An 11-minute (672 s) narrated study video in the style of Justin Sung: a priming question, the big picture first, five linked chunks, and a retrieval challenge at the end.
- Source text: https://hekm.co/article/1259
- Narration (full text):
../../transcripts/gharra-sermon-82.txt
Files
| Path | What |
|---|---|
gharra-sermon-82.mp4 |
Final video: 1920×1080 H.264, AAC audio, embedded English subtitle track |
gharra-sermon-82.srt |
The same subtitles as a separate file (214 cues) |
slides/s00.png … slides/s47.png |
The 48 rendered slides, in order |
build.py |
Generates the slides, narration audio, subtitles, and video clips |
Structure
The video is built around one question: Why do people waste the one window of time they are given?
Gifts ──tools for the test──▶ Window
Trap ──makes it feel far───▶ End
Response (taqwa) uses the Gifts, acts inside the Window,
sees through the Trap, prepares for the End
Rebuild
Requirements: Python 3, edge-tts (network access to Microsoft's TTS service), ffmpeg/ffprobe, and a headless Chrome.
Fonts go in ~/.fonts (override with FONTDIR): Amiri Regular and Bold, Caveat, and Inter, all from github.com/google/fonts (ofl/).
pip install edge-tts
python3 build.py # writes slides, audio, clips, subs.srt to ./build (override with OUT=...)
cd build
ffmpeg -f concat -safe 0 -i clips.txt -i subs.srt -map 0 -map 1 -c copy -c:s mov_text \
-metadata:s:s:0 language=eng -movflags +faststart gharra-sermon-82.mp4
PREVIEW=0,5,9 python3 build.py renders only those slides, with no audio. Set CHROME=/path/to/chrome if Chrome is not at the default path.
Notes
- Voices:
en-US-AndrewNeuralfor English andar-SA-HamedNeuralfor Arabic (rate −12%). These are generic AI voices; the video copies the teaching style, not Justin Sung's voice. - The narration is original English commentary. Arabic quotes are short lines from the sermon. The Arabic TTS may mispronounce some words, so it is worth checking by an Arabic speaker.