音视频监测 -开心体育

as prismon supports many different interfaces, protocols and codecs there is no easy answer to this question. for example, a single prismon prime server can support up to 24 3g-sdis or more then 80 st2110 with the recent server generation. but new and more performant server hardware become available regularly. detailed figures can be found in the datasheet and rohde & schwarz solution architects will help you to dimension your systems – especially for complex and mixed setups.

modern cots server platforms using state-of-the-art processors typically provide enough computing power to deliver the same high quality, low latency and 24/7 stable multiviewer images as dedicated special purpose hardware – even without additional gpu acceleration. this of cause requires software and a operating system highly optimized for real-time video processing. if implemented the right way, a software-based multiviewier is in no way inferior to hardware solutions. but they also deliver the high flexibility that is required in our ever-changing industry: easy extensions towards new codecs, formats or protocols. the possibility to use and simply replace, upgrade or repurpose cots hardware if required and the possibility to migrate a solution to the cloud are benefits that only software-based products can provide.

of course, achieving low i/o latencies in a software-based multiviewer running on cots hardware is more difficult than on special purpose hardware as cpu cycles and other hardware resources that are shared with operating system tasks or other applications. however, if done right, latencies, especially for ip i/o can be as low as for hardware solutions – typically between 2-3 frames.

of course, a traditional hardware-based or hardware accelerated multiviewer system cannot be deployed in the cloud. for such a solution it would be required to bring all the streams generated in the cloud back to the central hub to perform local multiviewing. if optimized accordingly, modern software-based multiviewer system can be deployed on different cloud environments directly. all signals can be processed directly in the cloud and combined to a multiviewer image – which is then fed back to the central hub through one of of the available streaming formats (srt, ndi,…).

having the possibility to distribute dozens of hd feeds on a single 100g connection is of course perfect for multiviewing applications. so by connecting a 100g enabled multiviewer solution to the st2110 100g network will allow the multiviewer to receive all the required feeds and transmit the generated multiviewer output on just a single cable.

this depends on the technical limitiations of a multiviewer product and the personal preferences of the operators. resolutions below qvga for a single tile will become very blurry and are typically not recommended. for a hd multiviewer screen this equals max. 36 channels. on an uhd multiviewer screen this equals max. 144 channels.

many modern multiviewer products offer open apis for remote control. if broadcast controllers provide a driver for a specific product api or if the api is based on an open standard (e.g. amwa nmos) many functions such as layout changes, switching of input signals or applying different monitoring presets can be automated or triggered by 3rd party broadcast controllers.

often the terms multiviewing and monitoring are used synonymously. by definition, multiviewing describes the pure visualization of multiple video signals alongside with audio bars and umd labels on a single screen. monitoring capabilities allow additional automated probing & checking of the different signals. those checks can be executed on different levels (content, protocol, signal) and detected errors will get reported through different mechanisms or highlighted on the multiviewer screens.

this can be achieved through different ways. one very efficient way is to create a multiviewer video at the remote location and to stream it back to a central hub using one of the different streaming formats such as srt or hls/dash. this will enable remote multiviewing in a decent quality and latency over (intercontinental) ip connections at around 3-4 mbit/s.

if the capacity of a single server is exceeded, prismon allows you to cluster multiple servers. through the available sdm feature, inputs can be pre-processed on one server and shared with many other servers for multiviewer generation. this enables efficient resource usage avoiding duplicates, infinite platform scaling and allows design of prismon system and with basically any i/o count

the sdm technology is based on a proprietary protocol, highly optimized for latency and throughput. proxy videos are shared using no additional compression to avoid quality degradations at the lowest possible latency. through the usage of unicast connections, no additional multicast-enabled infrastructure is required when using sdm.

each prismon system provides its own web-based user interface and a set of apis that offers complete control over the system and delivers sophisticated monitoring over signal and system health. furthermore, prismon provides the multiviewer control center extension – a single user interface to control multiple prismon instances with focus on layout generation, layout management and layout activation. it offers feature-rich graphical editors and a workflow optimized rooms & scenarious concept to activate a large set of mutiviewer layouts throughout an entire facility with just a single click.

similar to the question on capacity there is no single answer to this questions as it highly depends on the input and output formats. for example in pure st2110 setups it can be as low as 2 frames.