Deciding how to get started

Smilingreen

Known around here
Joined
Sep 17, 2021
Messages
1,459
Reaction score
5,842
Location
USA
Concerning at Andy's Amazon store the IPC-T2431T-AS 2.8mm camera:
I am thinking of getting this camera to start getting experience with setting up a camera with Blue Iris.
Is it an ONVIF compatible camera, and fully compatible with Blue Iris?

If this camera is wall mounted outside like with the PFB203W can it be exposed to rain?
It is ONVIF compatible and it has a IP67 rating, so can be exposed to the elements.
 

Herondas

n3wb
Joined
Apr 26, 2022
Messages
13
Reaction score
3
Location
US
Customer answers and questions took care of the ONVIF question.
 

TheWaterbug

Getting comfortable
Joined
Oct 20, 2017
Messages
386
Reaction score
646
Location
Palos Verdes
Basically AI cameras costs a little more because they have functionality to recognize human or vehicle and the non-AI cameras do not.
MHO, and YMMV, but in general you want the smarts in your computer, not in your devices. Your devices have been engineered to have tiny little brains, so they can be sold at the lowest possible price. This means performance and upgradeability will always be limited.

The brains in a PC can be tailored to your needs and upgraded or replaced as necessary.

Furthermore, if you buy N cameras with AI, you're paying for N brains. But the odds are that not all N brains will need to be active and computing at full tilt all at the same time. If you buy N cameras without AI and one computer, then you only need to pay for enough brains for the N streams that need simultaneous analysis. This is the same math that drives putting the smarts in the cloud, since now one giant brain can literally serve thousands of devices, whether those be cameras, robot vacuums, or whatever, but a lot of people (especially here!) don't trust the cloud, so we like to build our own NVRs out of commodity PCs.

Of course if you have 1 camera, the economics favor AI in the camera. But as the number of cameras grow, the economics quickly favor buying the brains in the PC.

The other math around here is that, if you think you will have 3-4 cameras, you will soon have 10-12.
 

wittaj

IPCT Contributor
Joined
Apr 28, 2019
Messages
11,790
Reaction score
20,980
Location
USA
MHO, and YMMV, but in general you want the smarts in your computer, not in your devices. Your devices have been engineered to have tiny little brains, so they can be sold at the lowest possible price. This means performance and upgradeability will always be limited.

The brains in a PC can be tailored to your needs and upgraded or replaced as necessary.

Furthermore, if you buy N cameras with AI, you're paying for N brains. But the odds are that not all N brains will need to be active and computing at full tilt all at the same time. If you buy N cameras without AI and one computer, then you only need to pay for enough brains for the N streams that need simultaneous analysis. This is the same math that drives putting the smarts in the cloud, since now one giant brain can literally serve thousands of devices, whether those be cameras, robot vacuums, or whatever, but a lot of people (especially here!) don't trust the cloud, so we like to build our own NVRs out of commodity PCs.

Of course if you have 1 camera, the economics favor AI in the camera. But as the number of cameras grow, the economics quickly favor buying the brains in the PC.

The other math around here is that, if you think you will have 3-4 cameras, you will soon have 10-12.
I generally agree, but type of camera, use case, and field of view need to be taken into consideration as well.

If someone has a lot of cameras and tries to use DeepStack on every one of them, the computer requirements can start to get extensive and/or delays in notifications.

Using the camera CPU for AI instead of BI motion detection then means that the BI computer CPU usage can drop a little and not spike trying to do AI. YMMV - probably more of an impact on an older unit than a newer unit.

Plus some series of cameras, like the Dahua 5442 series, only come with AI built-in, so there is not a savings to be had purchasing a non-AI 5442 camera.

Knock on wood, for my camera field of views, these AI check boxes are spot on in all of my cameras that have them! Has made scrubbing alerts in BI a breeze because even with how great motion detection is in BI, there are a few situations where I cannot knock out false triggers, especially at night with headlights bouncing off a hill for example. Trying to eliminate that and then I miss a real trigger. The camera AI doesn't even flinch at attempting to think the headlight bounce is a trigger, or motion lights turning on. Using DeepStack for those cameras would cause CPU/GPU spikes all night.

Again, a lot depends on the location and field of view and speed of objects at nighttime. Vehicles at night can be problematic for camera AI if the field of view is too tight because the camera needs to be able to identify the object, assess if it is a vehicle, and then trigger the camera, so there are instances where that might be an issue based on the specifics of what the camera is looking at. But DeepStack catches all of these as it is looking at images. I have tested this with my cameras as well.

And because I have a few "dumb" cameras without AI lol (so I have to use BI motion detection) and have some overlap with those and AI cams, I have been able to confirm every false and true trigger from those dumb cams were accurately triggered or not triggered in my camera with AI. To the point that I cannot see myself buying a new camera without that. YMMV. Just for redundancy, I still run a few cams with BI motion detection just in case an AI camera didn't pick something up. Plus I run 24/7 so I can always go back.

And even with camera AI, for a few of my cameras, I then supplement that camera with DeepStack in BI.

From my own personal experience - the true test....I have found the AI of the Dahua cameras to work even in a freakin blizzard....imagine how much the CPU would be maxing out sending all the snow pictures for analysis to Deepstack LOL. It is night rain or snow or insects with infrared that can max out a CPU doing DeepStack analysis on all the cameras.

My non-AI cams in BI were triggering all night. This event was also being ran through Deepstack and it failed to recognize a person in the picture, but the camera AI did in my 5442. The only triggers my AI cameras have are from human or car triggers and is doing so with a lot less CPU than sending pics to Deepstack. This pic says it all and the video had the red box over it even in complete white out on the screen:


1613268961041.png


I am using using DeepStack in concert with my AI cams for a few situations, and while some of that third party stuff is cool like tagging was it a dog or a bear, I don't need all that fancy stuff (now for LPR, yes I am using the tools created in those threads). If my camera triggers BI to tag an alert for human or vehicle and BI can accomplish what I need by way of a text or email or whatever, that is sufficient for my needs.

As always, YMMV.
 

Aronder

Young grasshopper
Joined
May 8, 2022
Messages
55
Reaction score
11
Location
United States
Thank you both for helping to jump start me.

In the camera in Settings / Network / Access Platform I see that P2P is enabled, and says:

"After enabling the function and connecting Internet, we will collect device information such as IP address, MAC address, name and serial number. The collected information is only used for remote access of the device. If you do not agree to enable the function, please cancel the selection of check box."

As it is enabled does this expose the camera through my router to the WAN? I doubt it but I don't have experience with P2P and thought to ask.
 

wittaj

IPCT Contributor
Joined
Apr 28, 2019
Messages
11,790
Reaction score
20,980
Location
USA
Yes, it will expose your camera if you have it connected thru your router.

Best practice is to turn P2P and UPnP off and put the cameras on a separate IP address subnet from your internet.

Most do so by using a dual NIC or VLAN.
 
Top