HD CCTV: a genuine alternative to IP?

05 Aug 09

The HDcctv Alliance only came into being in May of this year, and it’s planning to release version 1.0 of its interoperability standard in September. That’s a rapid turnaround – and one that has the potential to have a major impact on the worldwide CCTV industry.

The basic proposition is straightforward: replace standard definition cameras and DVRs with high definition-capable equipment, and use currently installed coaxial cable for transmission. Repeaters would be used for cable distances of over 100m. And that’s it.

Of course, there’s additional hard disk space to take into account. And there’s the question of whether the major manufacturers have gone too far down the megapixel IP route to start introducing HD equipment as well.

Security Installer editor Alan Hyder and I sat down with the Alliance’s Todd Rockoff recently to hear what the HDcctv Alliance is all about.

An IP rethink?

Rockoff: “People are rethinking the idea that IP cameras are about to take over the world.

“It’s not only about installers using coax. The user interface of a security system is based on the DVR. And that fact seems to have been forgotten by so many of our colleagues in the industry.

“The user interface is precious. The proposition to a true, coalface security director is: some IT guy walks in and says, ‘The firewall – configure the firewall and the security searching protocols this and that, and I’m gonna throw away your DVR, and you can have IP!’

“And the security guy says, ‘What are you going to do, cut off my arm, and go through this headache – why?’

“With HD CCTV, you have an alternative. 720p corresponds to megapixel, 1080p corresponds to 2.2 megapixel. That preserves the user interface, the usage paradigm.

“So it’s a zero training upgrade. The only difference perceived by the operator of the security system is when he’s configuring the recording settings for an alarm – the list of resolutions on the pull-down menu has some bigger numbers on it.

“Otherwise, it’s still a DVR. It smells, it feels like security stuff.”

A price to pay

He explains that there are some costs associated with the HD upgrade.

Rockoff: “You would have to buy a new DVR. With higher resolution there are greater costs. And that’s true in the IP world – there’s more data. However, it’s established, certainly for standard definition, that the DVR is a more cost effective way to achieve storage and networking than it is to put all the Moore’s Law-susceptible stuff into the camera.

“In fact, people replace DVRs regularly anyway as they move from Motion JPEG, to MPEG, to H.264. It makes sense. So there’s Moore’s Law churn, if you will, in the DVRs.”

Obviously standard definition cameras will need to be upgraded to HD, and they will be more expensive than current cameras. But…

The long haul

Rockoff: “Cameras need to have glass, to focus an image, they need to have an image sensor, to sense the image, and then some kind of processor, to render the image, and something to transmit the image. That’s the minimum thing that a camera should be. They can be good for five to ten years.

“As soon as you start putting secure communication protocols in that, let alone IP v7.0 or whatever, you’re guaranteeing, if you believe Moore’s Law continues apace, that within 18 months to two years, that’s an obsolete piece of equipment. Oops.

“Your amortization interval just took a nosedive. Which directly means that your total cost of ownership has multiplied, by a factor of several.

“So if an IP camera is a factor of three more than its corresponding resolution conventional camera, and the amortization interval is a factor of two shorter, pretty soon you’re looking at an order of magnitude lower price performance or cost effectiveness in the IP cameras.

“And so that’s been a frustration and that’s one of the reasons that the IP cameras have been so slow to take off, and that has not been addressed by the proponents in the industry.

“As I talk to people, these debates become almost religious. The IP camera salesman is saying, essentially: “It’s digible, stupid.” And that’s the sale.

“And you can’t intimidate people into buying this stuff, y’know: ‘You’ve got to stay with the latest greatest thing’. Well I’ve got news for you – High Definition CCTV is the latest greatest thing.

“It’s digital, it’s high tech, but it’s not expensive.

“Let’s come back to the storage point. For everybody the cost of storage is going down. HD compared to standard definition, for example, presents three to 12 times as much data per channel. 1080p60 is 12 times the number of pixels per second as compared to PAL video.

“The point about the way DVRs work is you’re not obligated to store all that. But you’ve got it all there in case you want to store it – if there is an alarm situation, you can store the full resolution. Also it’s there in the DVR for analytics purposes.”

Cheaper than IP?

In fact, Rockoff predicts, the cost premium for HD CCTV cameras over standard definition is “somewhere between negligible and modest”.

