Camera Calculation Quiz

JH
John Honovich
Oct 06, 2014
IPVM
GW
George Whittaker
Oct 07, 2014

looking at the multiple Aov questions on the test, and the Aov functionality of the camera picker, i get the impression that you favor working with Aov's rather than lens length. is that true? because i find it harder to deal with the angles.

JH
John Honovich
Oct 07, 2014
IPVM

George, insightful question / comment.

I used to use focal length in mms (e.g., 3mm, 6mm, 10mm, etc.) because that's what manufacturers typically provide. There's a number of reasons I think switching to angles is better:

  • Angles better express what physical area one is covering (90°, 60°, 15°, etc.)
  • Because of differences in imager size used, focal length alone can result in widths that are materially off (not huge, but notable).
  • Most importantly, with the trend to wider FoVs, the same lens length (e.g., 2mm) can mean 115°, 100° or 85°, etc. depending on the manufacturer and how its implemented. As such, using angles is more accurate.
GW
George Whittaker
Oct 07, 2014

all your points are valid. but they may have a slightly academic viewpoint to them.

because manufacturers identify camera lenses using focal length, the values have become familiar to those buying and installing them, and though they may not be as accurate, they are helpful.

also angular values are non-intuitive in certain cases, for instance given an AOV range of 100 to 14, how many times zoom is that? with mm its a trivial calculation.

also the slider that goes from high zoom on the left to low zoom on the right takes some adjustment. since for a given imager the angles map 1 to 1 to a focal length, why not put the mm values underneath the slider and angles on the top of the slider and have them change together. and reverse the order with low zoom on left. for the best of both worlds.

just one man's opinion

JH
John Honovich
Oct 07, 2014
IPVM

Hi George,

We do plan to display mms in addition to focal length in an upcoming improvement.

I don't see AoV to be academic because it is the actual angle / coverage area of a camera. Lens length is a mechanical metric that needs to be converted to coverage.

Btw, the answer to the zoom of "an AOV range of 100 to 14" is 100/14 or 7.14x. This has the added benefit that it is of the actual zoom range.

GW
George Whittaker
Oct 08, 2014

"AOV range of 100 to 14 is 7.14x"

shows you what i know!

to try and keep it simple i just went to the calculator after a refresh and noted what it said:

2.0 mm - 100.4°

then i added a 0 to make 20 mm, then it said:

20 mm - 13.7°

i rounded the numbers because i didn't want to make it seem crazy hard.

i figured that 2mm to 20mm was 10x. what did i do wrong?

JH
John Honovich
Oct 08, 2014
IPVM

On the 2mm, there's a rounding error in the calculator that we are pushing out a fix for tonight.

GW
George Whittaker
Oct 08, 2014

I don't think that fix affected this scenario, anyway 2.0 was already pretty decently rounded.

it looks like you are using something like AOV = 2 * arctan (sensor size / focal length * 2).

is that what is you intended, even when sensor size > focal length?

another reason to ditch focal length for manufacturers stated aov?

GW
George Whittaker
Oct 13, 2014

if i'm hearing you right then will a 2-8mm actually perform like a 3x (not a 4x) since the aov is the real measure?

2mm = 100 deg

8mm = 33 deg

do lens makers base their x claims on aov ratio?

JH
John Honovich
Oct 13, 2014
IPVM

Lens / camera manufacturers make their claims based on focal length ratio, not AoV.

For instance, check this Nikon lens - It has a 16.6 zoom ratio (based on a 18 - 300mm varifocal), but the AoV range is only 14.3 (based on 76° - 5.3°)

And, in surveillance, look at the Axis Q1675 - It claims an 18x zoom ratio (based on a 4.7mm to 84.6 varifocal), but the AoV range is only 14.75 (based on 59° to 4°).

From a practical performance perspective, clearly the AoV range / ratio is more important because that is the actual / visible difference.

GW
George Whittaker
Oct 13, 2014

one more quick question:

how does the calculator know what the aspect ratio of the sensor is, 4:3 or 16:9 ?

or doesn't that matter?

JH
John Honovich
Oct 13, 2014
IPVM

Generally manufacturers don't disclose the aspect ratio of the sensor they are using.

We list the aspect ratio of the max resolution, e.g., if it is 1.3MP, it is a 4:3, if it is 720p, it is 16:9, etc.

