USDX Fix Gap GUI — Validate UltraStar Deluxe GAP with AI

USDX Fix Gap GUI addresses a long-standing challenge in the UltraStar Deluxe community: ensuring that the GAP value in song files correctly aligns lyrics with the audio track. When GAP timing is off, singers experience mismatched lyrics and degraded gameplay. This project focuses on making GAP validation and correction reliable and efficient across large song libraries so that karaoke sessions feel smooth, precise, and enjoyable.

From an implementation perspective, the tool analyzes audio to detect onsets, beats, and silences, comparing these signals against lyric start metadata to validate or suggest corrected GAP values. A streamlined GUI guides users through loading a song, running the analysis, and applying fixes, with clear visual feedback to confirm alignment. The approach combines pragmatic audio signal processing (DSP) and light AI/ML heuristics to deliver accurate timing suggestions while remaining fast and practical.

My role spans Tool Design, Audio/ML Engineering, and Implementation. I defined the workflow for batch verification, designed the user interface for rapid inspection and correction, and implemented the analysis pipeline that computes suggested GAP adjustments. I also took care of packaging, documentation, and quality assurance to keep the tool reliable for community use across diverse song collections. I am not an expert in this domain, so I leaned heavily on AI‑assisted “vibe coding” — rapidly iterating with AI tools to explore approaches, refine heuristics, and validate ideas — to accelerate development and learning.