G1WH viewing angle?

[Snowslip]

I did a REAL viewing angle measurement test:
http://videobam.com/VXDRM
or

Initial calculations or what it SHOULD have been (all references are linked to http://www.carbidedepot.com/formulas-trigright.asp):
b=30cm, A=70 degrees => a=82.4cm
What was measured:
a=31cm, b=30cm (as per initial setup) => A=45.9 degrees.
The whole viewing angle = A*2=45.9*2=91.8 degrees. Or, if we factor in a measurement error, it looks like the actual viewing angle is 90 degrees.

Any comments?

Then, someone on another forum suggested that the specs call for a "diagonal angle".
From the way it is printed on the box, it is clear to me that it is supposed to be horisontal angle.
Anyway, I've done the estimations for the "diagonal angle".
Given:
- resolution: 1080H x 720V
- distance from the camera: 30 cm
- horisontal field of view @ 30cm: 62 cm

Calculations:
In order to deal with right-angle triangulars, I split the image into four equal quadrants. So, each quadrant has 540H x 360V resolution. From my previous measurements, 540 pixels correspond to 31 cm, thus 360 pix=31cm/540*360=20.7 cm.
Using http://www.carbidedepot.com/formulas-trigright.asp , let's find "c" for "a"=31 and "b"=20.7; c=37.3cm.
Then, let's find A for a=30 and b=37.3; A=51.2 degrees. Thus, the "diagonal" angle is ~100 degrees.
Therefore, even the "diagonal" angle is far off the specs: ~100 degrees vs. 140 degrees claimed, i.e. they claimed 40% wider angle.
Note that in my estimations I assumed that the lenses are ideally linear (no barrel distortion), which, as we already know, is NOT the case. So, if I run another real test, the results will be much worse.

 

[jokiin]

Measurements offered are diagonal viewing angles, what's printed on the box of these generic products often doesn't match what gets used though

 

[Snowslip]

Actually, "diagonal viewing angle" reminds me of "music power" or "peak power" that many manufacturers of amplifiers, CD-players, etc. print on the big labels on the boxes.
It is so funny to see a huge "300W" label on a boom-box with 10W power consumption rating printed in small letters at the back...

 

[Snowslip]

Measurements offered are diagonal viewing angles, what's printed on the box of these generic products often doesn't match what gets used though

SO, what are the REAL specs for this model?

 

[jokiin]

Not really, the diagonal measurement is a result of horizontal and vertical figures, it's not just a made up number, that doesn't mean what might be printed on a box is accurate though unfortunately

 

[winterseed]

I've left some firm comments on one or two Techmoan videos about this, it's a pet peeve of mine. I figure they advertise the specs of the LENS, regardless of how it interacts with the sensor. I don't quite understand why more people aren't bothered by it. 140 degrees sounds great but it doesn't correspond to the reality of capturing only about half of the scene in front of the camera.

Without direct side by side comparison, I haven't been convinced of any difference between the AOV of the G1WH and the G1W/-C. Is it possible that the new lens doesn't fit the sensor and as a result doesn't provide a significantly wider AOV?

 

[jokiin]

Possible the lens used is not correct for the sensor, you're talking about products built to a price rather than a standard, things aren't always ideal

 

[Snowslip]

I've left some firm comments on one or two Techmoan videos about this, it's a pet peeve of mine. I figure they advertise the specs of the LENS, regardless of how it interacts with the sensor. I don't quite understand why more people aren't bothered by it. 140 degrees sounds great but it doesn't correspond to the reality of capturing only about half of the scene in front of the camera.

Without direct side by side comparison, I haven't been convinced of any difference between the AOV of the G1WH and the G1W/-C. Is it possible that the new lens doesn't fit the sensor and as a result doesn't provide a significantly wider AOV?

I don't see what sense it makes to advertise the specs of one component of the system, and claim them as the system specs. E.g. I take the max operating temperature rating for a dummy 7805, which is 150C, and claim that the whole device can safely operate at 140C. And what about the electrolytic caps that are rated at 85C? Not to mention the Li-ION battery, sensor, etc...

Actually, as my test proofs, the actual AOV captures only 38% of what the claimed AOV is supposed to yeild (62 cm vs. 164cm of horizontal view)! That's insane.
It's like overrating the product by a factor of three. Sorry, it's called SCAM!

 

[Snowslip]

Possible the lens used is not correct for the sensor, you're talking about products built to a price rather than a standard, things aren't always ideal

The manufacturer MUST NOT LIE to the buyers.

 

[Snowslip]

Not really, the diagonal measurement is a result of horizontal and vertical figures, it's not just a made up number, that doesn't mean what might be printed on a box is accurate though unfortunately

I do understand how your "diagonal angle" is calculated, and I even re-calculated my results with regard to it.
What I do NOT understand is what sense it makes to advertise something that is not intuitively understandable by the end users. For human beings is natural to perceive the world in either horisontal or vertical plane. Even our eyes have higher resolution in these planes. (BTW, that what the JPEG compression is based on).
Lets' run a survey and ask the forum members what they picture in mind when they see "140 degrees angle of view" parameter.
And I repeat my previous statement: advertising the "diagonal angle" value without clearly stating so is misleading.

