NotoriousBRK

ISC West is 10x the size of ISC East, maybe more. If ISC East wasn't practically in our backyard I don't know that we'd exhibit at it. If you go to ISC West next year, PM me and I'll get you some passes to the show floor.

I've put hundreds of analog cameras on fiber. American Fibertek, Comnet, IFS, etc. all make good fiber stuff geared towards cctv/access control/intercom/etc. Who is talking about analog cameras?

I don't think it's any easier, especially when you take into account some of the enhanced flexibility you get from an IP system... The ability to cheaply backhaul multiple cameras over a wireless link (pair of Ubiquti radios, $150), ability to cover extreme distances without video degradation (fiber), remotely adjust camera image parameters and so forth. I feel like some people are looking at HD-SDI not for the topology, but more for an excuse to not have to learn IP in depth. IP networks are so ingrained in technology today that there is no way (IMO) to make a solid living going forward without being very IP/Ethernet proficient. Sure, if you're 5 years from retirement maybe you can scratch out the last few years without having to understand how a subnet mask really works. But if you're just getting into this, or have a lot of years left, your time is going to be better spent learning IP networking instead of wasting time on stuff like HD-SDI.

IP camera system "issues" and worries of network saturation are greatly exaggerated. HD-SDI is a novel concept. Most people that have tested it have found that it's not going to work with "just any" RG-59 cable, so be prepared to do lots of cable pulling and re-terminations. Most of the demo videos I've seen are not at all realistic. The HD-SDI DVR's are going to compress the video before recording it, using h.264. The live stream might or might not look better when compared to some IP cameras and VMS systems. But you can't record an uncompressed 1080p video stream without having literally hundreds of terabytes of storage. So, the recorded video (eg: what the customer will actually end up dealing with 99% of the time) will look exactly the same as video from a quality IP camera.

I use OmniGraffle Pro. And a lens calc wheel. This video is about a year old, but conveys the general idea: http://stream.myicvr.net/CameraLayout/

It's another cable, another termination, more cost, more parts, etc. I've also found PoE to be more resilient to voltage drops. Why run an extra wire when there is no need?

Seconded. I've used and test both (and have one of each brand on my house), both very rugged, solid builds, good support from the manufacturer. For most applications (unless it is a really close-range thing), I'd stay away from the "little LEDs sprinkled on a PCB" units.

it IS invisible I've yet to see a truly invisible IR illuminator. Most SD camera sensors taper significantly above 1000nm in terms of IR light sensitivity, and all the 940-950nm illuminators I've seen STILL have a visible glow to them. You *do* have to get closer, but when it is pitch black out you still see some red light. On top of that, the 940nm wavelength is (roughly) about 1/2 to 2/3's as powerful as 850nm light, so you need more power, meaning more cost.

Analog cameras will likely be cheaper, and simpler. IP ("digital") cameras can offer higher resolutions and other benefits in terms of expansion and flexibility. Very few people install analog cameras in new deployments unless they are going for absolute lowest cost or basic systems (in my experience). Analog has been around forever, there is a wide selection of products, any cctv installer can handle it. IP scares some people who don't understand networking. If your installer seems highly biased against IP, I would be cautious. Even if it's not the best option for your system, you at least want someone who can handle current technologies. Impossible to say without more details. In general, hard-wired is always better and more reliable. Various forms of wireless may make sense where cable runs are difficult or impossible. In almost every case the wireless component would be external to the camera. Cameras with built-in wireless are almost always in the "toy" category. If you already have smartphones (and a blackberry no longer qualifies as a smart phone), try to find a system compatible with your current phone(s). In general though, on most modern systems this is pretty easy to accomplish in a number of different ways. You are not going to be able to do remote recording (if this is what you are referring to) without a lot of cost and complexity. Video from each site would be recorded in one or more devices (cameras, DVRs, NVRs, etc.) at the local site. In some scenarios you can arrange to have some or all of the video backed up to a remote location. Unlikely. Most AIO systems cater to lowest common denominator. While your needs do not appear to be very high-end, they also seem a bit beyond basic. Proper system design is a factor of things like area to be covered, expected detail (anything you've seen on TV about enhancing images after the fact is pretty much garbage. If you don't record high-detail, you won't see high-detail later), lighting at the site(s), budget, and so on. From what you described, you will have several choices, starting most likely around $500/camera (installed) up to $2000 or $3000/camera (installed), but this varies somewhat by region and so on.

Do you know that your camera are capable of outputting a decent enough signal? I don't really deal much with DVR cards (or DVR's/analog stuff in general) these days, but for the most part a lot of that gear is commoditized, I think they are all of similar quality (provided you stay away from absolute junk). If your current DVR card isn't producing a suitable image, the problem might be the source, not the recorder...

