Megapixel and H.264: Made for each other?

| 0 Comments | 0 TrackBacks
As more models hit the market, megapixel cameras are gaining greater interest and attention from users and integrators. At the same time, more and more vendors are incorporating H.264 compression into their equipment.

Both technologies are instrumental in driving the trend toward network-centric security. Megapixel technology, in particular, while greatly enhancing the surveillance utility of CCTV systems, adapts them to fit the changing requirements of enterprise-level security, namely by gathering more data that can be used in the support of strategic and pro-active business processes.

Megapixel cameras, available from vendors such as Arecont Vision, Axis Communications, ACTi and IQinVision, have been shipping for more than twelve months. Their primary value proposition is that the increase in resolution they offer dramatically increases their range and field of vision. As users mull the cost of a digital to analog conversion, megapixel means analog cameras do not have to be replaced one-for-one. In some installations, the ratio can be one megapixel camera to three or four standard cameras. Michael Hodor, vice president of sales-North America for Arecont Vision, recalled a Wal-Mart installation where 7 megapixel cameras replaced 26 analog VGA cameras.

But megapixel cameras create a problem in that their image information generates huge video files that can tax bandwidth and storage capacities. An 8-megapixel camera can produce one terabyte of data every three days. Enter H.264, the video compression algorithm used in the broadcast industry and now being adopted in surveillance cameras and video management systems. While H.264 will work with any video format, it may prove invaluable to the economy of megapixel video. Simply put, it can effectively reduce data volume five to six times on average. Yet, because of the much higher amount of data contained in a megapixel image, clarity is not sacrificed for compression.

"When you start to do the math, you come out on top. Now with H.264, with the reduction of bandwidth and storage that happens, it's easy to make the justification," Hodor said. "It really is changing the game in terms of storage," Hodor said. "If a VGA camera is about 2.5 Mb/s at a high frame rate and I can take a 2.5 megapixel camera and run it at H.264 and be in the same ballpark--2.5 to 3 Mb/s--that changes everything."

"Money you save in bandwidth and storage is more than what you save with standard cameras and compression," added Paul Bodell, chief marketing officer with IQinVision.

Still, there are some variables to keep in mind. H.264 is an algorithm for dynamic compression. The more static the image is, the higher the compression rate will be. Movement, either in the image itself, or through a camera's pan, tilt and zoom operations, reduces the compression rate. That makes H.264 cameras excellent in environments where there is a steady unchanging image: Hallways, doors and parking lots, especially at night, as H.264 also functions well in low-light conditions. Their efficiency and savings decrease if used in environments where there is significant movement - public areas such as airports and train stations, building lobbies and major entrances and exits.

When all the of the tools that the H.264 compression standard supports are implemented, it offers 30 to 50 percent savings in bandwidth and storage compared to MPEG-4, depending on the level of scene activity, said Alastair Hayfield, an analyst with IMS Research.

"H.264 compression is an important technology for megapixel video surveillance camera vendors," Hayfield said. "In conjunction with better low-light sensitivity, megapixel lenses, IP video standards, and lower price points; it will strengthen the value proposition of megapixel video surveillance."

But H.264 is very processor-intensive, Hayfield noted, and different vendors implement different H.264 codec profiles. Although H.264 is a standard, the standard itself sets forth a series of profiles--a defined set of features for the codec to meet a certain set of specifications of intended applications. Choice of profile involves cost and feature trade-offs, so actual improvement in performance over MPEG-4 can vary among H.264 cameras.

In addition, because of H.264's dynamic use of bandwidth, IQinVision's Bodell recommends users employ variable bit rate (VBR) transmission over their IP and Ethernet networks. "VBR will preserve quality," Bodell said, because it will increase available bandwidth should the bitstream require it. A committed or fixed bit rate, on the other hand, "will degrade the image if the bandwidth gets congested" because it would throttle down traffic bursts, Bodell said.

Megapixel, H.264 and VLM

TimeSight Systems is attempting to leverage megapixel and H.264 through its concept of video lifecycle management (VLM). TimeSight manufactures NVRs that reduce the size of video files over time, taking them from highest information in the first few days of recording--the high-risk period when surveillance video is likely to be most needed--and gradually reducing the file size over time and as the risk period declines. TimeSight claims to be unique in applying a two-fold process to VLM-dynamic H.264 compression-over-time and motion optimized recording, where, as the name suggests, more storage capacity is reserved for images containing objects in motion.

Charles Foley, TimeSight CEO, says this stands to be a major solution to the storage requirements the proliferation of video presents. There are more than 100 million surveillance cameras in use worldwide, he said, and shipments double each year. He cited data that forecasts 65.6 million cameras will be deployed in 2013. Three related factors are at work, he said. More cameras are being deployed; video data is increasing because of all these new cameras; plus, there is more data generated per camera than ever before.

