Torque Video Systems Integrates New Transcoder And Adds AI

Singapore-based broadcast T&M solution manufacturer makes two significant announcements at IBC.

Torque Video Systems recently announced a business and sales distribution channel partnership with Vancouver-based NETINT Technologies, an innovator of computational storage and video processing system-on-a-chip (SoC) solutions. Under the agreement, Torque will integrate NETINT’s Codensity Video Transcoders into its DVStor product line. In addition, Torque will look after sales and marketing activities of NETINT Technologies’ products across Asia.

The DVStor already provides exceptional value for large volume recording of multi-stream digital content. Integrating the ASIC-based H.264/AVC and H.265/HEVC Codensity Video Transcoders into the DVStor enables new applications for disaster recovery playout, content repurposing and media analysis.

Codensity Video Transcoders offer high quality, scalable encoding for live video streaming, through a unique combination of ASIC-based encoding and the scalability of NVM Express (NVMe) cloud storage infrastructure. NVMe is also known as Non-Volatile Memory Host Controller Interface Specification. The transcoder supports up to 4K UHD resolution at a much lower price point compared to alternative encoding architectures. Modules easily plug into any NVMe-based server. The solution includes open-source FFmpeg compatible SDK, allowing service providers using FFmpeg to achieve quick and significant encoding capacity upgrades.

Torque Video Systems At IBC

Torque Video Systems announced it will launch a facial recognition module to its flagship DVStor Recording and Playout system at IBC 2019. Leveraging AI Deep Learning technologies, the new product performs rapid identification and cataloging of human facial features based on facial geometry metrics. Identified faces are stored in a database featuring a straightforward, easy-to-use operator GUI.

The massive capacity and capability of the DVStor can record and store hundreds of channels for months, providing a rich database of detected faces, and the video clips they were detected in. On the web-based GUI, detected faces are cross-referenced against video clips, allowing quick search of video clips containing one or more people. The GUI also facilitates adding metadata for complete identification.

Danny Wilson, Founder and CEO of Torque Video Systems said, “Even just a short few years ago, this kind of technology was almost magic. Fast forward to the present day, we are elated to bring these new and ultra-cool tools to our customers. The applications are wide ranging; from advertising to marketing, to law enforcement and government."

The new AI features are an optional module for Torque’s DVStor Recording and Playout system, which is globally deployed for compliance recording and disaster recovery playout applications.

You might also like...

Essential Guide: OTT (or is it ABR?)

Program delivery to mobile devices and smart televisions has fueled the growth for internet delivery. But one of the challenges broadcasters and media content providers face is that the internet was never originally designed to stream large amounts of video…

Bonded Cellular Spring 2020 Update

Need a live shot from inside an unmarked moving rental sedan during a thunderstorm? No problem.

Is Gamma Still Needed? - Part 1

Gamma is a topic that pervades almost all forms of image portrayal, including film, television and computers. Gamma has become a tradition, which means that its origins are not understood, and it is not questioned. Perhaps it is time that…

Audio Levels - Part 4

There are two basic reasons to know the level of an audio signal. One of these is more technical and one of them is more subjective.

Software-Defined Automation: Are We Nearly There Yet? Part II

Playout automation has been enabling fewer people to control more channels for decades but we’re not quite at the point where human interaction can be eliminated altogether. Since most linear broadcasters will either move to a software-based deployment for t…