- Nov 3, 2014
- 5
- 11
I am very grateful to the help i have received in this forum, especially from fenderman, I would like to contribute my experience with my fellow enthusiasts on this great forum on a larger setup of Blue Iris and IP Cams. You may find some info curious, such as very low resolution VGA IP Cams and a mix of different brands of cams (experimenting). It is because the setup is gradually being upgraded over time. First, the specifics.
36 IP Cams:
7 Foscam FI8910E VGA Cams (640x480, ALL@15FPS)
23 Foscam FI9805E 960P Cams (1280x960, 18@30FPS, 5@15FPS)
5 Hikvision DS-2CD2432F-IW Cams (1920x1080, All@30FPS)
1 Hikvision DS-2CD2232-I5 Cam (1920x1080, 30 FPS)
TIMELINE 1
Blue IRIS 3.x, 32 Bit. No Direct to Disk. No Optimization applied (limit refresh, DTD, reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs.
Note: System unable to handle the load. Crashes as RAM Usage approaches 3GB+. CPU pegged at 100% and unresponsive. Recording is choppy, pixelating and nearly unusable.
TIMELINE 2
Blue IRIS 3.x, 32 Bit. Optimization applied (limit refresh, SOME DTD (direct to disk), reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs. Win7
Note: System still unable to handle the load. All problems persists.
TIMELINE 3
Blue IRIS 4.x, 64 Bit. Optimization applied (limit refresh, SOME DTD, reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs. Win7
Note: System no longer crashes, RAM usage now exceeds 5GB. Quality issue remains, CPU pegged at 100%.
TIMELINE 4
Blue IRIS 4.x, 64 Bit. NO Optimization applied (limit refresh, DTD, reduce frame rate, etc).
Intel i7-5930K (Hexa Core, 3.5Ghz, Overclocked to 3.8Ghz), 16GB DDR4 RAM, 256GB System SSD, 1TB SSD for CURRENT CLIPS, 4TB 7.2KRPM HDD for STORED and ARCHIVE Clips.
Note: System stable, RAM usage exceeds 5GB. CPU at 70% when NOT recording. Triggering recording of 2 x DS-2CD2432F-IW will peg CPU to ~ 98%, presumably due to transcoding. Trigger multiple Cams causes CPU pegged at 100%. Most video recording qualities are fluid.
TIMELINE 5
Blue IRIS 4.x, 64 Bit. Optimization applied (limit screen refresh 15FPS, ALL DTD, but KEPT frame rates high as noted).
Intel i7-5930K (Hexa Core, 3.5Ghz, Overclocked to 3.8Ghz), 16GB DDR4 RAM, 256GB System SSD, 1TB SSD for CURRENT CLIPS, 4TB 7.2KRPM HDD for STORED and ARCHIVE Clips.
Note: System stable, RAM usage exceeds 5GB. CPU at 50-60% when NOT recording. Triggering recording of 2 x DS-2CD2432F-IW NO LONGER affects CPU significantly presumably due to NO transcoding. Triggering multiple recordings (5 or 6 HD cams) also has little effect on CPU. All video recording qualities are fluid and no visible loss of frames. FINALLY REACHED A STATE OF A GOOD SETUP.
CONCLUSION
1. Frame Rate Matters. At night, many cams auto reduce frame rates to 15FPS. CPU (I7-5930K) drops from 60% to 30%.
2. Screen Refresh Matters. It accounts for about 10% of CPU (I7-5930K) usage on my setup.
3. Megapixel matters. A 3MP cam triggering recording and transcoding will tax CPU signficantly. VGA cams, not so much.
4. PC Spec. Really Matters. In order to drive that many megapixels, it's important to get a beefy system. As Fenderman repeatedly reminds us, not all i7 are created equal. With this setup, inbound network traffic is about 12MB/S (about 100Mb/s) making gigabit ethernet a must. But SSD for CURRENT CLIP may be an overkill as writing 12MB/S (continuous recording, if needed) to HDD is within the comfort zone of normal magnetic HDDs.
5. DTD (Direct to Disk) Really Matters. Even if you have a stable setup with CPU at 50-60% usage WHEN NOT recording, triggering multiple HD cam transcoding will kill your high end CPU and start to cause dropped frames, pixelations, and a generally unresponsive system. Enabling DTD on as many CAMs as one can (if not on all) allows the video stream to be directly saved to disk without transcoding (CPU intensive work). One can research this site on the many good suggestions and caveats of using DTD (no time/weather video overlay, rotation of images, etc).
FINAL WORD:
I am by no means an expert by any sense of the word. I am just an enthusiast who likes to tinker. It is quite likely that my observations and conclusions above are erroneous or inaccurate. But I felt compelled to contribute to this site due to the help I previously received. It is my hope these notes will help others in some way. If you find this useful, please contribute to this thread so i know that there are interests in me providing future updates. Thank you.