Whether you use TimeSight's approach to VLM or not, the case for megapixel plus H.264 is compelling from a general storage and quality standpoint. From a purely apples-to-apples approach, a 40 kb megapixel H.264 compressed file is superior to a 40 kb VGA H.264 compressed file.

"You've got more pixels to work with," Foley explained. "Think of trying to capture an image of my face. If you have a hundred points to work with, that becomes a standard of quality. What if you have a thousand points to work with? You have much more granularity. Even if you compress it, you're not giving up the points or the pixels. You're giving up the degree of differentiation between them in terms of color, light or shading. So by having more pixels available to you, you've got a finer grain to work with as you increase compression."

A two-megapixel image, Foley said, can be half the size of a VGA image and still get much better clarity because there is six times as much data to work with at the outset.

"Bookend appeal"

Megapixel cameras and storage have "bookend appeal" in a sale, and integrators can use that as an advantage.  "When people start seeing what cameras can do," said Hodor, "wheels start turning." Arecont will often bring a storage vendor in as part of a proposal, he added.

In some cases, H.264 can also lighten the load on central processing units (CPUs) in NVRs and video management systems. For Arecont Vision, this is a competitive differentiator. "With H.264, [users] are used to seeing a reduction in image quality and an increase in CPU demand by virtue of the processing that's required and the decompression of that image," said Hodor. That is not the case with the Arecont line, he said, because of the efficiency of the algorithms working in the camera.

Hodor points to CPU usage data from partner VMS equipment. With Exacq Technology's systems, CPU usage with Arecont Vision's H.264 cameras dropped to 32 percent, compared to 96 percent usage with MJPEG. With OnSSI's VMS, CPU use decreased to 23 percent with Arecont Vision H.264 from 42 percent with MJPEG.

Although the networking concepts seem complicated, the advantages of integration tend to become clear once digital components are in place. The toughest barrier, Hodor said, is getting the analog-to-digital transition started. "The biggest obstacles we have are people's understanding of what this technology means and their ability to migrate over to this technology, not just megapixel, but the move from analog to IP," Hodor said. "It's really not that much different when you get right down to it. A network is just a network. We've all been on a network for decades. For more than a hundred years we've had telephones. This is something people just have to conceptualize and understand what can be done."

As more models hit the market, megapixel cameras are gaining greater interest and attention from users and integrators. At the same time, more and more vendors are incorporating H.264 compression into their equipment.

Both technologies are instrumental in driving the trend toward network-centric security. Megapixel technology, in particular, while greatly enhancing the surveillance utility of CCTV systems, adapts them to fit the changing requirements of enterprise-level security, namely by gathering more data that can be used in the support of strategic and pro-active business processes.

Megapixel cameras, available from vendors such as Arecont Vision, Axis Communications, ACTi and IQinVision, have been shipping for more than twelve months. Their primary value proposition is that the increase in resolution they offer dramatically increases their range and field of vision. As users mull the cost of a digital to analog conversion, megapixel means analog cameras do not have to be replaced one-for-one. In some installations, the ratio can be one megapixel camera to three or four standard cameras. Michael Hodor, vice president of sales-North America for Arecont Vision, recalled a Wal-Mart installation where 7 megapixel cameras replaced 26 analog VGA cameras.

But megapixel cameras create a problem in that their image information generates huge video files that can tax bandwidth and storage capacities. An 8-megapixel camera can produce one terabyte of data every three days. Enter H.264, the video compression algorithm used in the broadcast industry and now being adopted in surveillance cameras and video management systems. While H.264 will work with any video format, it may prove invaluable to the economy of megapixel video. Simply put, it can effectively reduce data volume five to six times on average. Yet, because of the much higher amount of data contained in a megapixel image, clarity is not sacrificed for compression.

"When you start to do the math, you come out on top. Now with H.264, with the reduction of bandwidth and storage that happens, it's easy to make the justification," Hodor said. "It really is changing the game in terms of storage," Hodor said. "If a VGA camera is about 2.5 Mb/s at a high frame rate and I can take a 2.5 megapixel camera and run it at H.264 and be in the same ballpark--2.5 to 3 Mb/s--that changes everything."

"Money you save in bandwidth and storage is more than what you save with standard cameras and compression," added Paul Bodell, chief marketing officer with IQinVision.

Still, there are some variables to keep in mind. H.264 is an algorithm for dynamic compression. The more static the image is, the higher the compression rate will be. Movement, either in the image itself, or through a camera's pan, tilt and zoom operations, reduces the compression rate. That makes H.264 cameras excellent in environments where there is a steady unchanging image: Hallways, doors and parking lots, especially at night, as H.264 also functions well in low-light conditions. Their efficiency and savings decrease if used in environments where there is significant movement - public areas such as airports and train stations, building lobbies and major entrances and exits.

