CameraGimp

You won't find what you are looking for in CCTV. Progressive scan won't work with our switches and recorders so there isn't a market for them. I'd suggest you google machine vision cameras. They are usually progressive scan and may be of use. FYI the Pixim imager (used in cctv) may be progressive, pretty sure it is frame capture, but it then uses its electronics to produce an interlace output so still not of use to you but it makes me look real smart.

Too much edge enhancement on number 4.

I doubt very much he knows what you mean, or why would he be asking in the first place. I'd guess you mean you need about 50-70 pixels of horizontal resolution on a face to make it out for identification purposes? If so he is saying you need to think about the field of view (lens focal length) as well as the camera camera resolution. A low resolution camera with a tight field of view can be better than a high resolution one with a wide field of view. You won't need a mega pixel camera, but get a high resolution conventional camera (480 - 520 TVL resolution) and select your lens so a face takes up 50 - 70 TVL of the 480TVL, or about a 1/10 of the screen. Can't help with low light performance. Hopefully someone else will offer a suggestion.

If the pictures on the literature are accurate then they look good. As the improvements are at the blue end I guess they won't add much to day/night cameras though?? I think the reduction in power usage (heat!) will be useful. Can't wait to see the marketing that comes from these. Can't believe Sony haven't given them a name ala Exwave.

Assuming you are looking at imagers from the same family the only difference if the imagers are different sizes is sensitivity. They are built the same, they have the same number of pixels so a bigger sensor will have bigger pixels and that will be more sensitive. Obviously comparing a 1/4" Exview to a 1/2" CMOS isn't fair and all bets are off.

The problem you have got is the camera is allowing the highlights to clip. Cameras are dumb, they don't know what it is in the scene that you want to see so they makes some best guesses. In this situation it is viewing a darkish scene with some small highlights (the number plate) the camera has to make a choice, bring the large area up (by applying AGC) or expose the small highlights at the detriment of the low lights. Most cameras will bring the dark areas up and clip highlights, CCTV cameras should see at low light. You might try a video auto iris lens and adjust the pk/av to pk or turn AGC off (but that will make things darker).

I can't see a link to the starlight camera but I also suspect it uses the technology that videotiger mentions. It has a number of names, depending on the manufacturer, frame integration, slow shutter, sens up, digital slow shutter (DSS). They all do the same thing. The camera keeps the shutter open for longer than 1/50s (1/60s NTSC). This allow the camera to see in less light but the downside is anything that moves will blur. Think Starship Enterprise going into warp. If you look at a still scene it looks great but anything that moves will not look so good. Slow shutter is good for applications where you don't want to identify a face. For example watching a perimeter fence for intruders. It is unlikely that it will give you an image that you can use to identify a face. Day/Night should give you low light performance close to that of a monochrome camera but with colour images during the day. You can probably get a daynight camera with slow shutter. I'm not a fan of slow shutter myself.

I looked at 360 cameras about 3 years ago. I thought the idea was good but the technology wasn't good enough to make the cameras useable in all but selected applications where the end user was aware of the weaknesses of the product. At the time I thought the Grandeye camera was the best but even that one wasn't good enough for mainstream applications. The cameras didn't have the resolution, sensitivity or frame rate required or expected by most. I think the idea of recording a 360 scene and being able to ptz through the recording is a no brainer but only if the picture quality is useful. It wasn't when I looked.

I looked at 360 cameras about 3 years ago. I thought the idea was good but the technology wasn't good enough to make the cameras useable in all but selected applications where the end user was aware of the weaknesses of the product. At the time I thought the Grandeye camera was the best but even that one wasn't good enough for mainstream applications. The cameras didn't have the resolution, sensitivity or frame rate required or expected by most. I think the idea of recording a 360 scene and being able to ptz through the recording is a no brainer but only if the picture quality is useful. It wasn't when I looked.

Just to clarify. HQ1 is the DSP, NOT the CCD.

