Scruit

For a residential application, if you had to choose between 7.5fps per channel in D1 or 30fps per channel in CIF, which would you advise your client to use? Your client wants to be able to use the video to help the police track down and prosecute anyone who breaks in. Each doorway has a camera where a person's face takes up about 25% of the image height. Assume you cannot change cameras, and assume there is enough HD space for 14 days of D1 recording, which is acceptable to the client.

Hmmm. I'll have to figure out how to mask the plate like inthe pictures above using my video editor - or repeat the test. Let me have a think about that.

MPEG4 and MJPeg are two different compression standards. My DVR uses MPEG4. D1 and CIF are two resolutions. The exactl resolutions can change from one manufacturer to another, but D1 is approximately 720x480 or 640x480 and CIF is approximately 320x240.

Those are D1 an CIF taken from the CCTV... Not sure what you mean by 'original'..? ---------------------------------------------------------- each pix "property" shows as 720x960 That's 2 720x480 images that I put into the same jpg image 480 x2 =960 . There was no quality loss putting them into the same image.

It's on a long gravel driveway that is only wide enough for one car. Right now the camera is focused on a point about 60' from the house using a 5-85mm verifocal auto-iris lens. Cars can sometimes pass by this point too quickly and result in a blurry image, so I'm going to move the 'killzone' to the end of the driveway where cars have to slow/stop as they enter/exit the drive onto the 55mph road. Hopefully the slower cars will result in a better image. I still have problems with the lower contrast during the day (in the night image, IR illuminated, it plate is crystal clear) so I'm using two channels on the DVR (loop out from ch1 into input of ch2) and the ch2 brightness/contract settings are adjusted for better contrast.

Those are D1 an CIF taken from the CCTV... Not sure what you mean by 'original'..?

I did the experiment: 1) Set DVR to D1 mode, max framerate, max quality. 2) Drive past License Plate Camera (best indicator of detail) 3) Replay video, pause, capture, save single image 4) Repeat above in CIF, max framerate, max quality. Result: Conclusion: I'm going to stick with D1.

The avermedia EB1304MOB in my car shows no appreciable difference between CIF and D1. I recorded a scene in both modes and took stills of both. After mixing them up I could not tell which was which. So I run in CIF mode on the car DVR. I'll do the same experiment with my home DVR. It'd be nice to have full motion video on all channels, but not i the image is much worse.

I'm not an expert - just a home user... There's so many other factors that you can't go by just TVL. The better cameras have much better light handling and DSP post-processing. Also you can to consider the resolution you are recording in... If you record in CIF then the extra resolution isn't going to help you much. I have a name-brand 480tvl camera and a non-name 480tvl and the difference is astounding when you start looking at changing light conditions etc. Then again, my short experience tells me that camera placement and focus is more important than the resolution... I tend to use cheaper cameras when precise detail is not important (ie an overview of the front garden), where the su bject of the camera is in a very predictable location with predictable lighting (indoor face-shot at a doorway) or when there is enough external light that the cheaper camera can still see. I use more expensive cameras where the location of the action is not as predictable but detail is still important (I have one camera covering two adjacent doors that at 10' apart. or outdoors where lighting is unpredictable. Finally, I have had a 330tvl camera produce images every bit as good as a the 520 tvl camera that replaced it. The 330tvl was set up just right, and the 520tvl was not. All other things being equal, the better tvl may give you better resolution -but there's no substitute for expert camera placement and lens selection. For that I defer to the experts here at this site.

My license plate cam can read a test license plate that is set out in the driveway under any light conditions encountered so far. I've been running 24hr tests where I have a computer saving an image every 5 mins and then I can run through them all quickly to ensure the LP is readable at night / day / sun / cloudy etc. that's fine. Problem is that when folks drive past the camera quickly I wind up with a slightly blurry image, like the shutter sheet is too low - also my CPD507HC is giving me an interleaved image (D1) that makes it even more blurry. I have an auto-iris lens. The stationary plate I put out there for testing is easy to read. I am about to move the LP camera to a location where poeple are driving towards/away from it (10' off the ground through) to minimize the side-to-side movement of the plate while it is in shot... but I'm thinking I can solve the 'speed of movement' issue by making the car slow down. And that's where I could use suggestions... Anyone got any ideas on how to make cars that come to my house slow down in the middle of the 150' driveway. I'm thinking that a speed bump would work but would look stupid in my gravel drive.

The driveway alarm is integrated with custom software I worte for my PC. When someone drives into my driveway the software takes a snapshot from teh CCTV (using a video capture card) and emails it to my phone and a coupel fo email addresses. it also sends another picture 20 seconds later so I can toll if the car had stayed at the house or left.

