Video Optimization OEM Applications

The Need for Wireless Video

As the use of wireless mesh networks has become more prevalent, the applications running over those networks have become more sophisticated. Wireless video has proven to be a popular application, especially in tactical networks where the collection and sharing of real-time visual information for surveillance and situational awareness is extremely valuable. Real time video is a demanding application with high bandwidth and stringent latency requirements. Carrying video in a wireless mesh adds additional challenges because the various network connections are not static but are constantly changing. Handling multicast video adds another layer of challenges.

A Brief Primer on IP Video

Most video compression techniques are lossy and degradation of picture quality due to compression is expected. Codecs (compressors-decompressors) convert a raw video stream into a compressed stream and vice-versa. Codecs compress video streams by removing both the spatial and temporal redundancies in the original video stream. They operate on the premise that much of the data present before compression is not necessary for achieving good perceptual quality. For example, the MPEG-2 standard video compression can compress video data by 15 to 30 times, while still producing a picture quality that is generally considered high-quality for standard-definition video. Video compression is a tradeoff between data size, video quality, and the cost of hardware required to decompress the video in a reasonable time.

There are three types of frames used in IP video compression: I‑frames, P‑frames, and B‑frames. A “frame” may span multiple data packets. An I‑frame is an "Intra-coded picture", basically a fully specified picture, similar to a static image file. P‑frames (predicted) and B‑frames (bi-predicted) hold only part of the image information, so they are smaller than an I‑frame and thus improve video compression rates.

The Challenge of Packet Loss

I/P/B frames can span multiple packets. What this mean in practice is that small packet losses can result in significant I/P/B frame losses. If video frames are lost in transmission, visible (and sometimes distracting) artifacts can appear. The type of frame lost can have a major impact on the effect of the video. For example, since the images derived from P-frames rely on an I-frame on which to base the predicted image, the loss of an I-frame can mean that the following frames cannot be reconstructed properly. Because of this, IP video is sensitive to, and cannot tolerate, high levels of packet loss. If there is no acknowledgement of video packets and retransmission of lost packets an impaired video stream cannot be remediated. The effects of packet loss also depend heavily on the picture content of the video. A video of a static scene is much more tolerant of lost frames than a dynamic image containing a lot of movement. In a dynamic image, which is more typical of tactical use, losing I-frames can cause significant difficulty delivering viewable video which could significantly impact safety in the field.

The Challenge of Multicast

Most video transmission is unidirectional by nature, that is there is one source of video and multiple receivers. Transmitting separate one-to-one video streams (or unicast stream) for each receiver implies replicating the video stream for each individual receiver. This wastes network bandwidth, leaving less bandwidth available for other video streams or applications. Bandwidth utilization can be optimized by sending a single stream addressed to multiple receivers (multicast stream). However, any loss in data for a multicast stream affects the video received by all viewers, whereas for a unicast stream, data loss only impacts the receiver for which the data is lost. Hence data loss in a multicast video stream is more critical.

A Fortress Solution for IP Video

Fortress FastPath Mesh™ handles multicast video as true multicast. The protocol determines the optimal path from the source to the subscribed viewers and then sends the data along that routing tree without unnecessary duplication of network traffic.

FastPath Mesh™ also provides several protocol mechanisms and settings to automatically, or manually, control the transmission mechanism over the air. This allows for optimization of both the video transport reliability and the overall bandwidth use. These mechanisms allow mesh to provide optimal video transport that scales with throughput and multicast subscriber number.

Because FastPath Mesh™ is media independent, it can be readily integrated into OEM solutions for optimized transport of IP video streams in a variety of applications. Fortress radios are Size, Weight and Power (SWaP) optimized, and Fortress FastPath Mesh™ utilizes the network bandwidth efficiently, particularly for applications such as multicast video without unnecessary duplication of resources. Integrating the Fortress radios and FastPath Mesh™ enables the Primes to offer the best-in-class wireless networking capability at a reduced cost. Fortress Technologies has the expertise to build custom-designed modules or work closely with Prime’s engineering teams to build a highly optimized solution that performs at par or better than an external add-on component. An OEM integrated solution using Fortress radios and FastPath Mesh™ enables OEM Partners and Primes to present an attractive offering to their customers.