Most branded manufacturers do quote those figures but even those figures are incomplete. First off they measure the light falling onto the test chart but they do not state the reflectance of the chart so you can fiddle your figures if you wanted. Imagine viewing a mound of coal or snow, one will reflect more light, you can do the same with charts. Second you should state the light source used and it's colour temp. Lux is visible light. There is nothing stopping me putting a little IR in my test scene. The lux meter won't see it but my day/night camera will. There are standards which detail how cameras should be tested. For example British Standard, BS EN 50132-2-1, I think, but as not everyone uses the same standard I always suggest you take any specification with a pinch of salt and test a camera yourself.

It doesn't matter which type of solution you go for as long as you have IR pass at night and IR stop during the day. That is all that matters if you want true colour and IR sensitivity. How the camera achieves it is your choice. From a theory point of view using two CCD's should give you the best performance, a mono CCD will be slightly more sensitive as it won't have colour filters but that will cost and the performance gains won't be massive. I've never come across a two lens solution. I don't quite understand what that is. Mechanical filter is by far the most popular method.

It is ok to use AGC with a DC lens but I set the level with it off. Here is how I would set it Assuming daylight. Turn AGC off Turn electronic iris or AES off. Check DC lens opens and closes, flick level to min then to max. The video should go black without AGC noise and then it should white out. If it doesn't check EI and AGC are off. Set DC level. You want 1v pk-pk on a typical scene, if you don't have the equipment you can set it with your eye. Turn AGC on and check the video doesn't change. If the picture get brighter you haven't set the level right (or there is a fault). AGC is slow to act so I is easier to see when you turn it off. Turn AGC on, wait 5 seconds, flick AGC off. If the video level drops then AGC was working and you have a problem. Turn AGC on and check it, pop your hand in front of the lens and look for noise. If you do that and the camera still misbehaves then I'm lost.

A 1/2" imager with 1.3MP can't be as sensitive as a 1/2" imager with 440K pixels can it? Surely there is a trade off between resolution and sensitivity.

I second CollinR. You need to decide what you want from the camera before you pick your camera. No camera can do everything. If you want super high resolution you will have to accept weak low light or low frame rates. You can't have all things, even with money, unless you combine cameras in the same location ie mega pixel with another on the same scene. So what do you expect. Frame rate, light level etc.

So both can be made into day night cameras. The 1500 with an extra chip to control the switch (a few pennies) and the 2500 without the extra chip. The six first gen early adopter partners all got the option to build in two unique features so each partner had unique selling features. The Pixim chips were quite flexible and I think the first gen had some programmable switches from the get go and I'm sure they could have triggered an IO to switch a filter in gen 1 if a partner had asked Pixim to code it for them. Or that's what I thought. Perhaps I'm mistaken.

Hi Kurt, I'd say a monochrome camera with an IR cut filter will be quite a bit less sensitive than one without. It would be about as sensitive as a similar colour camera. I worked for a camera manufacturer for 12 years and in that time we never once made a monochrome camera with an IR cut filter and I can only remember coming across one camera by another manufacturer with an IR cut filter. That was a Phillips or Bosch Dinion and that was weird at the time. It makes no sense to do it unless there was some machine vision / footfall application that needed the resolution of a mono as far as I could figure. Ten years ago people bought monochrome cameras for one of two reasons, they were cheaper (this is back in the day that you could buy monochrome multiplexers) or they wanted the low light performance. Nowadays colour cameras are cheap and you don't get mono muxes so you only need a monochrome for sensitivity or resolution but most end users don't look at a monochrome image and say "hey thats got great resolution" do they. Kind of makes you realise that 520tvl colour cameras might be more marketing led than end user led. Anyways.. It isn't easy to see from the pictures you've linked to but an IR cut filter is a piece of cyan glass (actually a crystal). It should be sat in front of the CCD. It looks like it is there because I can see the green/blue tint. If you had a true day/night camera there would be some kind of mechanical device that would slide the piece of glass out of the way. They should replace the filter with another piece of glass of identical optical characteristics (so you don't need to adjust the back focus). Only this piece of glass doesn't have the IR cut part and passes IR.

