AI Subtitle Studio includes a high-performance translation engine designed to translate entire subtitle files while maintaining timing, formatting, and context. By utilizing the DeepSeek AI provider (more coming soon), it can handle both individual files and massive batch translations.
1. How the Translation Process Works
The studio doesn’t just translate line-by-line; it uses a “Contextual Batching” approach to ensure higher accuracy.
- Smart Batching: Subtitles are grouped into batches (default: 15 lines).
- Context Awareness: The engine provides the AI with “Context Size” (default: 2 lines before/after), helping it understand who is speaking and the flow of the conversation.
- Parallel Processing: To speed up the process, the studio can run multiple translation batches simultaneously (default: 3 parallel batches).
- Metric Conversion: A unique feature that automatically converts imperial units to metric (e.g., miles to kilometers) within the translated text.
2. Supported Languages
The tool supports a vast range of languages for both source detection and target translation, including but not limited to:
- Common: English, Spanish, French, German, Italian, Portuguese, Russian.
- Asian: Chinese (Simplified/Traditional), Japanese, Korean, Thai, Vietnamese.
- Regional & Others: Dutch, Polish, Turkish, Arabic, Hindi, Swedish, Greek, and dozens more.
3. Translation Workflow
Step 1: Setup and Loading
You can start a translation in two ways:
- From Video: Add a video file; the studio will automatically scan for Embedded Tracks (inside the MKV/MP4) or Sidecar Files (external .srt/.vtt files in the same folder).
- Manual Load: Directly open existing subtitle files from your computer.
Step 2: Configure AI Settings
Before translating, you must ensure your AI credentials are set:
- API Key: Enter your DeepSeek API key in the settings menu.
- Acceleration: You can toggle GPU acceleration to speed up the local processing of subtitle files.
Step 3: Individual or Bulk Translation
- Individual: Select a file, choose your Source and Target languages, and click AI Translate. You can watch the progress in real-time through three phases: Preparing, Translating, and Finalizing.
- Bulk Mode: Select multiple videos/subtitles from your list. Set a global target language and click Bulk Translate to process your entire library in the background.
4. Advanced Features for Editors
The Translation Interface
Once a translation is in progress or finished, the studio provides a dual-pane editor:
- Sync Scrolling: Moving through the source text automatically scrolls the translated text to the same timestamp.
- Interactive Timeline: A visual timeline at the bottom allows you to “zoom” in on specific scenes to check subtitle density.
- Live Preview: As translations return from the AI, they are instantly populated into the translated blocks so you can begin proofreading immediately.
Auto-Detection
The studio attempts to guess the source language by:
- Checking for language codes in the filename (e.g.,
movie.en.srt). - Reading metadata from the embedded video tracks.
5. Troubleshooting
- “Missing API Key”: If translation won’t start, ensure your DeepSeek key is saved in the AI Providers settings.
- Image-Based Subs: Note that image-based formats like PGS or VobSub cannot be translated directly. You must first use the “Extract” feature (which may require OCR) to turn them into text before the translation engine can read them.
- “Permission Denied”: If saving a translated file fails, the app will offer a save dialog to choose a different folder.