“To have a good quality high def camera, the end price could be $50 more. Hey, it could be even $100 more than the corresponding standard definition camera – but that’s still cheaper than a standard definition IP camera.

“It will run over existing coaxial cable to 100m.

“For longer distances a repeater is used. If the coax is baked into the plaster, then that signal won’t be amenable to the upgrade.

“This is not the same as an IP upgrade. An IP upgrade is a forklift upgrade. Everything has to change.

“With high def, all the high def DVRs preserve the install base. In other words, they accept standard definition inputs, and they accept IP inputs – why not, right?

“But the point is, any coax cable which has sufficient signal integrity is a candidate to upgrade that camera to a higher resolution.

So you can put in your HD DVR, and let’s say only 30 per cent of the cameras [in a very complex installation] can take advantage of the upgrade. That’s fine. Thirty per cent of your cameras can now go from being a little PAL thing to HDTV, Blu-ray quality, coming back to the DVR.

“To the extent that one is willing to go to a little more inconvenience and add repeaters, you can upgrade more of those channels.”

Learning from broadcast

Rockoff says that HD CCTV is broadcast industry standard compliant.

“So it really takes advantage of the consumer trends that made displays so cheap. There’s a time honoured tradition of innovations for consumer TV finding their way into security.

“Be it the Videcon tubes, that were used as the first image pick up devices – they made it into security. Then of course there were CCDs. And every step along the way the security guys layer in different requirements, for example up the cable control protocols, ways of coding things in the vertical re-trace interval…

“So what we’re doing in HD CCTV is similar. We’re building on top of that SDI technology and then putting on metadata where the sender can identify who he is as a unique identifier or what type of product he is. Also a 720p camera sending 15 frames per second, for example, so the receiver can know how to reconstruct the video content from that.

“We’re also adding into our standard up the cable control, bi-directional audio, and up the cable power, so it is a one cable solution.”

Controlling the dome

He says one of the advantages is the lack of lagtime in controlling PTZ domes.

Rockoff: “With the control of PTZ domes, the signalling delay is tens of microseconds, which is not perceptible to a human. If you’re doing motion based compression, by definition the algorithm considers a group of pictures, typically like 15 in size, before it outputs its first frame. Because it’s looking backwards at the frames that went before.

“So that 15 frame time – that’s like half a second. And a human definitely notices that. If someone’s trying to track a suspect or an interesting person, in the lobby of an airport or through an aisle of a store, using the joystick – you just can’t do it with an IP-based speed dome.

“The vendors in that area again say, fine, change your user paradigm, and now start clicking on a mouse to preset targets, but that’s not the same as the smooth control that people have come to expect and enjoy.

“It’s a pain in the neck to focus an IP camera, let alone use it in anger.”

Analyse this

Are there compatibility issues with existing analytics software?

Rockoff: “There are no fundamental compatibility issues. What the analytics guys seek is high resolution, clean input data. So this plays right into their hands.

“Analytics do quite well today in the DVR. So HD CCTV is a path to extend that for 10 years or more.”

With big manufacturers having invested so much in IP, both in time and money, is the take up there going to be difficult?

Rockoff: “Well, that’s the nature of a disruptive innovation. It disrupts the competitive landscape. So yeah, there’s going to be a lot of guys who say, we’ve invested so much in the IP camera, we’ve just gotta go down that path to its inevitable conclusion. There will be people who will say that.

“There will be other people who may be selling IP cameras, who stand up and say, you know what – we’re re-thinking the whole value proposition, the return on investment proposition for these IP cameras, and here’s an alternative. And if we embrace an alternative we’re going to build our brand.

“The guys who understand the merits and take advantage of it are well position to outcompete those who are slower to come to the table.

“There are two scenarios here as I see it – and I’ve been around the world and I’m waiting for someone to tell me otherwise.

“One scenario is that some big guy – a GE, a Bosch, a Pelco – comes on board early.

“And the other scenario is that they don’t, and what happens is that Tier 2 guys, or Tier 3 guys – smaller guys, like Cop UK – get the opportunity to start selling high def equipment, and build their brand.

“CSST [China Security and Surveillance Technology] are huge, they’re like $500m, which is more than half of Pelco, and they own a whole bunch of brands. They’re in this with both feet, and their hands.”

HD standard

The HDcctv Alliance plans to release v1.0 of its interoperability standard next month, with full product launches taking place by April or May of next year.

