Back to corpus
architecturetechnical paper candidatescore 36

Stage 3: EXPAND + MASTER PLAN

**R1: Radeon 780M Compute Shader Compatibility** - Failure scenario: Unity compute shaders (LumeDepthReproject.compute, LumeOpticalFlow.compute) fail on AMD Vulkan/DX12 drivers. These were developed on Apple Metal (Mac5). - Probability: 20% - Impact: HIGH (no visual pipeline without GPU reprojection) - Mitigation: Test on K11 Day 2. If Vulkan fails, try DX12 backend. If both fail, use LUME format (CPU-reprojected in pointcloud_pub.py) instead of LUMD (GPU-reprojected). Wave 1 graceful-degrade path exists. - Validat

Open full HTMLDownload HTMLCorpus JSONPaper schema

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

**R1: Radeon 780M Compute Shader Compatibility** - Failure scenario: Unity compute shaders (LumeDepthReproject.compute, LumeOpticalFlow.compute) fail on AMD Vulkan/DX12 drivers. These were developed on Apple Metal (Mac5). - Probability: 20% - Impact: HIGH (no visual pipeline without GPU reprojection) - Mitigation: Test on K11 Day 2. If Vulkan fails, try DX12 backend. If both fail, use LUME format (CPU-reprojected in pointcloud_pub.py) instead of LUMD (GPU-reprojected). Wave 1 graceful-degrade path exists. - Validation: `pointcloud_pub.py --synthetic-depth` + Unity running on K11 produces visible animated surface. **R2: Tailscale Mesh Reliability for LUMM/LUMF to iOS** - Failure scenario: Tailscale has documented 100% packet loss episodes (Mac5 unreachable during some sessions). If LUMM/LUMF don't reach iOS, mocopi skeleton never enters the 128D vector, SAN loses body input, music becomes unresponsive. - Probability: 30% - Impact: HIGH (music disconnects from body motion) - Mitigation: (1) iOS local fallback: if no LUMM received for 3s, fall back to local CoreMotion/Vision-only 128D (dims [76:100] stay zero, as they are today pre-LUMM). (2) LUMF fallback: if no LUMF for 3s, ambient audio dims [104:108] stay zero. (3) Health check: lume_packet_inspector.py on K11 logs drop rates. echelon_relay.py pings Tailscale targets on startup. - Validation: Unplug Tailscale adapter on K11 while iOS is consuming. Verify music continues with degraded but not broken behavior. **R3: mocopi BLE Pairing Unreliability** - Failure scenario: Sony mocopi 6x sensors fail to pair via BLE to K11 Windows. BLE pairing on Windows is notoriously flaky. Without mocopi, LUMM :9702 is empty. - Probability: 35% - Impact: MEDIUM-HIGH (lose skeleton input, but depth + audio still work) - Mitigation: (1) Primary: use Sony mocopi app on phone, phone connects to sensors via BLE, phone sends to K11 via OSC :9500 instead of K11 BLE direct. mocopi_bridge.py already has an OSC listener path. (2) Secondary: skip mocopi entirely. Depth camera optical flow + audio transients still drive visuals. iOS still has CoreMotion/Vision for music. - Validation: Run system for 30 minutes with only Femto Bolt + UMA-8. Verify visuals and music are engaging without skeleton data. **R4: Music-to-Visual Synchronization Drift** - Failure scenario: K11 visuals react to local depth/audio sensors (~13ms latency). Mac5 audio reacts to iOS-processed data (~50ms). A hand clap is visible in depth at T=0 but audible at T=37ms later. Perceptible desync during fast transients. - Probability: 50% (will happen, question is severity) - Impact: MEDIUM (subconscious unease, not show-stopping) - Mitigation: (1) Local audio reactivity on K11 via LUMF means visuals DO react to the clap at T=8ms (via LumeAudioFftReceiver transie

Promotion decision

What has to happen next

Promote into a technical note or architecture paper with implementation anchors.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.