36 IP Cams:
7 Foscam FI8910E VGA Cams (640x480, ALL@15FPS)
23 Foscam FI9805E 960P Cams (1280x960, 18@30FPS, 5@15FPS)
5 Hikvision DS-2CD2432F-IW Cams (1920x1080, All@30FPS)
1 Hikvision DS-2CD2232-I5 Cam (1920x1080, 30 FPS)
TIMELINE 1
Blue IRIS 3.x, 32 Bit. No Direct to Disk. No Optimization applied (limit refresh, DTD, reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs.
Note: System unable to handle the load. Crashes as RAM Usage approaches 3GB+. CPU pegged at 100% and unresponsive. Recording is choppy, pixelating and nearly unusable.
TIMELINE 2
Blue IRIS 3.x, 32 Bit. Optimization applied (limit refresh, SOME DTD (direct to disk), reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs. Win7
Note: System still unable to handle the load. All problems persists.
TIMELINE 3
Blue IRIS 4.x, 64 Bit. Optimization applied (limit refresh, SOME DTD, reduce frame rate, etc).
Intel i7-2700, 12GB, 3TB x 3 HDDs. Win7
Note: System no longer crashes, RAM usage now exceeds 5GB. Quality issue remains, CPU pegged at 100%.
TIMELINE 4
Blue IRIS 4.x, 64 Bit. NO Optimization applied (limit refresh, DTD, reduce frame rate, etc).
Intel i7-5930K (Hexa Core, 3.5Ghz, Overclocked to 3.8Ghz), 16GB DDR4 RAM, 256GB System SSD, 1TB SSD for CURRENT CLIPS, 4TB 7.2KRPM HDD for STORED and ARCHIVE Clips.
Note: System stable, RAM usage exceeds 5GB. CPU at 70% when NOT recording. Triggering recording of 2 x DS-2CD2432F-IW will peg CPU to ~ 98%, presumably due to transcoding. Trigger multiple Cams causes CPU pegged at 100%. Most video recording qualities are fluid.
TIMELINE 5
Blue IRIS 4.x, 64 Bit. Optimization applied (limit screen refresh 15FPS, ALL DTD, but KEPT frame rates high as noted).
Intel i7-5930K (Hexa Core, 3.5Ghz, Overclocked to 3.8Ghz), 16GB DDR4 RAM, 256GB System SSD, 1TB SSD for CURRENT CLIPS, 4TB 7.2KRPM HDD for STORED and ARCHIVE Clips.
Note: System stable, RAM usage exceeds 5GB. CPU at 50-60% when NOT recording. Triggering recording of 2 x DS-2CD2432F-IW NO LONGER affects CPU significantly presumably due to NO transcoding. Triggering multiple recordings (5 or 6 HD cams) also has little effect on CPU. All video recording qualities are fluid and no visible loss of frames. FINALLY REACHED A STATE OF A GOOD SETUP.
CONCLUSION
1. Frame Rate Matters. At night, many cams auto reduce frame rates to 15FPS. CPU (I7-5930K) drops from 60% to 30%.
2. Screen Refresh Matters. It accounts for about 10% of CPU (I7-5930K) usage on my setup.
3. Megapixel matters. A 3MP cam triggering recording and transcoding will tax CPU signficantly. VGA cams, not so much.
4. PC Spec. Really Matters. In order to drive that many megapixels, it's important to get a beefy system. As Fenderman repeatedly reminds us, not all i7 are created equal. With this setup, inbound network traffic is about 12MB/S (about 100Mb/s) making gigabit ethernet a must. But SSD for CURRENT CLIP may be an overkill as writing 12MB/S (continuous recording, if needed) to HDD is within the comfort zone of normal magnetic HDDs.
5. DTD (Direct to Disk) Really Matters. Even if you have a stable setup with CPU at 50-60% usage WHEN NOT recording, triggering multiple HD cam transcoding will kill your high end CPU and start to cause dropped frames, pixelations, and a generally unresponsive system. Enabling DTD on as many CAMs as one can (if not on all) allows the video stream to be directly saved to disk without transcoding (CPU intensive work). One can research this site on the many good suggestions and caveats of using DTD (no time/weather video overlay, rotation of images, etc).
FINAL WORD:
I am by no means an expert by any sense of the word. I am just an enthusiast who likes to tinker. It is quite likely that my observations and conclusions above are erroneous or inaccurate. But I felt compelled to contribute to this site due to the help I previously received. It is my hope these notes will help others in some way. If you find this useful, please contribute to this thread so i know that there are interests in me providing future updates. Thank you.