“The DVR has to have HD inputs, the camera has to be HD because it’s driving the signal over the coax. In fact, there’s an opportunity for HD monitors as well.

“There are not yet DVRs available which can be used. The PC-based DVRs will be in mass production in September.

“The standalone or embedded DVRs will be early next year, maybe March, April, May timeframe for mass production.

“This is a new product category. So that’s the challenge here.

“No one can really sell high def cameras, without there being high def recorders. And repeaters. Quadders – four in, one out, for quad view. Fibre concentrators.

“The equipment exists, but it’s not standard compliant, because it’s like a precursor – it’s using the basic SDI technology – but it will be soon.

“To make it plain: we’re launching the product category in September, with a sketchy set of stuff which covers the whole spectrum, but you don’t have every flavour of lens, and all the cameras, that sort of thing.

“But you will have every flavour and lens and the cameras within 18 months.”

Not for nerds

Rockoff says there is still a reluctance amongst installers to get involved with IP cameras and networking. And he’s got an interesting possible explanation as to why…

Rockoff: “In the end market, I get the sense that… When I was in school, I don’t know about you, but when I was in third grade, I was the guy who wanted to thread the film in the projector, and I was the guy in the sixth grade who was writing Star Trek computer programs and stuff; and I’m the kind of guy who goes to Carnegie Mellon and gets PhD in Computer Science – and I’m the kind of guy who’s out there selling IP cameras, right?

“Well, the guys who install security equipment, they were the guys who were, like, throwing spitballs at me in third grade, and they were out playing football, and dating the girls… They don’t wanna have to be geeks.

“So there is a sense of, IP cameras are for nerds. And real security guys don’t do this.”

World domination

He claims that in four years’ time there could be as many as 15 million HD CCTV cameras in worldwide use.

“The numbers that are out there are reasonable – about 40 million security cameras sold each year, with about 90 per cent of that, today, being analogue stuff.

“The numbers that are in that discussion, in about 2013, have 55 per cent of the market still as standard definition analogue, 15 per cent IP – growing from maybe 10 per cent today – and the remaining 30 per cent: HD CCTV.

“In as short a time as four years. So with the market growth, that might be 15 million, or maybe more, HD CCTV cameras. And more importantly perhaps, twice as many HD cameras as IP cameras.

“So now you can deliver a broadcast-quality experience in a security environment.”

The Alliance – a potted history

Rockoff: “I began working in the Alliance as my full-time job on the 9th of May. Before that I was the EverFocus VP of global sales for the ODM business unit. That is, using the EverFocus factory and support structure for Tier 1 global brands.

“Before that I was at Pixim as vice president of sales for Asia. and prior to that I had various jobs, from director of sales to CEO of semiconductor companies, dealing with pixel sensors, image processors, in Silicon Valley.

“I transitioned there out of academia I guess back in ’95.

“We have six full fledged members. In fact I’ve made an agreement with COP UK to join. There’s a cooperation agreement and a licence agreement – to use the trademark – and we’ve only just got those from our lawyers, so we’re only really just getting started.

“But meanwhile, the members of the Alliance, from the middle of May, have been moving very rapidly on version 0.9 of the standard, and the draft of version 0.9 is now complete.

“Member companies are starting to test the interoperability of equipment using version 0.9 with a view to releasing it as version 1.0 at the end of September.

“The members that are already in – there’s a Silicon Valley company called Stretch, who make codecs; there’s a Canadian company who makes chips for transmission, called Gennum; there’s a Taiwanese manufacturer of cameras and DVRs, EverFocus; there’s a Korean manufacturer of equipment called Comart; and now we have CSST, China Security and Surveillance Technology.

“They’re China’s biggest security company. They make everything. So that means that they mostly sell conventional CCTV, because if you believe the market figures, IP cameras are less than 10 per cent of the market.

“So now we have COP UK, and Pixim joined, so that’s another Silicon Valley chip company, image sensors.”

The other current member is imaging solutions company Ovii. Rockoff says he is confident that other image sensing companies will also be joining in some form.

Getting involved

Rockoff: “As Pixim salesman I sat down with Craig Scott, who is a legend in the industry – he’s a force of nature, terrific guy to work with – and he made the observation about that critical difference between IP video, and IP camera.

“This is years ago now. He was a camera maker, and he pointed out to me, he says: I’m never going to make IP cameras. Why not?, I said. I was trying to sell him Pixim technology at the time, which is really good for IP camera, among other things.

