Formula to set a varifocal T5442T-ZE to emulate other fixed lenses.

[Perimeter]

I read countless threads on this problem and how it doesn't work. But I need to know fields of view so I can check beforehand if such a fixed lens would work in a location. Most manufacturers list the horizontal field of view angle. So I set the T5442T-ZE up and measured.
I set it up exactly 4 meters in front of a wall and focused on the center. Then I zoomed in to max (2250) and marked the side of the screen on the wall. Then I did some trigonometry and came up with a viewing angle of 47°. Which is exactly what dahua says in the specs sheet.
Then I zoomed out until I reached a brick end within a brick row and noted down the zoom setting. This I repeated until I reached a setting of 0. Then I measured the distances from the center to the bricks. Then I applied trigonometry.
There are some things to note: The zoom setting is not 100% reproducible. You have some tolerance when you return to the same zoom setting later. And more noteworthy, my horizontal viewing angle maxed out at 107°, which is less than the value Dahua states (114°). I have double checked my measurements, considered margins of error, but there are no extra 7° to be found. But there are further observations. The viewing angle seems smallest at the horizontal center line (which I used). The horizontal field of view is larger everywhere else than where I measured. That is because of visible lens distortion. The viewing angle may also change with set focus distance. That is why I mention the 4m. Also, picture quality at the outer edge at max. wide angle is really low. It is very hard to tell exactly.
For verification, I took another lens (different sensor, 3.6mm) and predicted where it's field of view would end according to my table. That corresponded very well with the observable FOV.
So here is the graph:

For practical purposes, here is a quick to use and easy to remember formula that converts horizontal FOV angle (x) into a zoom setting (z) to use with the T5442T-ZE:
z = -0,007537*x³+2,272*x²-247,36*x+9685

What needs to be done:
Someone has to cross check if their T5442T-ZE corresponds to mine. That is not yet sure. (Feel free to send me one and gladly do the work for you!)
The discrepancy at the max. wide angle setting should resolved. If the values given in spec sheets are FOV horizontal for the top of the frame, then my data would be rather meaningless.

Edit: Zoom settings are used with IE11 interface.

 

[tigerwillow1]

I've been wondering about the repeatability of the zoom settings, which would require a stepper motor and zero slop in the mechanism, or a position sensor of some sort. My assumption is that the cameras have neither. My own measuring exercise was with a few different fixed focal models, and it validated their horizontal FOV specs. In my younger years I was a trig whiz. Now I use online calculators.

 

[Perimeter]

I've been wondering about the repeatability of the zoom settings...

I constantly checked these during the measurement and after. The camera returned to roughly the same field of view, perhaps with 1/2" precision at 4m distance to wall. I was aiming for the gaps between the bricks. I could get back to each gap by dialing in the zoom setting. Where did you measure FOV? Along a line through the center of the picture or a line along the frame of the picture?

 

[tigerwillow1]

Where did you measure FOV?

I set up a ~100 foot long field of stake and rope triangles in my yard, and don't remember if I measured at the center. I would hope so since that seems the most logical. Looking at 6mm 5442 images I think it might not make much difference, with the 3.6mm cameras, obvious difference. The two measurements I wrote down were:

HFW4231E-SE 6mm
Camera spec 52 degrees, actual measurement 51.56 degrees
T5442TM-AS 6mm
Camera spec 56 degrees, actual measurement 54.2 degrees

 

[Perimeter]

My measurement of 107° max horizontal field of view is in close agreement with this measurement of 105°.

Dahua T5442T-ZE in-depth review

Hi guys, Today I present you a review of Dahua T5442T-ZE. This model is equipped with 2.7-12mm zoom lens and an excellent 4 megapixel 1/1.8 inch CMOS sensor. Compared to 5442TM-AS-LED and 5442TM-AS models, which I’ve already reviewed, this one has twice as much RAM available. The housing is...

ipcamtalk.com

 

[CaptainCrunch]

Please excuse any perceived sarcasm in my next question. Are you trying to achieve anything other than running tests to get the exact numbers for that specific camera?

I recently purchased 3 T5442T-ZE cameras to cover my back yard. There is a 14 foot gap in the roofline in the middle of the back of the house. The plan is to put a camera at each middle corner facing out. The FOV will go from the house and sweep out to the yard. The 3rd camera will be put on the patio and sit back 14 feet from the others. In planning this, I originally measured angles to determine what fixed focus cameras would work best. Instead, I went with variable focus cameras. That way I can get just the right combination of overlapping fields of view. Even though I measured angles to get an idea of what I needed, I'm sure my margin of error was at least 2 degrees. And because I will have overlapping fields of view, a 2 degree difference will not be noticeable in my real world application.

I would be surprised if all 3 of the cameras got the exact same field of view. Each lens moves inside of the camera. If one lens can extend out a fraction of a mm more than another camera, the result will be a different field of view by a degree or two or maybe more. But I don't care because I do not plan to have one fov stop at the exact point another fov starts. I plan to have overlapping fov. Is there a reason the average user should view the numbers on the website as an absolutely accurate every time for every camera number rather than as a general guide for that model? Again, I don't mean any sarcasm with this post. I read through it when it was first posted but I still can't figure out what information a user should have gotten from it.

 

[Perimeter]

Are you trying to achieve anything other than running tests to get the exact numbers for that specific camera?

I need to know fields of view so I can check beforehand purchase if such a fixed lens would work in a location.

I would be surprised if all 3 of the cameras got the exact same field of view.

The quoted reference agrees with mine with 0° difference zoomed in and with 2° difference at maximum wide angle. We would probably be even closer if we would have had the same distance to cover, focus distance changes the fov quite a bit.

 

[CaptainCrunch]

I posted my question after reading this thread and the linked thread. The only take away I can get is that someone should be cautious if they plan their camera deployment down to the degree with no overlap of fields of view and/or surroundings. I guess that's a thing although I don't believe it would be best practice. I'm just struggling to apply the information here to the real world.

And in regards to saying "I would be surprised if all 3 of the cameras got the exact same field of view," I would bet money they are each off by a degree or two due to any number of factors. But, regardless of any differences, 3 cameras should comfortably cover 180 degrees. Even going with the numbers in the linked thread, I will be good. Zoomed out all the way resulting in a 105 degree fov and planning for a 5 degree overlap with the wall, I'll have over 95 degrees of practical fov from each corner camera. That means if the middle camera is zoomed in all the way, I'll have 10 degrees of overlapping fields of view. So the few degree differences in each camera do not change anything. I could have achieved the same thing using the fixed focus cameras. Allowing for a 25% deviation from the number on the site would still allow for almost 10 degrees of overlap.