Karaoke-Maker API AI 驱动的卡拉 OK 解决方案

Input audio file (MP3/WAV)

Input vocal-removed audio file (MP3/WAV) for generating instrumental version

ASS subtitle file (must contain word-by-word effects)

Video resolution (e.g., 1280x720)

Whether to render audio visualization (disabling this will not render waves/spectrum)

Whether to force use spectrum when visualization is enabled; otherwise follow viz_type setting

Visualization type: waves (waveform, faster) / spectrum (frequency spectrum, slower)

Visualization area height (in pixels)

Position: top / bottom / center

Color in waves mode (0xRRGGBB or color name); color scheme in spectrum mode (rainbow/moreland/viridis etc.)

Visualization layer opacity 0-1

Optional: Directory containing required Chinese fonts (TTF/OTF) for ASS font matching

Output frame rate (reducing this value can significantly speed up processing, e.g., 24)

x264 encoding preset (ultrafast/superfast/veryfast/faster/fast/medium...)

x264 CRF value (larger value means smaller file size and faster speed, range is usually 18-32, recommended 26-30 for acceleration)

FFmpeg thread count (0 means auto-detect)

Filter graph parallel thread count (0 lets FFmpeg choose automatically)

Visualization layer refresh frame rate (0 means follow the fps value)

Optional: Comma-separated list of image paths (jpg/png/webp etc.)

Display duration per image (in seconds), only effective when image list is provided

Optional: Comma-separated list of video paths (mp4/mov etc.)

Whether to loop playback when video duration is shorter than audio duration

Optional: Title displayed on the first frame of the video

Optional: Author displayed on the first frame of the video