Audio Analytic Company Overview

Published Nov 03, 2010 00:00 AM
PUBLIC - This article does not require an IPVM subscription. Feel free to share.

In this note, we overview the products and positioning of Audio Analytic, a UK based startup developing sound analytics for use in security applications.

Audio Analytic offers two versions [link no longer available] of their sound analytics: Full and Lite. The Full version provides sound classification while the Lite version delivers more basic loudness detection. The Lite version is positioned for low noise environments, "comparable with other level detection systems on the market place." The Full version targets specific sound detection in areas where "sounds are likely to occur which our equally or greater in loudness to the specific sound of interest which would otherwise trigger a loudness detection systems."

For the full version, Audio Analytic provides various 'sound packs' that match specific sounds. Examples of sound packs/applications include:

  • Monitoring aggression - ‘Hospitals, Prisons, Petrol Stations and Convenience Stores’
  • Monitoring car alarms - ‘large car parks attached to hospitals, airports, train stations, shopping centres, office and residential parking’

Computing load for the analytics is likely to be fairly low (especially relative to historical needs for more demanding video analytic applications). Using the Axis ARTPEC-3 as a baseline, the full version is estimated to consume 5-15% of the processor. Using a 2 core i7 Intel chip, about 100 channels of the full version can be processed.

Their sound analytics do not record audio nor play it to users. Sound is analyzed as a trigger to send an alert to a security monitoring system.

In the video surveillance market, available sound analytics are limited. One good example is Mobotix's built-in audio analytics which we used briefly and seems to work well for basic sound level detection. However, we've seen very few higher end sound analytics offered with surveillance. Gun shot detection, such as SpotShotter is a notable and prominent exception.

Audio Analytic's products are developed but not yet deployed in production products. (They are in the news recently because of their October 2010 Series A funding announcement [link no longer available].) Audio Analytic seeks to license their analytics to camera and VMS/NVR.

Given their stage and business model, pricing has not been set. We see 2 potential structural options:

  • Selling as a premium add-on (like video analytics is usually done today). The Full version would cost a few hundred dollars for specialized applications that want to use them.
  • Bundled by a camera manufacturer and provided to customers for free. Camera manufacturers could offer this on their higher end models as a differentiator. We are seeing this currently as a tactic by a number of larger manufacturers with analytics (e.g., Pelco and Sony).

We do think sound analytics has the potential to provide incremental value for applications where video analytics would be difficult (e.g., detecting aggression or car alarms). Obviously, this is early stage so more definitive judgement awaits further development and production release.