When I took pictures of a fender-bender I witnessed and testified in court, I was asked by the plaintiff's lawyer to explain how the pictures got from my phone the printout. This was to show he judge that I hadn't messed with the pictures. I just said that the pictures were taken by my cellphone with the defendant and plaintiff present, emailed to my home computer then printed on a Sony photo printer. Judge was happy with that - although the defendant never challenged the pictures and the bodyshop manager said that the damage to the plaintiff's car was exactly like picture I took. There really was no challenge. I'm not concerned about the 'overall' cameras being challenged in cvourt if someone breaks into my house - but I can see them simply denying that they are the person depicted in the pictures. That's when I hope my 'identification' cameras (LP, close-up of face at each doorway etc) will help me out. I can see stills being challenged more than the video... And the CPCam usb export is in a prioprietrary format with a watermark and the claim is that they can't be tampered with. Hopefully I can use the proprietary player as evidence that the video is good. Of course, I do expect that someone stealing a couple thousand bucks in property from house would require less validation and verification than if I caught a murder or rape etc on tape.

I don't get a lot of turn-arounds - although I get enough that the driveway alarm is set about 20' in so that it doesn't alert me to folks just turning around. I don't mind the turn-arounds. I *do* get about one car a month entering the driveway and doing something strange. A couple times the car just stopped halfway down my driveway for a few minutes in the middle of the night, and then sped away when I turned on my bedroom light. Anotehr time a guy drove up to my house and then spent a couple of minutes staring in through the windows of my car, then then left - the was not 'for sale' nor should anyone think that. One time a lady drove up to the house and spent 10 minutes walking around the back of the house, around the barn, my workshop and pool. I didn't see her until she was leaving, and she claimed she thought the house was for sale. No for sale sign or anything... Just lots of weird things like that. If someone is going to steal from my house then I'm assuming they will drive up to the house. Unless they know there is a license plate camera then they will not want to make multiple trips downteh 120' driveway (about 180' to the workshop where the last burglary occured). They'd be at the house too long and the ADT alarm will hopefully be triggered. The other consideration with the aiming the license plate camera at the end of the driveway is that passing cars may trigger the motion record - but that's not a problem. The new DVR has been in for a week or two now and it's only used 25% of the HD so far. Plus there's only about 100 or so cars per day pass the house.

The LP camera is currently mounted under the eave next to the front door - so it's looking across teh driveway at an angle. It's also easy to reach up and disable. The new location is in a true housing with heater/fan and is out of reach / attached to the house and is perfectly inline with the driveway - although about 2' higher. If I drop it any further then it will be in reach again... The housing is installed/wired etc and I'll move the LP camera to that location this weekend. If the cars are still moving through shot too fast for the camera then I'll consider slowing the cars down. Another thought... Rather than having my 'kill zone' halfway down the driveway... This new lens has enough zoom that I can have my killzone at the very end of the driveway - and everyone has to come to an almost complete stop before exiting the driveway because theyr's exiting from an 8' wide gravel driveway into a 55mph road next to a blind hill. Duh me - that's a location where cars HAVE to slow down, if not stop. I never thought of this becuase the lens I had up until last week didn't have enough zoom to reach that far (was limited to 60', driveway is 120') but this new lens goes up to 85mm. I'd probably aim for about 15' - 20' short of the road (about where the car's rear plate will be while the driver is waiting for a gap in traffic). I'm sure my 85mm lens will give me a one-car-width view at ~100' I can't believe I was considering building something to slow the cars down when the cars already have to stop, just 40' away.

Do you use motion detect exclusively? Or do you use manual recording? I wanted to use manual + motion on this CPD507 because I can set the manual framerate to be very low and motion record to max. The manual says that motion record can be set to group mode, meaning "the channels that are recording" will share the max framerate. ie Channels 1-4 = group 1. They share 30fps Channels 5-8 = group 2. They share 30fps. The manual gave the example that if only 1 channel is recording then it gets 30fps. If two are recording then they get 15fps each. Fair enough. Problem is that when I have manual recording set on then ALL channels are recording, and it appears that even though only one channel in the group has motion, that one channel does not get 30 fps, it's actually closer to 7.5fps. I guess I can get around this by recording motion only (no manual record) and use the pre-event option also. I just have to make sure my motion setup is good.