Based on skimming the manual, that model appears to have the BNC connector cable replaced with a balun-ized cable assembly. You should be able to find the same camera with a regular BNC output. Or, worst-case, put another balun at the camera to convert it from UTP to coax.

As a rough estimate, you're going to need at least 75 pixels per foot to get really good image detail. You might be able to get by with 50ppf in really good lighting. It's easiest to figure this across the horizontal FOV. On your 1080p camera: 1920 pixels across the HFOV / 755 pixels per foot = 25.6 feet. In any area of the field of view where that camera is covering an area 25 feet wide or less you should get good image detail (provided you have a good lens, proper focus, etc.). If you are trying to cover an area 100 feet wide (for example), then you are spreading each pixel over an area 4x as large as what you need to in order to maintain good image detail.

For small installs (around 24 cameras or less), any basic modern switch will be just fine. I would stick with a name brand (D-Link, Netgear, Linksys, Cisco, etc.) just to be safe, but you don't need any special features or functions for basic stuff.

I have Mac Mini's on all my TV's that serve this purpose (plus other things like playing DVDs, interfacing with the home automation system, etc.). Easiest way is to have a static page/bookmark for the cameras and common layouts selectable via hotkeys over an IR remote, or using the Mobile Mouse app. Hook the 'Mini to an HDMI input on your TV, and the rest is pretty easy. YMMV.

The lack of budget, and high expectations make this sound like a recipe for disaster. I would refer this job to my most-hated competitor.

Avigilon has been doing some great business, they have solid products, and are obviously looking to raise money to continue to grow the business. In this market, they have many things in their favor in regards to an IPO, and could end up doing very well, certainly much better odds than raising more VC. Also, they threw a helluva party in Orlando this week

Get an analog to IP encoder and then send the IP video over a wireless IP link, into a decoder on the other end and then into the DVR.

Blackberries are horribly for most streaming video. Some do better than others, but many can't handle resizing a stream, so if the resolution of the stream is larger than the screen (highly likely) you may have issues.

If you had access to a linux box you could do it with cron (for scheduling) and either wget or curl (http://linux.about.com/od/commands/l/blcmdl1_curl.htm) for formatting/sending the http commands.

Do you mean at what point would I try to "cap" the bandwidth to prevent overrunning the NIC? I would personally try to keep it at 400Mbps or less, but I think in general you will find that the HDD sustained write speed (especially with occasional reads if someone is going to be watching or searching recorded video at times) is going to be the bigger limiting factor. Still I think the question may be too broad, since the best answer may apply differently based on the scope of the system. A gigbit Ethernet can sustain a max data throughput of around 750Mbps. Barring other data, I'd usually derate any such spec by 1/2, eg: if a device says it can support X, I'd try to engineer a system based around that device for 1/2 X, unless X is some radically simple data point (like a max operating temp, if a camera says it operates to 122F, I don't mind operating it at 100F).

If your Internet router has multiple ports, then connect both switches to the router directly. I wasn't sure if you had the typical 4-port switch style router, or the typical 1-port ISP-provided style. In the grand scheme of things I doubt you could tell ANY difference in either scenario though.

You are most likely worrying unnecessarily. Creating 2 IP subnets won't change anything, it is the switch fabric, or backplane, that would be the potential bottleneck. Not sure what kind or qty of IP cameras you are going with, but 30Mbps per camera (MJPEG) would be a bandwidth estimate on the high-side of things. 6-10Mbps if you're using h.264 (and I'm assuming 2-3 megapixel cameras here. If you're only doing 720p, then it's even less.) If the 16 port switch had 16 cameras x 30Mbps, that would be 480Mbps, or about 2/3 of your effective available gigabit LAN bandwidth. The other typical stuff you do (eg: Internet crap) would have PLENTY of available bandwidth, especially since that 480Mbps of traffic would only be between the cameras and NVR. Easiest way to handle this is to put all the cameras and NVR on one dedicated switch. Connect that switch to the first switch, and put your other PC's, Internet router, etc. all on that first switch. Put everything in the same IP subnet to keep it simple.

Is there a particular reason you're opposed to IP cameras? Basic IP cameras plus a point to point Ubiquiti wireless setup would pretty fairly cheap and easy to implement.

It may not be at your CompUSA, but check out the Apple wireless devices. The Airport Express ($99 in the states) can either create a wifi network, or expand/strengthen an existing network. In the scenario you describe it may be near impossible to have a single device radiate high-power in multiple directions, plus have enough sensitivity to pick up a signal from a laptop or tablet that is NOT a high power device. In the scenario you describe the proper solution is normally multiple access points creating a single network vs. 1 single AP covering the entire footprint.

You could do this with ffmpeg running on a little linux server. Not sure if there is a Windows port or not.