When all the of the tools that the H.264 compression standard supports are implemented, it offers 30 to 50 percent savings in bandwidth and storage compared to MPEG-4, depending on the level of scene activity, said Alastair Hayfield, an analyst with IMS Research.

"H.264 compression is an important technology for megapixel video surveillance camera vendors," Hayfield said. "In conjunction with better low-light sensitivity, megapixel lenses, IP video standards, and lower price points; it will strengthen the value proposition of megapixel video surveillance."

But H.264 is very processor-intensive, Hayfield noted, and different vendors implement different H.264 codec profiles. Although H.264 is a standard, the standard itself sets forth a series of profiles--a defined set of features for the codec to meet a certain set of specifications of intended applications. Choice of profile involves cost and feature trade-offs, so actual improvement in performance over MPEG-4 can vary among H.264 cameras.

In addition, because of H.264's dynamic use of bandwidth, IQinVision's Bodell recommends users employ variable bit rate (VBR) transmission over their IP and Ethernet networks. "VBR will preserve quality," Bodell said, because it will increase available bandwidth should the bitstream require it. A committed or fixed bit rate, on the other hand, "will degrade the image if the bandwidth gets congested" because it would throttle down traffic bursts, Bodell said.

Megapixel, H.264 and VLM

TimeSight Systems is attempting to leverage megapixel and H.264 through its concept of video lifecycle management (VLM). TimeSight manufactures NVRs that reduce the size of video files over time, taking them from highest information in the first few days of recording--the high-risk period when surveillance video is likely to be most needed--and gradually reducing the file size over time and as the risk period declines. TimeSight claims to be unique in applying a two-fold process to VLM-dynamic H.264 compression-over-time and motion optimized recording, where, as the name suggests, more storage capacity is reserved for images containing objects in motion.

Charles Foley, TimeSight CEO, says this stands to be a major solution to the storage requirements the proliferation of video presents. There are more than 100 million surveillance cameras in use worldwide, he said, and shipments double each year. He cited data that forecasts 65.6 million cameras will be deployed in 2013. Three related factors are at work, he said. More cameras are being deployed; video data is increasing because of all these new cameras; plus, there is more data generated per camera than ever before.

Whether you use TimeSight's approach to VLM or not, the case for megapixel plus H.264 is compelling from a general storage and quality standpoint. From a purely apples-to-apples approach, a 40 kb megapixel H.264 compressed file is superior to a 40 kb VGA H.264 compressed file.

"You've got more pixels to work with," Foley explained. "Think of trying to capture an image of my face. If you have a hundred points to work with, that becomes a standard of quality. What if you have a thousand points to work with? You have much more granularity. Even if you compress it, you're not giving up the points or the pixels. You're giving up the degree of differentiation between them in terms of color, light or shading. So by having more pixels available to you, you've got a finer grain to work with as you increase compression."

A two-megapixel image, Foley said, can be half the size of a VGA image and still get much better clarity because there is six times as much data to work with at the outset.

"Bookend appeal"

Megapixel cameras and storage have "bookend appeal" in a sale, and integrators can use that as an advantage.  "When people start seeing what cameras can do," said Hodor, "wheels start turning." Arecont will often bring a storage vendor in as part of a proposal, he added.

In some cases, H.264 can also lighten the load on central processing units (CPUs) in NVRs and video management systems. For Arecont Vision, this is a competitive differentiator. "With H.264, [users] are used to seeing a reduction in image quality and an increase in CPU demand by virtue of the processing that's required and the decompression of that image," said Hodor. That is not the case with the Arecont line, he said, because of the efficiency of the algorithms working in the camera.

Hodor points to CPU usage data from partner VMS equipment. With Exacq Technology's systems, CPU usage with Arecont Vision's H.264 cameras dropped to 32 percent, compared to 96 percent usage with MJPEG. With OnSSI's VMS, CPU use decreased to 23 percent with Arecont Vision H.264 from 42 percent with MJPEG.

Although the networking concepts seem complicated, the advantages of integration tend to become clear once digital components are in place. The toughest barrier, Hodor said, is getting the analog-to-digital transition started. "The biggest obstacles we have are people's understanding of what this technology means and their ability to migrate over to this technology, not just megapixel, but the move from analog to IP," Hodor said. "It's really not that much different when you get right down to it. A network is just a network. We've all been on a network for decades. For more than a hundred years we've had telephones. This is something people just have to conceptualize and understand what can be done."

No TrackBacks

TrackBack URL: http://www.securitysquared.com/cgi-bin/mt/mt-tb.cgi/107

Leave a comment