Camera light sensitivity specifications

[Fiona 0]

Stanislav, I read both articles including the Nikolai Uvarov article.

 

He has taken issue with the impropriety of minimum illumination (lux) claims.

 

He is saying that the published low lux specifications in most cameras are deceptive and he has demonstrated this in the case of the Watec 1/2" CCD by proving that there are insufficient photoelectrons to produce any image. (One photoelectron.)

 

Question: Do Nikolai Uvarov's remarks about a practical minimum illumination of 0.15 luxes at S/N 24 dB relate only to 1/2" CCDs? In other words, exponentially, 0.15 luxes would be insufficient for 1/3" CCD?

[Stanislav 0]

Question: Do Nikolai Uvarov's remarks about a practical minimum illumination of 0.15 luxes at S/N 24 dB relate only to 1/2" CCDs? In other words, exponentially, 0.15 luxes would be insufficient for 1/3" CCD?

 

Yes, 1/3" CCD has smaller pixels therefore it collects less number of photons, therefore its sensitivity must be less.

However there are other factors:

1.Smaller pixels can have less "noise that is inherent in the CCD ", less number of noise electrons (it is my supposition). Actually the sensitivity is limited by the number of noise electrons and the number of active photons.

2.All calculation in this article made for "source of white light with uniform energy distribution ". For a source with maximum in IR range ( tungsten halogen lamp ) the sensitivity will be higher, as the greater part of "valid" photons will be out of visible range.

3.Sensitivity for 17dB SNR will be higher than for 24dB SNR.

 

In my own tests I got the practical results for 1/3" ССD and a tungsten halogen lamp:

1. Sensitivity of black-white cameras with high resolution CCD image sensors (752x582 effective pixels) 1.3" IT CCD, Sony Super HAD CCD is 0.06-0.15lx at signal/noise ratio =17dB, lens aperture F1.2.

2. Sensitivity of black-white cameras with high resolution image sensor (752x582 effective pixels) 1.3" Sony ExView HAD CCD is 0.04lx at signal/noise ratio =17dB, lens aperture F1.2.

 

Please see other results of my tests here

[Soundy 1]

I agree with Soundy.

Stop it, you're scaring me.

[Soundy 1]

All professional tests are performed on stands, not in outdoor field. All camera parameters can be measured on laboratory stand.

This is the problem, though: low-grade manufacturers often DON'T BOTHER TESTING. They DO use the on-paper, theoretical specs... or publish specs "inspired by" other similar cameras. IR ranges are calculated based simply on the number of LEDs they can cram in, with no accounting for how well they're aimed, how even the output is of the individual LEDs, how steady the drive current is (20%-tolerance current-limiting resistors can cause a lot of variance), and so on.

 

Manufacturers fudging the numbers IS an important consideration as well: a "minimum lux" number may be the lowest level at which an image is detected, rather than the lowest level at which you get a USABLE image, for example.

 

You can make a lot better profit margin when you DON'T do the testing for a camera you're trying to sell in part of a $200 package.

[Stanislav 0]

All professional tests are performed on stands, not in outdoor field. All camera parameters can be measured on laboratory stand.

This is the problem, though: low-grade manufacturers often DON'T BOTHER TESTING. They DO use the on-paper, theoretical specs... or publish specs "inspired by" other similar cameras. IR ranges are calculated based simply on the number of LEDs they can cram in, with no accounting for how well they're aimed, how even the output is of the individual LEDs, how steady the drive current is (20%-tolerance current-limiting resistors can cause a lot of variance), and so on.

 

I wrote the same above:

Right tests are accurate. But parameters in the specifications weren't obtained from real tests. Many manufacturers don't test their cameras! I have in my customers several camera manufacturers. They really don't know how to test cameras They don't need to test them. These parameters became more marketing than technic.

 

Manufacturers fudging the numbers IS an important consideration as well: a "minimum lux" number may be the lowest level at which an image is detected, rather than the lowest level at which you get a USABLE image, for example.

 

Therefore we need to fix all important conditions:

1. Signal/noise ratio

2. IRE

3. Lens aperture

4. Light source

5. Exposure time

6. ?Resolution

7. We should disable noise reduction when testing.

[J-Telectro 0]

All I can say is wow! What a fantastic forum! It would have taken me months to try to

locate all this information.

 

I was wondering if I should replace some Bosch LTC0355 cameras that I have. I had seen

cameras with better low light performance specs and thought that the LTC0355 might be

obsolete.

Maybe they are better than I thought.

 

From what I have read here:

 

1. Many camera manufacturers inflate their sensitivity values.

 

2. Since most (reputable) manufacturers are using the same CCD sensors their cameras should

have similar basic sensitivity as long as tricks aren't employed such as prolonged shutter

times or reduced resolution.

