AI Agent Hub
Back to skills
Japanese Video Subtitle Burner icon

Japanese Video Subtitle Burner

Design & Media Updated 2026.08.29

Paste the following prompt into your AI chat to install this skill:

Please install @user_918cb67a/video-japanese-subtitle according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem It Solves

Japanese videos often contain BGM, sound effects, and colloquial speech, making manual transcription expensive. openai-whisper's ja→zh translation can also sound stiff. This skill packages Japanese recognition, Chinese translation, and subtitle burn-in into a reusable local pipeline.

Workflow And Key Capabilities

The process follows six steps: extract audio, transcribe with Whisper, translate, convert SRT→ASS, burn subtitles with ffmpeg, and spot-check frames with ffprobe. Key behaviors include:
- Transcribe, then translate: Whisper handles Japanese speech-to-text, while an LLM handles translation, falling back to MyMemory if needed.
- Burn-in stability: Subtitles are converted to ASS with white text, black outline, and bottom center alignment; Windows path issues are avoided with relative paths and cwd=ass_dir.
- Encoder fallback: It prefers GPU NVENC and falls back to libx264 -preset ultrafast on failure.
- Resume support: Each step checks whether the output file already exists, so completed steps can be skipped on longer videos.

Limits To Keep In Mind

The scope is Japanese to Chinese subtitle processing, useful for engineers doing local or semi-automated video localization. The source notes that Whisper base can struggle with mixed audio, and faster-whisper is a future improvement direction. If the input is not Japanese or you need other language pairs, you must adjust the Whisper language parameter and translation flow.

Use Cases

  • Process Japanese tutorial or Vlog videos, convert spoken content into Chinese subtitles, and burn them into MP4 files.
  • Batch-process 30-minute Japanese clips with resumable transcription, translation, and subtitle-rendered output.
  • Run ffmpeg/Whisper on Windows while avoiding path, GBK, and SRT parsing failures that break burn-in.
  • Keep subtitle entries aligned when LLM output is unstable by falling back to MyMemory and padding missing entries.

Best For

  • Video editors localizing Japanese content who need spoken audio quickly turned into subtitled master files.
  • Creators publishing Japanese tutorials who need stable long-audio transcription with BGM or sound effects.
  • Media automation engineers who need to chain Whisper, LLM translation, and ffmpeg into a resumable pipeline.
  • Operations teams without a dedicated Japanese translation team who need LLM/MyMemory fallback subtitles.