/ec175/spectrometry_public.git
cd spectrometry_public
python -m venv .venv
Turn the environment on. On Windows that's .venv\Scripts\activate; on macOS or Linux it's
source .venv/bin/activate. Then pick one:
🌪️ Render one of my scenes. Every engine has a cli.py. This is the wind tunnel clip from
the top of this page, at preview size:
pip install -r engines/wind-tunnel/src/requirements.txt
cd engines/wind-tunnel/src
python cli.py list
python cli.py tri_foil_air_fast --preview
Your video lands in engines/wind-tunnel/src/out/. Drop --preview for the full 1080×1920
version. Each engine page has a Render a video section with its own
commands.
🎬 Make a molecule video. Chemical Profiles is a finished manim scene: one molecule and its spectra.
pip install -r chemical_profiles/requirements.txt
cd chemical_profiles
manim -ql -r 540,960 chemical_profile.py VerticalProfile_Aspirin
📺 Put the CRT look on your own clip. The Oscilloscope film filter works on any video you already have.
pip install -r engines/oscilloscope/src/requirements.txt
python engines/oscilloscope/src/filter_cli.py your_clip.mp4 filmed.mp4 --glitch 0.8
🧪 Play with a simulation. Every engine page starts with a short Start here section you can paste straight into Python. Wind Tunnel is a good first one.
Anyone can use this code, for videos, school projects, your own engine, whatever you like. It's MIT licensed. The only thing I ask is that you leave a trail back here:
Made with spectrometry.mp4 engines by Clancy Earl - github.com/ec175/spectrometry_public
[](https://github.com/ec175/spectrometry_public)
If you post something you made, tag @spectrometry.mp4. I'd love to see it.
Using an AI coding agent? Point it at
AGENTS.md. It explains how the repo fits together and how to credit it. For a formal citation, useCITATION.cff(on desktop, it's the Cite this repository button).
No. Every engine runs on a normal CPU. If you have an NVIDIA card, the engines that support it use it automatically for the maths (CuPy) and the video encoding (NVENC), and quietly fall back to the CPU if anything is missing. All the timings in these docs came from one RTX 2060 with an i7-9700K, so read them as rough ratios.
No. Most manim projects do, but the chemical profile scenes use plain text rendering (Pango) everywhere, so you can skip the whole LaTeX install.
Most of them, yes. The scene scripts for 78 of my compositions are in the engines' src/ folders,
and python cli.py list shows them. Some are held back: scenes built on photos or music I
don't own, the missile and ammunition studies, sets that aren't finished, and the "not public
yet" engines above. Each
engine's catalogue marks which scenes ship. My finished videos also have music and edits that
aren't part of the engines.
Yes. Everything is written for vertical 1080×1920 because that's what I post, but the aspect
ratio lives in each engine's RenderConfig, never in the physics. Wind Tunnel even has a
right flow direction for the classic landscape tunnel.
No. They're representative simulations: peak positions come from literature values and group-contribution estimates, so they're good for learning and for video, but not for citing. Check against a real reference before you trust any band.
Open an issue with
the command you ran and the full error. Most problems are either ffmpeg missing from your PATH or
two engines sharing one virtual environment (use one per engine; Academia needs numpy<2).
engines/ one folder per engine: README, catalogue, stills, source
chemical_profiles/ a finished manim scene, the quickest way to a full video
docs/ the website, plus the preview clips on this page
tools/ rebuilds the catalogues and the stills pages
AGENTS.md instructions for AI coding agents