An Ir cut filter on a monochrome would give a true greyscale not improve resolution. In the same way IR skews colours it will also skew luminance. it's just we like it to do this in low light surveillance applications. Imagine you look at a person wearing black cotton pants and a black fleece jacket with a colour camera you will see black pants and and equally black jacket. But if you look at the man with an IR sensitive mono he will have black pants and grey jacket because fleece reflects IR and so doesn't appear black. So an IR filter won't alter resolution or sharpness but it will make sure things look true to life.

Any imager can be a true day night. Assuming you mean IR sensitive at night and not IR sensitive during the day to be true day night then true day night comes as a result of moving the piece of crystal sat in front of the imager and it has nothing to do with the imager. So the original Pixim imager could be a true day night if you wanted it to be as could any imager that is IR sensitive (which is all of them).

IR CCTV cameras do not work like your binoculars. I think (not 100% sure) but you need to look at light intensifiers to get good information. An IR CCTV camera is sensitive to IR and so can see it. The camera just turns that into a TV picture (in simple terms) where as light intensifiers take weak light, use magic to make it brighter and then show it out visible to the human eye. Basically they magnify the light where as IR CCD's see different wavelengths invisible to the human eye and so appear to see in the dark.

As well as the size of the lens, another reason is they use similar mechanics and lenses for camcorders so they get massive economies of scale. There are quite a number of different dome manufacturers but not that many camera / lens module manufacturers. It is quite sewn up.

Is that last picture real or have you touched it up to highlight the fault because it looks a lot worse than I would expect and considerably worse than the previous pictures you posted. I still can't help you but what you are describing and the first pictures posted sound like pixel or blemish faults. They are a fact of life and most camera manufacturers will not warranty against them. However that last picture does look bad, a lot worse than I would expect. If it is blemishes and the camera is new you might have a case to get a replacement. Usually a blemish is small and not too noticeable especially when you are viewing a normal scene, they get lost in the pattern whereas that last picture they look very noticeable indeed. Here is what I know about blemishes. If you've got one it will always appear in the same place, the don't move. They can appear more pronounced as the ccd warms up, so they may not be visible on power up but appear after a few minutes. They are usually white but you can get different colours. If that describes your dots and they are not acceptable I suggest you contact the manufacturer.

They are faulty pixels and there is nothing you can do about them. How they happen is open to discussion. Pixels can fail in a number of ways such as black, white or somewhere in between and you normally see them at low light and high gain.

The poor picture from the old camera could be because it hadn't been touched in ten years. A ten year old HR mono camera will have a higher resolution than a colour camera today. When you fit your new camera and it looks better it will be because you've just adjusted it, not because it is a better camera. Backlight compensation (BLC) is a feature that was developed because CCD cameras have a small dynamic range. BLC does not increase dynamic range, it just allows us to use the small range we have better whereas WDR is a camera with better dynamic range so it shouldn't need BLC. WDR is closer to how we see, BLC tries to guess what we want to see and expose that correctly but it will expose other bits incorrectly as a consequence. Can't say. Depends on the light on the scene. From your previous post you will probably be ok. But this is where experience pays.

Before you spend your money on a WDR camera. Were you happy with the pictures from the camera that failed? If so I don't think WDR is necessary. Todays colour cameras are going to be as good if not better than a ten year old mono camera for WDR. By all means spend the extra money if you want but do you need to. You should be able to reuse your old lens. You have to make sure the camera you fit has the same size imager as the one you take out or you'll get a different field of view (most likely 1/3") and your new camera has to take a CS mount lens (very common). If you change from a mono to a colour you will probably lose sensitivity, even a ten year old mono will beat a modern colour (low cost one anyway). So be aware that your night time images may not be as good but it sounds well lit so I hope it won't be an issue. In which case I'd suggest you go for any middle of the road 1/3" high resolution colour camera. It should take the lens, power, video and bracket you already have and fit in a matter of minutes.