 

[winterseed]

For cheap models especially, sellers would probably claim that they don't know the exact AOV and are simply providing the available specs. I consider it to be more of a 'buyer beware' situation for now. Reviewers should list "measured horizontal AOV" alongside "advertised/claimed AOV" and link to how the measurement was taken.

Much of the review process is already cutting through marketing BS, I would very much like to see AOV/FOV added to the list. To watch a review video from an otherwise reputable reviewer, only to have them rattle off the advertised lens spec like it was the word of God and then go on to tout the benefits of a wide FOV even as I'm watching sample clips which are clearly less than 90 degrees wide really troubles me.

I think barrel distortion (the "fish eye look") tricks people into thinking the view is actually wide.

 

[jokiin]

figures are provided by the lens manufacturers, the results can vary even if the same lens is used with different sensors, look at number 8 under spec of this lens as an example of differing diagonal viewing angle depending on the horizontal and vertical figures http://alaudoptical.en.alibaba.com/...Lens_Security_Camera_Lens_M12_Board_Lens.html

 

[winterseed]

That makes perfect sense. The problem, as Snowslip said, is that that figure means nothing to the end user in a fixed configuration product like this. It's particularly useless in the case of the G1WH because the wider lens spec is a deciding factor in peoples' purchasing decision; I've seen at least a few people choose the G1WH over the G1W-C because of the wider lens.

I tried to illustrate the difference between 140 degrees (red) and 90 degrees (green) in the attached images. This is just to compare what people might think a DVR boasting a "140 degree wide angle lens" will capture with something more realistic. Personally I think the 'sweet spot' is somewhere between 100 and 120 degrees.

I feel like I should point out that the G1WH doesn't have a bad field of view; most products have misleading descriptions and 90 degrees seems to be about average. The new lens simply isn't the big improvement the number suggests.

 

[Snowslip]

figures are provided by the lens manufacturers, the results can vary even if the same lens is used with different sensors, look at number 8 under spec of this lens as an example of differing diagonal viewing angle depending on the horizontal and vertical figures http://alaudoptical.en.alibaba.com/...Lens_Security_Camera_Lens_M12_Board_Lens.html

Good point. From what I can see on that page, the lenses are supposed to yield 142 degrees "diagonal" angle for 1/2.7" sensors and ~125 degrees for 1/3" ones. That sort of makes sense.
Then, what is the difference between G1W and G1WH? Lenses? What are the sensor sizes in G1W and G1WH?

 

[jokiin]

Sensor sizes are the same I believe, they may use different lenses although the real world results may not be much different even if the listed specification is

 

[Snowslip]

That makes perfect sense. The problem, as Snowslip said, is that that figure means nothing to the end user in a fixed configuration product like this. It's particularly useless in the case of the G1WH because the wider lens spec is a deciding factor in peoples' purchasing decision; I've seen at least a few people choose the G1WH over the G1W-C because of the wider lens.
I tried to illustrate the difference between 140 degrees (red) and 90 degrees (green) in the attached images. This is just to compare what people might think a DVR boasting a "140 degree wide angle lens" will capture with something more realistic. Personally I think the 'sweet spot' is somewhere between 100 and 120 degrees.
I feel like I should point out that the G1WH doesn't have a bad field of view; most products have misleading descriptions and 90 degrees seems to be about average. The new lens simply isn't the big improvement the number suggests.

Nice and visual pictures. However, note that those 25 degrees on either side "eat" about 2/3 of the field of vision (linear)! Going back to my experiment: at 30 cm from the camera, it was able to capture two objects 62 cm apart (31 cm on either side from the center line). The expected distance between two objects that corresponds 140 degrees angle of vision is 164 cm. 62/164=37.8% ONLY!

But since jokiin clarified that the horisontal angle of view is supposed to be ~120 degrees, the expected distance is 52*2=104, i.e. the coverage is 62/104=60%. Thus, we loose 40% of horisontal field of vision, comparing to what was advertised.

 

[Snowslip]

Sensor sizes are the same I believe, they may use different lenses although the real world results may not be much different even if the listed specification is

So, basically, these are pretty much the same cameras, right?
It looks like the manufacturer changed the lens for a "better" one and advertised its specs to boost the sales.
However, he did not check its compatibility with the sensor and make sure that the new lens are actually better than the old ones.

One technical thing is still not clear to me: what is the size of the sensor if the camera has even narrower angle of view comparing to already "oversized" 1/3" sensor's specs?

 

[mel]

http://datakam.com/download/Manuals/Aptina AR0330_DS_G.pdf


Optical format 1/3-inch (6.0 mm)
Entire Array: 6.09mm
Still Image: 5.63mm (4:3)
HD Image: 5.82mm (16:9)
Active pixels 2304(H) x 1536(V): (entire array):
5.07mm(H) x 3.38mm(V)
2048(H) x 1536(V) (4:3, still mode)
2304(H) x 1296(V) (16:9, sHD mode)
Pixel size 2.2um x 2.2um

 

[winterseed]

This compares 90, 120 and 140 degrees and is cropped to a perpendicular line so it fits with the linear measuring method.