Here's an updated summary in English for Reddit/Facebook/ipcamtalk:
Blue Iris - Progressive Live View Delay (~16 hour cycle) - SOLVED WITH WORKAROUND
System specs:
Live view delay starts at 0 seconds after BI service restart, then progressively increases over ~16 hours until reaching 1-2 minutes. Restarting the BI service immediately fixes the delay until it accumulates again. System occasionally freezes requiring a wait before becoming responsive again.
What we verified/ruled out:
Scheduled automatic BI service restart twice daily via Windows Task Scheduler:
Despite all resources being available (CPU, RAM, GPU, disk, network), BI progressively accumulates a buffer over ~16 hours that causes live view delay. The buffer resets completely with a service restart. No obvious bottleneck identified. Could be a memory leak in BI? A known bug with 4K RTSP streams?
Question for the community:
Has anyone experienced this exact pattern — progressive live view delay that accumulates over ~16 hours and resets with a service restart, with no resource saturation? Any insight into the root cause would be greatly appreciated!
System specs:
- Intel i5-12600K, 32GB DDR5 4800MHz
- ASUS PRIME Z690-P (DDR5 version)
- 11 Lorex E891AB cameras (4K H.264, 15fps, 8192 Kbps) + 1 E892AB (disabled - clock issues)
- Blue Iris 6.0.9.3
- CodeProject.AI with YOLOv5 6.2 .NET and Face Recognition uninstalled
- Windows 11, HP ProCurve 3500yl-24G POE switch
- Storage: 18TB HDD for recordings D: of 5 cameras, 20TB HDD E: for recording of 6 cameras, SSD for DB C:
- Realtek Gaming 2.5Gb NIC
Live view delay starts at 0 seconds after BI service restart, then progressively increases over ~16 hours until reaching 1-2 minutes. Restarting the BI service immediately fixes the delay until it accumulates again. System occasionally freezes requiring a wait before becoming responsive again.
What we verified/ruled out:
- CPU usage: ~28-35%

- RAM usage: ~68%

- GPU usage: ~14%

- Disk usage: ~8%

- Network: occasional spikes to 10-29ms but not consistently correlated with delay

- No thermal throttling (CPU temps 45-55°C)

- Power plan: High Performance

- Single camera test: delay still present (~40-60 sec) — not a multi-camera resource issue

- CodeProject.AI disabled: delay persists

- GOP/I-Frame interval: set to 15 on both main and sub streams

- Receive buffer: reduced to 1.0 MB on all cameras (was 6-20MB)
- Limit decoding unless required: unchecked on all cameras
- Hardware accelerated decode: Intel® +VPP (global setting)
- Hardware decode per camera: Default (inherits global)
- Stream buffer time: 2.0 sec
- Decoder compatibility mode: unchecked
- Use RTSP/stream timecode: unchecked
- Send RTSP keep-alives: checked
- Max AI analyzers: reduced to 1
- Bit exact: unchecked
- CPU usage factor: 1.000
- Main Stream: H.264, 4K (3840x2160), 15fps, CBR, 8192 Kbps
- Sub Stream: H.264, 704x480, 15fps, CBR, 256 Kbps
- GOP Main Stream: 15 (1 keyframe per second)
- GOP Sub Stream: 15 (was 1 — caused excessive bitrate ~3644 kB/s)
- NTP time sync enabled with correct timezone (EST/EDT)
- Firmware updated to V18 on all E891AB cameras
- DB repair completed successfully (DBRegenerate: success)
- Startup apps disabled: Microsoft Teams, Edge, Mobile Devices, Logitech Download Assistant
- Face Recognition and YOLOv5 .NET uninstalled from CodeProject.AI *********
- Sub stream GOP corrected from 1 to 15 (critical — was causing abnormal sub stream bitrate)
Scheduled automatic BI service restart twice daily via Windows Task Scheduler:
- 03:00 and 15:00 every day
- Command: cmd.exe /c sc stop BlueIris && timeout 10 && sc start BlueIris
- This keeps the delay within acceptable range (~0 seconds) as long as BI has been running less than ~16 hours
Despite all resources being available (CPU, RAM, GPU, disk, network), BI progressively accumulates a buffer over ~16 hours that causes live view delay. The buffer resets completely with a service restart. No obvious bottleneck identified. Could be a memory leak in BI? A known bug with 4K RTSP streams?
Question for the community:
Has anyone experienced this exact pattern — progressive live view delay that accumulates over ~16 hours and resets with a service restart, with no resource saturation? Any insight into the root cause would be greatly appreciated!
Last edited: