Dynamic Video Encoding
Demand for ever higher resolutions is something that is driven by forces external to the security industry. Where once 4CIF and D1 images were more than acceptable, today user expectations are for HD and increasingly 4K UHD video. There’s nothing new in this; end users have always wanted higher resolutions and real-time video. Because older technologies featured compromise because of inherent limitations never meant that customers were happy with what they got. Today things are very different, but that does have an impact on bandwidth. Benchmark looks at one smarter solution: dynamic video encoding.
Talk to many of the self-proclaimed experts about networked video surveillance, and it won’t be long before the subject of bandwidth management comes up. There are many theories about the best way to manage the load on a network, but the reality is that in order to minimise bandwidth requirements, you have to reduce the amount of data being transmitted. That’s somewhat ironic, given that the reason for investing in HD and 4K UHD video is actually to get data!
While the increases in video resolution certainly add to the challenge, the bigger consideration is that video pumps out 25 (or increasingly 50) frames per second, every second of every minute of every hour of the day. With streaming video, there is no respite, no lull, and whilst modern technology makes event-based recording a simple reality, a huge number of applications are still streaming video around the clock.
Historical approaches to bandwidth management are techniques that were used in the days of analogue recording to enable archiving of multiple video feeds to a single magnetic tape. The challenge back then had nothing to do with bandwidth. The VCR could only record a single input of video, so multiplexing was used to effectively create a single stream from multiple cameras.
Time-lapse and field recording were commonplace and – to a point – accepted, but in today’s industry we can acknowledge what they really are: reduced frame rates and reduced resolutions. For some reason, these flawed techniques which cause obvious degradation to video quality form the core strategy of many peoples’ approach to bandwidth management.
There is a major issue with the idea of reducing frame rates or resolution where HD or 4K UHD video has been specified. Both video formats are established standards which have been created outside of the security industry, and which all stakeholders in the surveillance sector – installers and integrators, specifiers, consultants, manufacturers and even industry bodies and associations – have no authority to change. It simply is not acceptable for anyone to deviate from the standard if a user has specified the use of HD and/or 4K UHD video.
An intelligent approach
Many people think in terms of bandwidth reduction, but that’s not the issue. The first thing to point out is that considerations should really be about bandwidth management. The goal is to use the available resources to ensure the delivery of the best quality video possible when it is required. In short this means we’re aiming to achieve the highest resolutions and highest frame rates from the allocated bandwidth for important security video.
By way of an example, if one zone of a site had a bandwidth budget of 50Mbps and included 10 HD cameras, common practice would be to ‘share’ that budget out between devices. In truth, 5Mbps is a bit tight for an HD1080p stream. If an incident happens in a busy scene, that level of bandwidth isn’t going to deliver images which provide the highest evidential quality.
What has to be considered is that typically, such an approach would also see nine streams from that zone of nothing important. In fact, 90 per cent of the bandwidth budget is being used for nothing of importance.
The challenge is to ensure that cameras delivering images with no great importance use less bandwidth, thus allowing detail and quality to be preserved for devices which are capturing video of interest. Because of the unexpected nature of security, it’s impossible to know which cameras will capture critical video at any given time, so the issue requires a smart solution.
Thankfully, the recent increases in video resolution coincided with a leap forwards in processing capabilites, and this spare capacity allows the implementation of dynamic video encoding. The result is that many of today’s higher resolution cameras have spare processing power, and some innovative manufacturers are using it to address the bandwidth issue.
As with any processing functionality, manufacturers have different names for the process, and there are a host of features with monikers alluding to smart and intelligent streaming. Probably the safest terminology to describe the processes is dynamic encoding.
Whilst there will be differences in each vendor’s implementation, the concept of dynamic video encoding allows a forensic engine to examine captured video streams. The goal of dynamic video encoding to identify static parts of the scene and those with a high degree of change and/or detail. The process is not too dissimilar to that carried out by predictive compression engines.
Areas with low detail and no change are considered to be unimportant. After all, nothing much is happening. As a result, higher levels of compression can be applied. What’s important is that video isn’t removed or compressed to the point of being unusable. The quality will still be acceptable and for proof of continuity it will be fine.
Where change is detected and high levels of detail exist, the algorithm sees the video data as important and therefore increases the available bitrate. This ensures that important parts of the video retain detail and sharpness. As a result, video containing faces, vehicle number plates, identifying elements such as tattoos or logoed clothing, etc., will be preserved and streamed with a higher bitrate. However, video of an empty car park can be streamed with far less load on the bandwidth.
Because the decoding is dynamic, it reacts to the scene content. If someone walks into the car park, or if a vehicle enters of leaves, the bitrate can be increased to ensure the video can be used for identification and evidential purposes.