GW
George Whittaker
Oct 14, 2014

given what you just said, is it correct that the calculator gives the same PPF in both these examples?

1/3" sensor and 10mm lens at a dist of 100ft with max res. of 1.3MP = 26.7 PPF

1/3" sensor and 10mm lens at a dist of 100ft with max res. of 720p = 26.7 PPF

JH
John Honovich
Oct 14, 2014
IPVM

Yes, because PPF is calculated based on the horizontal pixel count. Both 1.3MP and 720P have the same horizontal pixel count - 1280.

See The Definitive Guide to PPF.

GW
George Whittaker
Oct 14, 2014

stop my train of thinking when it jumps the track

they do have the same horizontal pixel count, 1280

they do have the same diagonal imager size, 6.0mm, 1/3"

they don't have the same horizontal imager size, (16:9 vs 4:3)

a wider horizontal imager size results in a wider horizontal FOV

a wider horizontal FOV will have a lower horizontal PPF

PPF Guide helped alot!!!

JH
John Honovich
Oct 15, 2014
IPVM

They do have the same horizontal FoV. What differs is the vertical FoV only.

See: Aspect Ratio 16:9 vs 4:3 Shootout

(1)
GW
George Whittaker
Oct 15, 2014

now i'm confused. why do lens manufacturers show different hfov angles then?

also in that aspect ratio shootout you say that most of the so-called "HD" sensors also have a higher (vertical) resolution 4:3 mode, therefore 1280x720 would not be their max. res., as I stated. following your 'Beware' advice, i'm wondering if because the article is > 2 years old, maybe native 1/3" (6.0mm) 16:9 sensors were not common back then, but now things have changed?

here is a picture of a camera i own, its an everfocus 3220 HD. its highest stated resolution is 1920x1080. it claims a 1/3" sensor. this is a picture i took of it.

what do you make of that?

JH
John Honovich
Oct 15, 2014
IPVM

Different HFoV angles are shown because the angles are a factor of two things: lens length and imager size. Notice that the multiple HFoV angles are listed under different imager sizes. The bigger the imager, the wider the FoV, everything else equal (and that pattern is shown on your excerpt).

On the other point, some sensors are HD only. However, even with the new 4K cameras, most are supporting 12MP max resolution. Why? Same thing. 4K / 8.3MP is 16:9. 12MP is the full 4:3 with a taller FoV.

GW
George Whittaker
Oct 15, 2014

"Notice that the multiple HFoV angles are listed under different imager sizes"

yes, but notice that multiple HFoV angles are also listed under the SAME imager size and focal length, in this case:

1/3" WIDE HFoV 100°43' for 4:3

1/3" WIDE HFoV 111°8' for 16:9

why do they do it?

JH
John Honovich
Oct 15, 2014
IPVM

That I don't know. Have not seen that before.

Avatar
Ethan Ace
Oct 15, 2014

I have. Panasonic lists this on some of their new cameras, too. It seems to come into play when camera manufacturers use a 16:9 sensor, but crop it to 4:3 at certain resolutions. So since they're cropping the sensor horizontally, but maintaining the same vertical pixel count, the effective width of the sensor is smaller, changing the HFoV.

Generally, things work the other way, with a 4:3 sensor cropped to 16:9 (3MP cropped to 1080p for example), which maintains the same sensor width and doesn't introduce this issue.

Avatar
Ethan Ace
Oct 15, 2014

Here's an example. Different HFOV on the 1080p model when using 4:3, same on 720p:

GW
George Whittaker
Oct 16, 2014

thx. interesting.

"So since they're cropping the sensor horizontally, but maintaining the same vertical pixel count, the effective width of the sensor is smaller, changing the HFoV."

almost the same vertical pixel count. it looks to be expanding a little in the vertical direction as it crops the horizontal, no? from 16.0 to 17.8 Tele

SU
Stanislav Utochkin
Oct 19, 2014

It seems WV-SFV631L has sensor with near to 16:10 aspect ratio. To get 16:9 it is cropped in vertical, while in horizontal full width is used. To get 4:3 it is cropped in horizontal. It follows from the size 5,421*3,405mm and angles.

JH
John Honovich
Oct 19, 2014
IPVM

Stanislav, good find.