 

3. Specs don't mean much, you can only really tell by testing the camera.

 

Thanks!

[mike_va 0]

Bosch 0355 is not a bad camera (and inexpensive). It is limited in adjustments, like default shutter that Bosch gives you on their better cameras. At least they provide useful specs.

 

For a while I would put up everything against that. 0620 is better. Other ones that will beat it are CNB BBM-24F, Mintron 1/2" mono, some WATEC (but limited in other ways).

 

Also try some faster lenses f1.0. There is only so far these cameras can actually go, without lying or using sense up.

 

Also note that not all cameras will have the same sensitivity to IR.

[Fiona 0]

Stanislav, you have posted much information and I am still studying it.

 

To maintain the integrity of the arguments advanced by David Elberbaum and Nikolai Uvarov, could you post an objective measurement of any camera that fulfills the parameters that you, Elberbaum and Uvarov specify.

 

It is important to set the criteria in order to compare like for like and remove the subjectivity that prevails.

 

Question: Do you have a test sheet of a known or common camera model? Could you post such a sheet on this thread please?

[Stanislav 0]

To maintain the integrity of the arguments advanced by David Elberbaum and Nikolai Uvarov, could you post an objective measurement of any camera that fulfills the parameters that you, Elberbaum and Uvarov specify.

It is important to set the criteria in order to compare like for like and remove the subjectivity that prevails.

 

I developed real testing procedure. It is described step by step here. This procedure has been published in 2007 with VideoCAD6 and corrected in 2011.

Using this procedure I tested cameras and got these results.

 

These results fulfill the criteria:

1. Signal/noise ratio - 17dB

2. IRE - is measured while testing. It depends on max AGC gain of the camera on test. See Page 10 "Parameters, limiting image quality at defining sensitivity" for details.

3. Lens aperture - F1.2

4. Light source - tungsten halogen lamp with nominal supply voltage

5. Exposure time - 20ms

6. Resolution was not checked in this sensitivity testing.

 

 

Question: Do you have a test sheet of a known or common camera model? Could you post such a sheet on this thread please?

 

You can find some real tested models in the Table of Camera Models in VideoCAD demo. But these models are not widely known. My supplier offered me these models at the moment of testing (in 2007). Actually only one model in my test gave result equal to spec value - PELCO MC3710H-7X. Its sensitivity was 0.07 lx at 40IRE with the lens F1.2. The camera has 1/3" SONY Super HAD™ CCD

[Soundy 1]

1. Many camera manufacturers inflate their sensitivity values.

It's not just that... the root of the problem is, there IS NO defined standard for measuring and publishing camera light sensitivity specs in the CCTV industry. Fiona and Stanislav could have the best, fairest, most accurate test methods in the world, but there's nothing that forces the camera makers to use those tests, or any tests. As it stands, a camera that claims ".01 lux" could be referring to the level at which it gets a usable image... or the level at which the sensor gets measurable output. There's no standards body that requires them to say what kind of output they're looking at.

 

This really reminds me of when I was working in car audio in the early 90s... coming from a live/studio audio background, I was used to seeing power output specs for amps, that fell under various industry bodies and specifications for testing and reporting output numbers. Power specs were supposed to include information like the test frequency used and the total harmonic distortion produced at given power levels, and maximum power was supposed to be measured at a maximum level of THD (once the THD hit that point, that was assumed to be max rated power). Even home audio fell under the same guidelines for testing.

 

In car audio, all that was out the window. The industry had no standardized testing, and nobody to enforce it if they did. Same problem: a manufacturer could literally make up their numbers, even if they didn't make any sense. One $50 amp a friend bought claimed to be "200W" output, with nothing to back it up (it had a 10A fuse on its power line... even the most basic math for that is impossible: 10A @ 12V provides a maximum 120W, PERIOD, even before you include losses). Meantime, my car's subs were powered by an amp that was rated 30W/channel, yet came with a 20A fuse... and new, sold for around $500. And it SHOOK my car, with only two 8" subs. I blew the output transistors on it once, and there was NO noticeable distortion before it went (after I repaired it, I put a fan on the amp... never fried it again after that).

 

One of the biggest problems we saw, was that the cheaper manufacturers that DID measure their amps, would measure them while powered with 14.5-15V (which car electrical systems usually output while running with no load), which allowed them to generate more output power for a given input signal... and thus stamp the amp with a bigger number, since a design that gave you 120W output (for the sake of easy math) with a 12V source, would give you 145W output with a 14.5V source. The GOOD ones, the trusted competition brands, used regulated power supplies, and measured with a solid 12V source, which at least gave you CONSISTENT numbers, and gave you the same output whether they were getting 15V or 10V.

