Virenesis Music Studio — AI Music & Lyric Production Dashboard
The Challenge
Generating high-fidelity AI music often produces pitch instability, vocal echoing, static background beats, and imprecise language enunciation. Existing tools lack stem isolation and dual-version comparison workflows.
Technical Stack & Engines
The Solution: SFT Audio Pipeline & High-Tech Fourier Visualizers
Architected a real-time web application integrating the ACE-Step 1.5 SFT model and Gemini 3.6 Flash. Every generation produces two distinct track variations (Version A & Version B) with dynamic BPM extraction, auto-detected Tagalog/OPM vs English lyric phrasing, Suno/TopMediai pacing meta-tags, and 100% strict stem separation.
Dual-Version Generation
Generates Version A and Version B simultaneously with distinct random seeds, custom BPM variations, and unique vibe tags.
Strict Stem Isolation
Instant zero-bleed stem extraction allowing users to play Full Mix, Clean Vocals, or Pure Instrumental tracks independently.
Suno/TopMediai Pacing Formatter
Injects structural section headers (`[Verse 1 | slow tempo | clear vocals]`) and auto-detects lyric vocabulary.
Fourier FFT Spectrum Visualizer
Dual-mirrored HTML5 Web Audio FFT spectrum analyzer with neon glow, peak gravity dots, and oscilloscope sine overlay.
UI Screenshots
Main Generation Workspace & Cyber Tagger Controls
High-Tech Fourier Transform Spectrum Analyzer & Player Deck