Related, Panasonic lists the effective pixels of the WV-SFV631L as being "approx. 2.4 megapixel". That's an atypical size / count and appears related to the cropping you reference.

SU
Stanislav Utochkin
Oct 19, 2014

About YV2.8x2.8SR4A-2:

I suppose, Fujinon got the 1/3" 16:9 sensor not by cropping in vertical the "standard" 1/3" 4:3 sensor (4.8*3.6mm).

They used "non standard" 1/3" 16:9 sensor with the diagonal size equals to the diagonal of "standard" 1/3" 4:3 sensor (6mm). So the sizes of 16:9 sensor should be 5.2*2.9mm. It corresponds with TELE angles (f=7.6).

WIDE angles don't correspond because of strong lens distortion in WIDE mode. Here is VideoCAD model of this lens in WIDE mode ( 1/3" f=2.8mm 4:3 AH=100 deg AV=73 deg)

GW
George Whittaker
Oct 19, 2014

as for non-standard 16:9 "1/3 sensor, the critical "standard" is just the 1/3" part, not so much the aspect ratio. the lens simply provides a circle with a 6.0mm diameter which will cover the imager's 6.0mm diagonal of any aspect ratio.

in this case it looks like fuji is providing AOVs for both 16:9 and 4:3 at their largest possible size, perhaps because native HD sensors, like the one from the everfocus camera shown above, are becoming more popular.

and if people only want 16:9 images, then they are better off with a native sensor because the size of the pixels themselves are bigger. and bigger pixels equals more light capture.

SU
Stanislav Utochkin
Oct 19, 2014

Thus there is a confusion with "standard" because of presence of two different aspect ratios: image sensor's and image's one. For camera's view area calculation only image's aspect ratio is important. For exact calculation we need to know real sizes of the active sensor's area. When we work with different aspect ratios, use of sensor format in inches without understanding can lead to error. This error is neglected by the most of lens calculators.

JH
John Honovich
Oct 19, 2014
IPVM

"we need to know real sizes of the active sensor's area."

But most manufacturers do not specify the sensor's area beyond the generic inch statement, right? You will always see something like 1/4", 1/3", 1/2.7" etc. but Panasonic actual listing of the dimensions of the scanning area (e.g., 5.436 mm (H) x 3.336 mm (V)) is rare, no?

SU
Stanislav Utochkin
Oct 19, 2014

Unfortunately manufacturers often give not enough
info for modeling. Therefore we can choose one of three ways:

1.Accept bad accuracy and use reserve. Actually hight accuracy has no sense in CCTV practice because there are a lot of factors we can't take into account.

2. Choose right manufacturers with needed info.

3. Measure camera parameters independently.

GW
George Whittaker
Oct 19, 2014

no, i think there is a fourth way that is better.

that is, as long as you are using the sensor at its true highest resolution, you then just use the aspect ratio of the resolution with the known sensor diagonal to calculate sensor width.

this works in all cases i know of except when a camera intentionally doesnt use all pixels in its highest res, which seems rare.

SU
Stanislav Utochkin
Oct 19, 2014

The fourth way is "calculate needed info by analysing info provided by manufacturers". This way requires understanding and a little mathematic knowledge. It contradicts with basic principle of popular CCTV tools - it must be as simple as possible:)

GW
George Whittaker
Oct 19, 2014

when i said "you just use/calculate" i didn't mean "you personally" of course!

in John's calculator at least, the resolution and image format has already been defined so no new information is necessary. and the calculator does the calculating.

for absolute backward compatibility one could have a "Native resolution?" checkbox i suppose, but using non-native resolutions will often make wrong angles anyway.

regardless, treating all fractional inch formats as being 4:3 aspect ratio seems unwise going forward, as they are not the most efficient use of silicon real-estate when the target displays are HD, and so one would expect the number of display optimized (non 4:3 ratio) imagers to increase.

Avatar
Max Shumeyko
Oct 16, 2014

Interesing quiz.

But I would say it is not easy to detemine the pixel density by watching a photo.

JH
John Honovich
Oct 16, 2014
IPVM

This is a quiz about using the IPVM camera calculator to do calculations. The photo was pulled directly from the calculator. One can compare to the calculator for the answer.

New discussion

Ask questions and get answers to your physical security questions from IPVM team members and fellow subscribers.

Newest discussions