VTD – AI software for automatic video translation & voiceover

1. 🎙️ Vocal Separation & Background Music Retention Technology (Demucs AI)

Unlike ordinary video translation tools that destroy the original audio of the video, VTD integrates Demucs AI (the world's leading audio source separation technology):

• Automatically separates the original voice from the background music and sound effects (BGM & SFX).
• Overdubs the new narration precisely over the old voice position while the background music, engine noises, thunder, laughter... remain 100% intact..

2. 🤖 Super Accurate AI Extraction & Multilingual Translation

• Speech-to-Text (STT): Use the Faster-Whisper extremely powerful offline. Supports models ranging from ultra-lightweight (tiny, base) to highly specialized and accurate for Vietnamese (medium, large-v3).
• Smart Translation: Supports completely free offline translation with NLLB-200 (Meta AI) or translate using Gemini AI / QTran-Gemini (optimizing grammar and natural phrasing like a human translator).
• Glossary (Advanced Dictionary): Allows you to build your own terminology library by topic (e.g., Xianxia, Wuxia, Technology…). The AI will automatically translate accurately based on your context.

3. 🔊 Natural, Emotionally Expressive AI Voiceover (Piper TTS)

• Integrated Piper TTS (ONNX) ultra-fast offline mode.
• Provides a diverse collection of voices (Southern and Northern Vietnamese, male and female) with natural intonation and proper pauses, completely free of any “robot-like” sound.
• Smart Audio Sync: Automatically adjusts the speed of the AI voice to perfectly match the timecode of the speaking character in the original video.

4. ✏️ Interactive Translation Editor

Before exporting the final video, VTD provides an extremely intuitive editing interface so you can take full control of your product:

• You can easily edit individual translated sentences and adjust the start/end timecodes for each line.
• Freely change the voice for individual sentences or adjust the speaking speed separately to achieve perfection.

5. 🖼️ Video Frame & Vertical Video Design (Canvas Editor)

An exceptionally optimized feature for mobile content creators on platforms like TikTok, Shorts, and Reels:

• Convert to 9:16 Vertical Aspect Ratio: Automatically centers horizontal (16:9) videos within a vertical frame.
• Insert Custom Background Images: Fill the top and bottom empty spaces with artistic blurred background images or your own custom designs.
• Catchy Titles (Top/Bottom Text): Automatically renders titles at the top and bottom with auto-line-wrap technology and striking text border shadows.
• Custom Subtitle Styles: Choose Vietnamese fonts (Arial, Segoe UI, Impact…), adjust font size, apply advanced HSL color blending, and create professional opaque boxes.
• Watermarking: Supports inserting text or image watermarks with adjustable opacity to protect your video copyright.

6. 🛠️ Practical Content Creator Support Features

• Video Trimming: Select the desired time range to work on (e.g., from minute 01:00 to 03:00) directly on the app.
• Insert Outro & Midroll Ads: Allows inserting introductory clips/images at the end of the video or placing ads in the middle of the video at predetermined timestamps.
• On-Screen Text Scanning (OCR): Integrated EasyOCR / PaddleOCR to scan and translate text appearing on the original video screen, automatically supplementing missing subtitles.

⚙️ RECOMMENDED SYSTEM REQUIREMENTS

📥 DELIVERED PACKAGE INCLUDES:

• Full application Source Code written in Python, highly optimized and easy to upgrade.
• Complete UI interface with a pre-built executable file (.exe) to run immediately.
• Detailed step-by-step instruction file for installing offline libraries (Whisper, NLLB, EasyOCR, Piper TTS).