DJ Dojo

About the project

What this is

DJ Dojo is a VR “deck lab” built for Quest where you can jump in, grab a pair of decks and a mixer, and actually practice mixing without fighting the controls or the UI. The goal wasn’t to reinvent DJing, it was to make a setup that feels good in-headset, runs well on standalone hardware, and doesn’t scare new people away.

How it’s put together

Under the hood it’s a Unity project targeting Quest with:

C# deck/mixer components for transport (play, pause, cue, loops), pitch/tempo, EQ, filters and crossfader curves. Audio is routed through Unity’s mixer with separate groups for each deck, cue, master and FX so metering and visualizers stay clean.
Meta avatars + tracked hands so you show up as “you” by default and can literally grab knobs, faders and jogs. Grabbables use tuned colliders, clamped angles and a bit of smoothing so you don’t accidentally slam the crossfader because your hand twitched.
Quest‑friendly visuals: a small set of stylized “dojo” stages sharing materials/shaders, baked lighting where possible, very few dynamic lights, and lightweight waveform/level shaders so the GPU doesn’t melt.

Interaction-wise, everything is built around “don’t let the user feel dumb”:

Controls are big, with sensible dead zones and sensitivity curves. Beginners can make broad moves; people who know what they’re doing can still ride gain/EQ precisely.
Critical actions (load track, hard stop, etc.) have visual confirmations and short debounce windows so you don’t ruin a mix with one bad grab.
Waveforms, meters and cue points are visually consistent across the whole rig, so once you understand one deck, you understand all of it.

Training / performance side

On top of the core rig there are a few systems for actually getting better:

Practice routines that spawn predefined track pairs, set BPM targets, and watch how long you keep phase within a tolerance. At the end you get a basic “you held this mix in time for X seconds” kind of summary.
Session tracking that logs session length, attempts and success rate locally. It’s intentionally low‑key right now, but it’s the data backbone for future progression (belts/grades, etc.).
Performance mode that strips UI down to the essentials and frames the view so Quest capture/streaming doesn’t look like a debug build.

Things that were harder than expected

Feel > features: it’s easy to wire up a bunch of knobs; it’s harder to make them feel trustworthy. Most time went into collider tuning, rotation limits, input smoothing and haptics timing.
Style vs. framerate: pushing stylized, anime‑ish lighting and shaders on Quest without tanking FPS meant cutting a lot of “cool” ideas and going back to simpler materials and baked light.
Explaining DJ concepts: the app tries to teach beat‑matching and phrasing through short, playable drills instead of walls of text, which means the UX has to do a lot of heavy lifting.

Where it’s going next

Short‑term plans:

Build a real progression layer on top of the existing practice metrics (belts, goals, unlocks).
Add a lightweight coach/ghost mode that can show “suggested moves” over the deck instead of more UI.
Harden multiplayer so small dojo sessions with friends are stable, with synchronized FX and better session recovery.

It’s still early, but the core loop is there: put on the headset, stand in front of a rig that feels like actual hardware, and slowly get less bad at mixing without fighting the tech.