Performance HUD
When playback isn't smooth or you want to understand how Orbit is performing on your system, the Performance HUD shows real-time diagnostic information. It's useful for investigating glitches, comparing different settings, and confirming that your system is handling the workload.
Opening the HUD
- Menu: Debug → Toggle Performance HUD (⌘⌥D)
- Shows a floating window with real-time performance metrics.
Metrics to watch
| Metric | What it means | Healthy range / guidance |
|---|---|---|
| CPU Usage | Host CPU usage sample | Low and steady; spikes with render time suggest contention |
| Sample Rate / Buffer Size | Device clock and buffer length | Buffer 128–512 frames typical; unexpected changes = device switch |
| Render avg/p95/max (ms) | Engine render time per buffer | Stay well below buffer duration (<25% is comfortable); p95/max near buffer size risk underruns |
| Load (%) | Render time as % of buffer duration | <50% steady is good; >75% sustained or spiky warrants attention |
| Underruns | Buffer underrun count | 0 and not increasing. A single blip might be benign; rising counts during playback need attention |
| Faults | Engine fault count | 0; investigate any non-zero value |
| Active Objects | Objects currently contributing | Matches ADM expectation; big jumps without timeline changes suggest ADM/mute/solo issues |
| Event/Prep/Log Queues | Internal queue depths | Near zero or stable; growth correlates with contention or I/O stalls |
| Blocks | Total buffers rendered | Reference counter to align timing between runs |
| Transport | Engine transport time (s) | Should track playback; sanity check HUD sync |
When to use
- Investigate glitches: watch underruns and render p95/max.
- Compare routing/mode changes: ensure active object counts and queues stay stable.