TIMELINE
Mission
Become the global gateway to music: the brand people think of when they want to try playing, anywhere, anytime.
Values
Become the global gateway to music: the brand people think of when they want to try playing, anywhere, anytime.
Values
Playful – Music should be joyful, not stressful.
Accessible – No gatekeeping, no excuses, music belongs to everyone.
Fearless – Encourage everybody to try, even if they’ve never touched an instrument.
Creative – Always experiment, always push for new ways to play.
Brand Story
Gamsing’s founder Bruce’s first stage wasn’t a club or a garage.
It was a beat-up kitchen radio, in a small village in rural China.
He was six when he first heard Van Halen, Lynyrd Skynyrd, and Led Zeppelin, and fell hopelessly in love with rock music.
He wanted to play, but the thought stayed locked inside.
Drums cost too much. Guitars were too far away. Music lessons weren’t even a conversation his family could have.
The years rolled on, school exams, endless overtime, crowded buses, a tiny rented apartment, diapers and midnight cries. Life became a long, noisy grind.
At forty, Bruce sold his first company and, for the first time, felt truly free.
One sleepless night, he thought about that six-year-old boy sitting by the radio, wishing he could play.
The next morning, he signed up for drum lessons.
For the first time in his life, he sat behind a real kit, and it felt electric.
But it also made him wonder:
“Why should anyone have to wait three decades to feel like this?”
The Air Drum Kit is his answer.
A pocket-sized stage for anyone who’s ever tapped a beat on the table, air-drummed in traffic, or turned up the radio and wished they could play along.
Bruce is still chasing that same rush he felt in the kitchen.
Only now, he’s handing the sticks to you.
AI SMART DRUMMING AI SMART DRUMMING
From motion to meaningful rhythm.
From motion to meaningful rhythm.
01. AUDIO UNDERSTANDING & AI RECOGNITION
Understand Every Beat
The system extracts time‑frequency features, drum onset information, and short‑term energy changes, converting raw audio into structured data that AI can interpret. A two‑stage CRNN model then performs hierarchical recognition: Stage 1 identifies silence and active drum performance regions. Stage 2 distinguishes steady rhythm sections from transitional drum fills. Background noise, decaying sounds, and low‑level interference are effectively filtered out.
Core Technologies: Audio feature extraction, drum onset detection, short‑term energy analysis, and two‑stage CRNN classification.
02. INTELLIGENT SEGMENTATION & BOUNDARY OPTIMIZATION
Segmentation That Follows the Performance
Instead of cutting audio at fixed time intervals, the system converts frame‑level predictions into continuous rhythm, fill, and silence sections based on the actual musical structure. Each initial boundary is refined through energy analysis and localized searching:
- Snaps cut points to natural energy valleys
- Relabels and restores continuous silence regions
- Merges short fragments caused by false detection
- Divides overly long sections into more meaningful musical units
Core Technologies: Temporal segmentation, candidate refinement, energy‑valley alignment, silence relabeling, fragment merging, and long‑section splitting.
03. GROOVE CLASSIFICATION & TEMPO ANALYSIS
Turn Audio Clips into Meaningful Rhythm Structures
After segmentation, the system analyzes the drum‑onset distribution, accent positions, note density, and timing relationships within each section. It can then:
- Compare structural similarities between rhythm sections
- Group similar grooves into the same rhythm pattern
- Identify repeated or closely related sections
- Estimate tempo and correct common half‑time or double‑time errors
- Produce structured results with groove and tempo metadata
Core Technologies: Onset‑based rhythm representation, similarity calculation, pattern clustering, tempo estimation, and reference alignment。




