https://blog.xijiang.org/tags/entertainment/ Recent content in Entertainment on Breeding and genetics Hugo en-us Tue, 28 Jul 2026 13:37:05 +0200 https://blog.xijiang.org/posts/karaoke-generator/ Tue, 28 Jul 2026 13:37:05 +0200 https://blog.xijiang.org/posts/karaoke-generator/ <h1 id="karaoke-environment-manual">Karaoke Environment Manual</h1> <p><img src="banner.jpg" alt="Karaoke Generator Banner"></p> <p>Welcome to the <strong>Karaoke Environment</strong>. This tool suite allows you to download videos from YouTube, separate vocals from instrumental music using AI models (<code>Spleeter</code>), generate timed lyrics/subtitles using speech recognition (<code>Whisper</code>), convert Traditional Chinese characters to Simplified Chinese (<code>OpenCC</code>), and render ready-to-sing Karaoke video files with burned-in subtitles.</p> <hr> <h2 id="-environment-directory-structure">📁 Environment Directory Structure</h2> <ul> <li><strong><code>karaoke-setup.sh</code></strong>: Main setup script to initialize environment structure, Python virtual environment, and executables.</li> <li><strong><code>.venv/</code></strong>: Python Virtual Environment containing all required AI models, libraries (<code>yt-dlp</code>, <code>whisper_timestamped</code>, <code>spleeter</code>, <code>opencc</code>, <code>tf_keras</code>, <code>torch</code>), and CLI tools in <code>.venv/bin/</code>.</li> <li><strong><code>songs/</code></strong>: Output directory storing intermediate assets (downloaded <code>.mp4</code>/<code>.wav</code>, separated vocals, generated <code>.srt</code> files).</li> <li><strong><code>result/</code></strong>: Final Karaoke video outputs: <ul> <li><strong><code>result/instrumental/</code></strong>: Pure backing tracks + original video with subtitled lyrics.</li> <li><strong><code>result/quarter-vocal/</code></strong>: Backing tracks + 25% low-volume original vocals + original video with subtitled lyrics.</li> </ul> </li> <li><strong><code>mastered/</code></strong>: Location for finalized songs.</li> <li><strong><code>pretrained_models/</code></strong>: Cache directory for pretrained stem separation and speech recognition models.</li> </ul> <hr> <h2 id="-setup--initialization">🛠️ Setup &amp; Initialization</h2> <p>To initialize or reproduce this environment, run:</p>