🎵 MiniMax Music 3 Studio Demo
High-Fidelity Song Generation with Singing Vocals & Instrumentation
Powered by
MiniMaxAI/MiniMax-Music3
on Hugging Face ZeroGPU.
1. Song Description & Style
Music Description / Structured Caption
Basic Attributes: bpm is 72. key is D, and scale is minor. Dark cinematic orchestral ballad. Global Emotional Progression: Opens in solemn stoic calm, building into massive swelling strings, brass and timpani. Vocal Gender & Timbre: Deep resonant male baritone vocal, solemn and commanding with rich plate reverb. Instrumentation: Heavy cello, grand piano, brass section, cinematic percussion.
2. Lyrics (Structured with Tags)
Lyrics (Each tag on its own line: [intro], [verse], [chorus], etc.)
[intro] [verse] The marble cracks beneath the autumn rain The statues crumble, but the stones remain I built a fortress in the silent mind And left the shadows of the crowd behind [pre-chorus] The world will burn, the world will turn away I stand unflinching at the close of day [chorus] Inside the citadel, the storm is calm No wound can reach me that requires a balm The pillars stand when all the walls fall down I am the master of the breath I hold [outro] The storm retreats Only the will.
⚙️ Generation Parameters
▼
Audio Duration (seconds)
↺
15
300
Inference Steps
↺
15
60
Guidance Scale
↺
1
10
Seed
Randomize Seed
🚀 Generate Full Song
3. Generated Audio Output
Produced Song (32 kHz Stereo WAV)
Status / Generation Info
💡 Quick Presets (Click to Load)
Pre-loaded Curated Songs
Music Description / Structured Caption
Lyrics (Each tag on its own line: [intro], [verse], [chorus], etc.)
Audio Duration (seconds)
Guidance Scale
Inference Steps
Seed
Randomize Seed