 

The best indicator we found, other than reputation, was actually price. It was (and probably still is; I've been out of it for years) very much an industry where you got what you paid for. CCTV is very much the same, especially when you look at the range of analog cameras, from the craptastic $15 IR bullets, to the likes of the Panasonic SD5s that retail for close to $800, but do everything except make your toast in the morning.

[Stanislav 0]

1. Many camera manufacturers inflate their sensitivity values.

It's not just that... the root of the problem is, there IS NO defined standard for measuring and publishing camera light sensitivity specs in the CCTV industry. Fiona and Stanislav could have the best, fairest, most accurate test methods in the world, but there's nothing that forces the camera makers to use those tests, or any tests. As it stands, a camera that claims ".01 lux" could be referring to the level at which it gets a usable image... or the level at which the sensor gets measurable output. There's no standards body that requires them to say what kind of output they're looking at.

 

The problem is deeper. Actually, there are several standards and it is possible to develop any necessary standard, but the CCTV industry doesn't need standards.

 

This is a commerce problem. And problem of CCTV designers knowledge. The most of buyers of cameras look only on sensitivity value and don't want to know more. While cameras with 0.00001lx in spec are sold better than cameras with 0.04lx in spec, the 0.0001lx will be written in camera's spec.

 

That forces honest manufacturers to resort to tricks too. In other case they would lose money. It is impossible to explain to all customers what is the sensitivity and why its value in their spec is lower in 10 times when the customers look only on LUX value. But the honest manufacturers can't lie directly. Therefore parameters in spec of honest manufacturers are right in general, but only skilled engineer can understand this. For example, sensitivity is corresponds to long exposure time or to low resolution.

 

Thus we again come to the designer's knowledge.

 

I don't want to force manufacturers, I just offer tools and articles for designers which can help them to increase their knowledge.

[Soundy 1]

And so we come back to the three most reliable indicators: reputation, experience... and yes, price.

[Stanislav 0]

And so we come back to the three most reliable indicators: reputation, experience... and yes, price.

I would like to divide experience + reputation and real knowledge. As I wrote above "Many designers consider lighting and sensitivity very complicated field which can't be understood. In spite of this fact the designers do their work ". These designers have good reputation and good experience but lack of knowledge of lighting and sensitivity.

 

Also the price is generally out of this question. The sensitivity is determined by used image sensor which is not only component of camera price. Cheap and expensive cameras (with the same sensor type) have really close sensitivity in spite of difference in price and difference in values in specifications.

[tomcctv 190]

Also the price is generally out of this question. The sensitivity is determined by used image sensor which is not only component of camera price. Cheap and expensive cameras

 

 

 

in cctv Price (consruction cost of camera) is a big factor .... you have to put price of each component together.

 

cheap ....... bad lens (plastic)

 

good quality glass optical lens

[Stanislav 0]

Also the price is generally out of this question. The sensitivity is determined by used image sensor which is not only component of camera price. Cheap and expensive cameras

 

 

 

in cctv Price (consruction cost of camera) is a big factor .... you have to put price of each component together.

 

cheap ....... bad lens (plastic)

 

good quality glass optical lens

 

Certainly these factors are important in CCTV as well as the price. But we speak here about sensitivity. So the brand name, lens material (with the same aperture), box material, inputs, outputs etc have no influence on camera sensitivity, but have influence on camera price.

Edited January 28, 2012 by Guest

[tomcctv 190]

have no influence on camera sensitivity.

 

 

where does all the collection of information start for any camera ??? Lens

[Stanislav 0]

have no influence on camera sensitivity.

 

 

where does all the collection of information start for any camera ??? Lens

 

Lens aperture (F-number) determines amount of light passing through it and thus only the aperture have influence on sensitivity. For example, there are very expensive varifocal megapixel lens with F1.4 and low cost lens with fixed focal length and F1.2. The F1.2 lens will provide better sensitivity in 1.36 times.

 

I also compared lens resolution and come to a conclusion that cheap fixed lens is better. But we speak here not about resolution .

[rory 0]

As I wrote above "Many designers consider lighting and sensitivity very complicated field which can't be understood. In spite of this fact the designers do their work ". These designers have good reputation and good experience but lack of knowledge of lighting and sensitivity.

I can look at the sun and tell you what time it is, how's that?

[Soundy 1]

Also the price is generally out of this question. The sensitivity is determined by used image sensor which is not only component of camera price. Cheap and expensive cameras (with the same sensor type) have really close sensitivity in spite of difference in price and difference in values in specifications.

The SENSORS in both cheap and expensive cameras may be the same, and thus the same price... the actual performance of the cameras themselves will tend to vary greatly with how well the resulting signal is processed. And so yes, price IS often a good indicator of a CAMERA'S image quality ("camera" referring to the complete unit, NOT JUST THE SENSOR.

