SmartRTSP uses hardware-accelerated video decoding for all streams, keeping the baseline power usage low. The primary battery variable is AI detection — how often and how deeply the app analyses incoming frames. Performance Mode controls this trade-off directly.
On-device AI detection analyses video frames locally — nothing leaves your network. The three performance modes adjust how aggressively detection runs, directly controlling the balance between responsiveness and battery life.
Performance Modes
Reduces detection frequency to the minimum needed to catch meaningful events. Frame analysis intervals are widened, and intensive detection (person, face) runs only when motion has been detected first via cascade detection.
The default mode. Detection runs frequently enough to catch events reliably, with cascade detection filtering out empty frames before running person or face models. Good balance for everyday home monitoring and security use cases.
Detection runs at maximum frequency with the shortest frame analysis intervals. All detection types (motion, person, face, sound) operate at their lowest possible latency. Recommended when detection speed is the priority and a charger is available.
Battery Usage at a Glance
| Mode | Battery / hr | Detection Frequency | Best For |
|---|---|---|---|
| Power Saving | ~5% | Minimal — motion-gated only | All-day battery, overnight |
| Balanced | ~10% | Moderate — cascade filtered | Everyday home & business use |
| High Performance | ~20% | Maximum — all detection types | Plugged-in Mac or iPad station |
Battery estimates are approximate and vary by device model, number of cameras, stream resolution, and ambient temperature. Measured on a single RTSP stream at 1080p with detection active.
Cascade Detection — 60–80% CPU Savings
Cascade detection is SmartRTSP's two-stage approach to AI analysis. Instead of running every detection model on every frame, a fast lightweight motion check acts as a gating stage. Only frames where motion is detected proceed to the more computationally intensive person, face, or sound detection models.
~100ms
~600ms
~800ms
When no motion is detected, the pipeline stops at the first stage. The expensive models never run on empty frames — saving 60–80% CPU compared to running full detection unconditionally.
Tips for Reducing Battery Drain
-
✓
Use sub-stream URLs in grid view. Lower-resolution sub-streams require less decode work per frame. In a multi-camera grid, switching each cell to the camera's sub-stream URL cuts power draw noticeably while keeping the view clear enough for monitoring.
-
✓
Disable detection types you don't need. If you only need person alerts, turn off face and sound detection in Settings. Each active detection type adds to CPU usage — fewer active detectors means lower drain.
-
✓
Reduce camera frame rate at the source. Log into each camera's web interface and lower the frame rate from 30 fps to 15 fps. Detection catches events reliably at 15 fps, and the app decodes half as many frames per second.
-
✓
Use H.265 streams where possible. H.265/HEVC streams carry the same video quality at roughly half the bitrate of H.264. Smaller packets mean less Wi-Fi radio activity, contributing to overall power savings.
-
✓
Reduce screen brightness for always-on setups. The display is one of the biggest power draws on any iPhone or iPad. For a dedicated monitoring setup, reduce brightness significantly or enable auto-lock and rely on notification alerts instead of a live view.
-
✓
Enable Background App Refresh for background monitoring. Go to iOS Settings → General → Background App Refresh → SmartRTSP and ensure it is enabled. This allows SmartRTSP to continue monitoring and delivering alerts when the app is in the background, without requiring the screen to stay on.
Frequently Asked Questions
Which performance mode should I use for everyday monitoring?
What is cascade detection?
Why is person detection slower than motion detection?
Does SmartRTSP use hardware video decoding?
Download SmartRTSP Free
On-device AI detection, three performance modes, and hardware-accelerated streaming — free for iPhone, iPad & Mac.