Stanislav

Fast reverse gear:) It is good illustration of uneducated designers who trust in your approach "easy intuitive design" with "easy to use simplified 3D design software” you promote. Their projects look glamorous with girls and boys but make laugh and angry installers in the field. It is also good illustration that inspection on real place is obligatory. Design software is an auxiliary tool. Project quality depends on designer skill and work. It is very good illustration of your approach, Maxim.

Good way to find the best positions for your cameras is modeling in special CCTV design software. You should draw your building in 3D quickly and roughly without detailing. But you must keep accurate dimensions. It will not take much time. Then you should place cameras at possible places, enter camera heights and basic camera parameters: Lens focal length, Image sensor format, Resolution in pixels on horizontal and vertical. After these actions you will be able to play with cameras, change places and camera parameters and see 3D models of images from your cameras. Thus you will select cameras and position fit your needs. In this way you can achieve better result than by following advices from other persons who did not visit your place. Real dimensions are great important. Your pictures can not contain all dimentions. Other persons have not information you have. Any demo version from VideoCAD Starter to VideoCAD Lite will be suitable. You can use demo versions for free within 30 days. VideoCAD Professional demo version is not suitable for such tasks. Professional version allows much more but its demo version has considerable restriction (it does not allow to change lens focal length). VideoCAD Professional is intended for professional using. VideoCAD Lite demo version is good free choice for such amateur tasks.

IVk-RgLR0jI

Today, all Lens calculators and CCTV design software (except VideoCAD Professional 8.1) calculate view angles of wide angle lenses with errors, because they do not take into account the Lens distortion. About Lens Distortion Additional errors are expected in calculating the view angles of modern cameras because of the image sensor size is considered in inches from the time of vidicons not taking into account existence of the modern image sensors with aspect ratios different from 4:3 and/or using not whole pixel area for scanning. See Specifying active area size of the image sensor

What do you think about such Fields of view? What lens calculator can calculate them correctly?

VideoCAD 8 Professional can simulate panoramic cameras. See details http://cctvcad.com/videocad_help/index.html?prim_fisheye.htm You can also model wide-angle cameras with strong lens distortion. See Modeling lens distortion

Here is an idea of choosing frame rates based on the desired frequency of getting in the frame objects with the known positions and velocity. http://cctvcad.com/videocad_help/index.html?prim20_framerate.htm

Thank you, MaxIcon Here are several images for better understanding sharpening. Level of one line from the images is shown as a curve. Normal image with a little noise. No blurring and no sharpening. Pay attention on vertical curve without overshots at the line drop point. Blurring. Line at the drop point is not vertical. Any lens has limited resolution. If the lens resolution is unsufficient, it adds blurring. But lens blurring doesn't reduce noise because of the noise is added on the image sensor. Little oversharpening. Line at the drop point is strictly vertical as on the normal image. But overshots appear. There is optimal sharpening level which can correct lens blur and approximate normal image. It allows manufacturers to use cheaper lenses. Images with little oversharpening look better. But sharpening as any other image processing can't restore detail lost because of limited lens resolution, it can make images which will look like a normal image but this corrected image will not contain all details. Pay attention also on increased noise level in comparison with normal image. It is unnoticeable on image but it is visible on line curve. Noisy image without sharpening. The same noisy image with the same little oversharpening. With noisy image oversharpening corrupts it. To prevent this effect, many cameras turn off sharpening automatically when the image becomes noisy.

Noise reduction and sharpening are not contrary each other. They are different procedures. There are a lot of noise reduction algorithms including 3D noise reduction. It is not simple blurring, but image blurring is a side effect of NR. General advantage of NR inside camera is traffic reduction in low light conditions. Sharpening is image processing which can correct low resolution of lens in some degree. But noise increasing is a side effect of sharpening. Thus you should try to use sharpening if you use a lens with not enough resolution in good light conditions. But you should not use sharpening with low light noise images. It makes the image more noise. If your camera works in low light condition, using noise reduction can decrease traffic and disk space requirements. In case of limited bandwidth or disk space using noise reduction can increase image quality in whole.

Here is expected difference between f=2.8mm/F2.0 and f=.3.6mm/F1.8 in low light. When the light is sufficient there will be no difference in contrast and noise. f=2.8mm/F2.0 f=.3.6mm/F1.8

Cheaper tool with less capabilities - VideoCAD Starter ($85) Cheaper tool with more capabilities - VideoCAD Starter II KIt ($198)

Please see my article Illumination and camera sensitivity in CCTV. Therein is some information how these images are simulated. *IR illuminators can be modeled too.

You need not a tester. Just download VideoCAD Starter, you can use it for free in 30 days and model your situation in details. There is a full free software - CCTV Design Lens Calculator, it can help you too.

The difference between F1.2 and F1.4 depends of current light level and camera sensitivity. If there is a bright region on the scene then current camera exposure can be limited by this region and changing lens will not improve image at all. If the scene is low illuminated uniformly, you can expect theoretically difference as shown on the images. But real difference can be a little more because varifocal lens probably transmits less light than equivalent fixed F1.4, it has more complicated construction and more optical losses than fixed lens. Also the F1.4 value in varifocal lens spec F1.4, 3.5-8mm can be applied to 3.5mm focal length. With 8mm focal length the F number can be worse.