[Fiona 0]

The mandate of any forum is to act as a watchdog against the outrageous claims made by manufacturers.

 

In the research of camera specification documents, it is now necessary to go back years to find evidence of manufacturers making the slightest effort to specify the most basic parameters behind sensitivity claims. The absence of any reasonably rigorous Standard that may be applied to any low lux camera claim ought to have been the first order of business on this forum when it was first established years ago. The most basic parameters or conditions that obtain in low lux claims are seldom if ever mentioned and all the while our fascination with cameras continues unabated.

 

There ought to be general consensus on the importance of verifying all sensitivity claims. Instead, outright deception by marketing departments has been allowed to pass unchallenged. Supposedly intelligent remarks are made about infinitesimally small low lux numbers while the accounting departments in Shenzhen and Guangdong split their sides laughing.

 

As a basic first step, any claim to low lux sensitivity ought to specify the parameters which obtain (and often dramatically distort) even honest manufacturer’s claims:

 

For example:

 

Exposure time: Shutter Speed - half second exposure times give excellent low lux ratings and the manufacturers who admit these nonsensical speeds are not overtly lying. But ½ second exposure times are generally nonsense. What? Is the interloper going to freeze on the spot to assist the camera? Often enough now, no exposure times are mentioned in lux claims and no other critical parameters are mentioned at all.

 

These are the two seminal documents posted by Stanislav:

 

Unravel and Undo the Unreal CCTV Camera Specifications by David Elberbaum

http://www.cctvcad.com/Files/cctvfocus37_unravel.pdf

 

The Secrets of Higher Sensitivity CCTV Cameras by Nikolai Uvarov

http://www.bezopasnost.ru/upload/iblock/af0/Nikolai%20Uvarov%20CCTV%20focus%2023.pdf

Edited January 29, 2012 by Guest

[rory 0]

The SENSORS in both cheap and expensive cameras may be the same, and thus the same price... the actual performance of the cameras themselves will tend to vary greatly with how well the resulting signal is processed. And so yes, price IS often a good indicator of a CAMERA'S image quality ("camera" referring to the complete unit, NOT JUST THE SENSOR.

A few years back when I checked I found several versions of the Exview chip, all with different prices.

[Fiona 0]

Further to the discussion on sensitivity claims are the specifications for two of the lowest lux cameras offered by Panasonic and JVC:

 

The Panasonic WV-CL930 or WV-CL934 ½” CCD and the JVC TK-C95 10U ½” CCD camera which appears to the replacement to the JVC TK-C1480BE 81BE ½” CCD mentioned by Uvarov.

 

Both claimed sensitivity ratings appear meaningless.

 

The first table is by Panasonic and refers to the WV-CL930 or the WV-CL934

 

The second and third tables describe the JVC TK-C95 10U

 

The Panasonic WV-CL930 manual offers less information.

 

This is simply an example of Stanislav's, Elberbaum's and Uvarov's argument about ill-defined lux ratings. This post is not about the comparison of cameras. It is about the comparison of ill-defined sensitivity claims.

[Stanislav 0]

As I wrote above "Many designers consider lighting and sensitivity very complicated field which can't be understood. In spite of this fact the designers do their work ". These designers have good reputation and good experience but lack of knowledge of lighting and sensitivity.

I can look at the sun and tell you what time it is, how's that?

 

It is really Great!

[Stanislav 0]

Also the price is generally out of this question. The sensitivity is determined by used image sensor which is not only component of camera price. Cheap and expensive cameras (with the same sensor type) have really close sensitivity in spite of difference in price and difference in values in specifications.

The SENSORS in both cheap and expensive cameras may be the same, and thus the same price...

There is a contradiction in this sentence "cheap and expensive cameras " vs " thus the same price"

the actual performance of the cameras themselves will tend to vary greatly with how well the resulting signal is processed.

May be. It depends what do you imply on the"actual performance ". But I speak here not about indefinite " actual performance". I speak about real defined parameter - SENSITIVITY (same exposure, same resolution, same signal/noise ratio, same lens aperture). This sensitivity will be the same with the same sensor.

And so yes, price IS often a good indicator of a CAMERA'S image quality ("camera" referring to the complete unit, NOT JUST THE SENSOR.

Sorry.. I must to remind you once again: I speak here not about indefinite " image quality ". I speak about real defined parameter - SENSITIVITY (same exposure, same resolution, same signal/noise ratio, same lens aperture). Please see Page 9 Camera sensitivity and Page 10 Parameters, limiting image quality at defining sensitivity.

[rory 0]

Ya xochu tebe soxranit tvou nervu

Ostav Rory v pokoe

sdes pochti ni y kogo net tools and obrazovania

teby ponyat

Looks like you have a really big hard on for me, troll.