“He says, because there’s a difference between IP cameras and IP video and there’s no case for IP cameras in general. So I took that on, and I get it.

“In other words, within a premises like this [London hotel] where you typically have less than a hundred metres of cable, to pre-compress the video information and then to decompose into packets and then to inject those packets into an occasionally congested network, doesn’t make scientific sense for a stream that’s continual, 24/7. It never stops.

“On the other hand, to get out of the premises you want to remotely store some of the video, remotely control it, observe what’s going on from anytime, anywhere, that kind of idea – yeah, leverage IP. That’s an infrastructure that’s planetary in nature.

“So there’s the closed circuit, secure premises, and then there’s the kind of more open, integrated over distance enterprise. Those are different requirements.

“A term that’s bandied about by computer scientists in the security industry is ‘the intelligent edge device’ – you hear that a lot. I personally have a big problem with the word ‘intelligence’ – people are intelligent, designed systems are not.

“So that’s not just recording in the camera, but analytics in the camera, or everything it takes to be a web server, in the camera. But it’s a theorem, that by putting everything inside the camera, because it is susceptible to Moore’s Law, not only does it increase the power consumption and therefore decrease reliability and increase the cost, it also shortens the lifespan of the camera. So that’s a lot of bad stuff, for dubious gain.”

Getting the word out

Rockoff: “Then, when I was with EverFocus, I met with the guys from Gennum, the link company. A little about the link: Serial Digital Interface is broadcast industry standard. All the high def content in the broadcast industry is carried in the studio over coax cable. I think it’s RG6 – it’s a very high grade of cable. Interface-wise it’s the same as RG59 which is what we use in security.

“Broadcast is all about very high quality, about 3 gigabits a second serial digital interface to convey the high def image. But then the guys from Gennum said ‘Do you know what? We can make this very cost effective for security, in the order of 5, 10, 15 dollars chip cost, per link, because the volumes are so much higher in security.’

“At the same time I had some Tier 1 guys say to me, if there’s any way to get high resolution for our install base, lets us know about it, because that’s a real pain point. So we put that together.

“When I was at EverFocus I started working on High Def CCTV – just really the concept. And I remember debating the concept with several Tier 1 manufacturers.

“And what they came back with was ‘This is interesting, and from the fundamentals it seems like you’re actually correct.’

“However, it doesn’t matter, they said, as recently as a year ago. It doesn’t matter that you’ve identified this approach which is fundamentally superior to megapixel IP cameras, because we believe the transition to IP cameras will have concluded before any HD CCTV equipment can make it to market.

“That was the statement then.”

Tenacious Todd

Rockoff: “But I kept working on it. And we sat down this year at ISC West – EverFocus, in our ODM meeting room, had the industry’s first HD CCTV system demo.

“We showed unadulterated live view of a 1080p30 camera. We showed the ability to lay up standard def and high def images properly on a high def display. We showed the ability to record and play back high def, and people saw that and said: ‘Oh my god, when can I get it?’ And that was it. It was so easy to sell.

“So we had this system level demo at ISC West, and some of the charter members just happened to be in the room: EverFocus, Stretch, Craig Scott’s company Ovii, the guys at Gennum – and we just sat down and said we need to have some kind of industry group – an alliance.

“Let’s call it the HDcctv Alliance, where we all get together. The real thing is we have to create the standards, so that there’s no risk to the equipment makers. And we have to promote the value of HD to the end market, again to minimise risk, so once you have the equipment, people will readily take it.

“I’m finding that it’s very easy to sell – but the word has to get out.

“After ISC I went away and thought about it and I said: you know what? I’m enjoying working for EverFocus and being a salesman, but I think high def is gonna be huge. It was coming together for me, so I said I’m going to invest myself in high def and see where this takes us.

“So I started the Alliance.

“The guys who are in at this point are all strategic – in other words, they haven’t been hung up on their attorneys reviewing the documents, they’re just in. But going forward of course we’ll have a more formal structure.

“Some people are amazed. Here we are with version 1.0 published in September. We started in May. So basically four months from when we really started – four months from nothing to having a full blown spec and a full blown alliance – that’s unprecedented.”

Get the latest stories first with info4security's newsletters: Click to signup

Post and bookmark this story at the following sites:

What is this?

del.icio.us	Digg	Reddit	Google	Newsvine