Of course I should write "The thinner your cable is - the more its resistance " Thank you for this correction. Thinner cables have worse quality and less price. They have more losses in hight frequences, than thick cables. Actually thin cables have more losses in direct current too, but in hight frequences their losses are much more. Thin cables have worse amplitude-frequency characteristic. Not exact. Loss in high frequance is caused by resistance too. "Shunt capacitance per unit length, in farads per metre." and "Series inductance per unit length, in henrys per metre" don't lead to losses. These parameters determine "Characteristic impedance in ohms (Ω).". Losses are caused by "Series resistance per unit length, in ohms per metre" and "Shunt conductance per unit length, in siemens per metre.". "Series resistance per unit length, in ohms per metre- The resistance per unit length is just the resistance of inner conductor and the shield at low frequencies. At higher frequencies, skin effect increases the effective resistance by confining the conduction to a thin layer of each conductor." "Shunt conductance per unit length, in siemens per metre. The shunt conductance is usually very small because insulators with good dielectric properties are used (a very low loss tangent). At high frequencies, a dielectric can have a significant resistive loss." See details: http://en.wikipedia.org/wiki/Coaxial_cable#Fundamental_electrical_parameters Thus the Series resistance and Shunt conductance are raised on hight frequences. Because of thinner cables have more "Series resistance per unit length" and "Shunt conductance per unit length" than thick cables, they have more losses in high frequences.

See also CCTV Design Lens Calculator. Together with usual functions of Lens Calculators, it offers several new tools, such as calculating field of view projections in 3D, simulating camera resolution using photos and images including megapixel cameras up to hundreds of megapixels and other. CCTV Design Lens Calculator is free

With VideoCAD it is possible to simulate in 3D the motion blur in dependance of shutter time and velocities of objects. See details: Video surveillance of moving objects You can also measure real maximal exposure time of each sens-up mode using only an analog oscilloscope.

I made such scheme 15 years ago for one of customers. I used multichannel RF modulator, IR remote control and had developed special device to control it. In that years I was keen on electronics and even the modulator had been made by my own hands

Your DVR produces standard VGA output. Further processing depends on the capability of your TV. For example my TV can several screen modes, it can stretch and cut images with different aspect ratios.... but all these amazing options don't work with VGA input! They work well with antenna input, AV input, S-video, component input, HDMI.. But image from the VGA input can be shown only as a " letterbox " May be your TV can process VGA input or not, you need to test it with VGA input exactly.

Generally varifocal lenses have more optical aberrations and optical loss, less resolution than fixed focal length lenses. I suppose it is better to avoid varifocal lenses in those cases where they are not necessary. It is very simple to calculate appropriate focal length using lens calculators or CCTV Design software.

In reality there are many factors which have influence on the identification (view angle on the face, light, contrast, compression, lens resolution, noise, range of heights, shutter speed for moving persons etc). But it is difficult to take into account all these factors. Only the pixel density (spatial resolution) can be simple calculated. Because of many influencing factors weren't taken into account, it can't be exact values of the pixel density for all possible cases. Nobody can guarantee that with 40 or 80 pixels per face the face will be identified. But with 80 pixels this is more probably. 35, 40, 80, 83 pixels are very rough empirical data. But we need the data for our work. To get more accurate result we have to adjust these values according to real conditions. Sharpening can correct image blurring from insufficient resolution of lens. But it is difficult to choose exact level of sharpening. Therefore in many cases we see images with excessive sharpening. The excessive sharpening becomes an additional distortion which decreases possibility of identification.

Try a polarized filter, but I don't think that it can help you. To minimize the glare on the windshield try to choose better locations for the camera and the light source.

Please see my article about person identification criteria, may be it will be helpful. In the other article there is a comparison between megapixels vs shutter speed in low light.

May be your cameras are True Day/Night ones with removable IR cut filter. Such cameras are sensitive to IR only in B/W mode. From some reasons like a camera bug or illumination features (for example on the scene there is a bright region) these cameras sometimes don't switch to b/w mode. Thus the IR cut filter remains and cuts IR part of light.

Not in all devices all inputs have independent ADCs. Some frame grabber chips have a multiplexer (simple switch) between several inputs and one shared ADC. This scheme limits the maximal frame rate from multiplexed non-synchronized cameras because the ADC must wait for the moment of starting frame from the next multiplexed camera. That is the reason of difference between max frame rate of "one camera for one chip" and "several multiplexed cameras for one chip". For example for one camera - 25fps, but for several cameras in sum - 12.5fps. I tested a PC-based video recording system several years ago. The system has a mode which allows to considerably increase number of frames from several multiplexed cameras. But in this mode the system became sensitive to sync. When I connected synchronized cameras - the images were good, but with non-synchronized cameras their images were unstable on vertical.