Audio emerges as key element to sophisticated security systems
In this new age of security where analytics and Big Data are the engines that run today’s sophisticated systems, a technology partner once viewed as an afterthought has now taken on a critical role in detection and prevention. Advanced audio platforms have supplanted the ubiquitous intercom and have emerged to handle everything from alarms and verification to gunshot detection, even aggression detection in myriad market sectors.
SouthComm Security Group Editorial Director Steve Lasky sat down for SecurityInfoWatch.com (SIW) with Louroe Electronics CEO Richard Brent at ISC West here in Las Vegas and posed several questions related to technology and industry trends related to the audio world. Here is that discussion.
SIW: Many industry professionals are not familiar with the concept of audio analytics and alarm verification. Please discuss, in detail, some background on the growth of this security solution and what is trending in today’s integrated world.
Brent: Advanced security systems are searching for devices and sensors beyond video that provide better information and intelligence into a security event. When an alarm is activated, operators need to verify the alarm before dispatching a security guard or first responders. The cost of responding can be prohibitively expensive, especially with some municipalities charging fines for being called to a false alarm. Using audio to detect different types of sounds, playing back the sound that triggered the alarm, and pairing with video are all steps to be used to verify alarms. Audio analytics can provide early indicators of an event in progress by identifying different types of sounds like aggressive speech, gunshots, glass breaking, and car alarms. Using a system that integrates audio and analytics alongside more traditional security technologies provides an added layer of information that can be used for detection, verification, and evidence.
SIW: Rudimentary audio alarms systems have been integrated with video cameras for markets like car dealerships and warehouses. But stepping beyond this, where has the concept of audio detection migrated into other markets and how is it integrated with other products?
Brent: We see the adoption rate of audio used with video growing in a number of different markets. The law enforcement market is perhaps the most common. If an officer conducting an investigative interview with a suspect is able to obtain a verbal confession from the suspect, using video by itself isn’t of value to the prosecution. The court will want to hear the suspect making the confession. The same is true of any evidence obtained in the interview such as names, dates, or places. For this reason, we see many law enforcement projects specify the use of audio. The integration of audio can be done either by connecting an IP microphone to the network and letting the VMS handle synchronization with video or by connecting a microphone to a compatible camera’s audio inputs.
Analog systems will connect with wire directly to a DVR where the audio can be paired with video. These days there is not a technology obstacle to adding audio. Years ago we would receive support calls from installers saying that they had a 16 channel DVR but only one audio input. Now we see DVRs with an audio input for each video input. On the IP side, the only limitation is how many switch ports are available. Technology providers are seeing audio integration support become not just commonplace, but essential to providing a more advanced solution.
SIW: Some applications for audio almost border on science fiction when you consider the implantation of advanced audio analytics and aggression detection. How can audio analytics be used in conjunction with other Big Data analytics for physical security applications and what markets are best served? As far as aggression detection, what is the concept behind this and where is it employed?
Brent: We see aggression detection as a technology breakthrough for the security industry. Almost all of the analytics you’ll see at a tradeshow are reactive in nature. This is true for both video analytics as well as many audio analytics like gunshot or glass break detection. A serious event has already taken place by the time the analytic is activated and so the security team can only respond. However, aggression detection works by picking up sound patterns that indicate someone is agitated or has a higher likelihood of committing an assault, an attack, or physical violence. This enables the security team to proactively respond and de-escalate a situation because they have received an early indicator of a potential security threat.
It’s important to note as well that the detector works by analyzing sound patterns and energies at various frequencies. It does not detect words, phrases, or voice identification. Aggression detection has been deployed in various law enforcement facilities where officers want to have early notification of potential physical violence by suspects, inmates, or detainees. It has also been installed in emergency rooms and hospital billing areas – high-stress environments that are prone to outbursts.
SIW: With the advent of more IP-based devices being interconnected, how can devices that were once considered one-trick ponies, like the ubiquitous intercom, benefit from audio analytics when integrated with video and alarm devices?
Brent: The key benefit of using audio analytics is to receive early notification of an event. Video devices are great for capturing visual evidence of an event. However, cameras lack the broad capability of early threat detection and do not provide the capability to respond to an event. Using the example in your question, an analytics-enabled intercom or two-way audio device can hear sound patterns, automatically analyze them for security warnings, and activate an audible alert, siren, or live audio from a monitoring center. It’s one level of deterrent to have cameras installed and visible, it’s an entirely more effective level of deterrent to have a two-way audio system with live speech from a monitoring center instructing a suspect “you in the green shirt, leave the area immediately.”
The combination of analytics, two-way audio, and remote monitoring made possible by IP-based audio analytics expand the capabilities of systems integrated with video and alarm devices, and may lead to faster resolution of security events. Audio technologies enable a monitoring center to intercede before a situation escalates, which is not a capability of video only systems.
SIW: I see where you are showcasing a new solution at ISC West – the new Digifact A. Please discuss the technology behind this launch and what solutions it brings to the table for end users that will enhance their security mission.
Brent: For years we’ve been asked for a network-based audio solution. In our experience, our customers want an easy to install and easy to configure product. Given that most of the industry is video focused and has less experience with audio, we needed to make the new Digifact A product as straightforward to use as possible. The Digifact A has a quad-core processor which means an end user can be running multiple analytics at the same time without any degradation in quality or accuracy. Our intention in using such a powerful processor for audio, which for security applications don’t require much processing power, was so that when end users want to add future audio processing technologies to their security system, they would not need to replace the device but only need to perform a software update. Plus, by running the analytics on the edge, a system requires less bandwidth and processing power on the server. Add to the performance capabilities that the physical installation of the Digifact A is much faster than our analog Verifact A microphone. Connecting CAT5 cable to an RJ45 jack goes much faster than connecting bare wire to a screw down the terminal block. We anticipate that both integrators and end users will prefer using the Digifact A for the above reasons.
SIW: Where is the future of advanced audio analytics headed?
Brent: Different types of sounds are continuously being reviewed and classified for use in future types of analytics. We see growing interest in drone detection, especially for perimeter security applications. We are always working on improving the audio pickup of our microphones so that the input to the analytic algorithms has quality data to process.
Beyond the future of the technology development, we see users adopting audio analytics as a way to receive an early indicator of a threat, and then through their VMS or monitoring center notify those in a given area to evacuate, lockdown, or take an appropriate action given the nature of the threat. Just as DVR manufacturers began adding more audio inputs to match the number of video inputs years ago, we see VMS manufacturers being more resolute in providing two-way audio capabilities, and prioritizing analytics integration support into their own software development.