My manual record is set to 15fps. Alarm record is 120fps total. This is in Frame/D1 mode. As far as having two cams showing he same thing in the same group... I *thought* it would work that way, but it doesn't. if I turn manual record off (and the green record icons on each channel turn off) then any motion in ANY ONE channel will make 9 channels with-cameras-attached start recording (green record icons all appear together) and share 120. (in groups of 30fps /4channels. of course) It doesn't use the 120fps JUST for channels with motion. I observed this behaviour, and confirmed with CPCam support. I wish they ekpt the "alarm fps is just for channels with an alarm condition" feature from the 576, but alas no. ie if I have two cams on the front door and two on the back door. Common sense would say put ONE front and ONE back in group 1, and the OTHER front and OTHER back in group 2. If someone goes to the back door then group 1 will give the front camera 1fps an the rear cameras 29fps, and group 2 will do the same for the other 2 cameras... BUT IT DOESN'T!!! Motion on any camera will make all channels record. So the 2 front cameras and 2 back cameras will get 15fps each regardless of is ther are split across group 1 and 2 as above, or if you have the 2 fronts in group 1 and the 2 backs in group 2. I really hope CPCam considers fixing that to leave non-motion channels at manual record rate. Right now I have my screens split across the 4 groups as 2, 2, 2 and 3. The license plate camera is being looped out of the 507HC to a seperate 4ch DVR that does 30fps and I have the camera going into ch1 then looped into ch1. ch1 has brightness/contrast settings that work for night, ch2 has brightess/contrast settings that work for daytime. The camera itself can see a license plate perfectly will under all lighting conditions, day/night, car with hi-beams aimed at camera etc - the brightness/contrast just provide a little extra help for reading the plate. The 507HC record the LP cam in default brightness/contrast. This gives me 3 seperate recordings of the same license plate image (with no quality loss - because I'm looping instead of just splitting). The hope is that at least one of the recordings has the 'right' brightness and contrast. I cannot read the plate on my car - but that's by design. My wife's car is easy to read the plate on, every time. Since adding the auto-iris lens to the LP camera I have read every plate that has come into my driveway (except where I couldn't even have read the plate with the naked eye)

I just use the maximize button. Then Esc to get back to a window.

I have two more options I guess... 1) Chicane... Put two offset obstructions that force cars to swerve left and the right. Would be a PITA for the propane truck, though. 2) make the driveway thinner, force cars to drive slowly and carefully though a width reduction. I've been planning to install a brick border to keep the gravel inside the driveway - I could use the edging to create a chicane or width restriction that would not look out of place - not like a speed bump etc. The areas where the edging creats a width restriction could have planters or a bird house etc. Ideally though I'd want to be able to remove the width restriction temporarily to allow the propane truck to back up down my driveway... Hmmm 3) option 3 is to move my camera so it's more central to the driveway so that the license plates are not moving trhough the FOV at 45deg - they are moving almost directly towards or away from the camera. This would hopefully reduce the effect of the motion blur.

Ok, so they took you on your word that the images and chain of custody was as you described it - instead of requiring you to produce actual evidence that the images were original (ie bringing the DVR in to court) I've been wondering what challenges I may face in court if I ever need to use my CCTV video in a prosecution.

The night images are the best. My problem is mostly during the day when there is enough ambient IR light that the contrast between the letters and background is not as high. I was thinking of something weird... I could take a loop out of the channel that has the license plate camera on it and feed it into another DVR channel - then I could alter the brightness and contrast on both of those channels so that one of them gives me best settings for day, and the other gives me best settings for night. Right now I'm going through all the different combinations of electronic Iris & backlight compensation settings and doing a 24hour test with each. So far I have seen absolutley no difference between BLC On and Off in this application. Today/tomorrow I'll be trying the electronic iris on (even though it has an auto-iris) to see if the shutter speed goes up or down... What I really need is to be able to manually set a higher shutter speed and have the camera compensate with the auto-iris lens.

For now I'm going to use the loop-out into an older 4ch DVR that I have that does not have the same interlacing issue. Case in point, today an unknown pickup truck entered my driveway, spent one minute parked beside my house, then left. The license plate on the back was almost illegible, but the 'test' plate was easy to read. I need to figure out a way to make the camera use a higher shutter speed. (Turns out the pickup was the electric meter reader - he apparently traded his red Ranger for a white GMC)

Definitely sounds like a hard drive problem - if it can't even get the data off the HD. Have you tried a different HD?

Did you have to *prove* that the images were not tampered with (ie watermark etc), or demonstrate the imge recovery steps? Or just state that the images are as on the DVR.

Check your local laws first. I'd be amazed if there was any school in the US that would broadcast images of children inside a classroom out on the public internet. It's one thing to keep the images on closed circuit and available on to a limited number of administrators - but letting anyone who claims to be a parent log in .... Nuh-uh.