Switching off for good?

As content distribution shifts online via IP networks, is the end nigh for matrix switchers? Steve Montgomery investigates the likely outcomes.

The demise of the video matrix switch in professional AV installations has been predicted. It is expected to be replaced in the future by online content distribution over  IP networks. If this is true, the questions facing the industry are when will it occur and  how should manufacturers and integrators address the industry in the near term in order to future-proof products and installations?

Nick Mawer, marketing manager at Kramer, believes that: “Eventually we will route all our signals on the network; closed AV systems based on proprietary technologies will vanish. As the AV industry converges to IT we’ll see more manufacturers  picking  up  with  this  shift  and introducing a multitude of AV over IP products and systems. As with every new technology it will have its own adoption rate, like the move from analogue to digital.”

Stijn  Ooms, technology director for Crestron EMEA, agrees: “Whilst it is true that the network is playing an increasingly central role in the AV industry, the reality is that matrix switchers will continue to be a valuable asset for some time to come. The main reason for this is that currently-available codecs are not sufficiently efficient to stream  effectively without using a great deal of bandwidth. Insufficient bandwidth leads to a myriad of problems for other applications on the network such as email, access to the internet and IP telephony.”

The response from Crestron is a hybrid product. Ooms continues: “The DigitalMedia  switcher combines a traditional matrix switcher with streaming capability that can  receive streaming signals as well as send them when the application requires it.”

Matrix switchers are the backbone of AV systems. Their function is to route the outputs of computers and other image sources to display devices located within the same room, or across buildings, campuses, and even around the world.

The type of matrix switcher required will depend on the application  for  which  it  is  intended. To meet an extremely wide range of potential applications, matrix switches are available in several sizes and forms. From small HDMI video switches used to manage multiple sources onto multiple  displays, to massive extender-switches capable of delivering many video streams over long distances.  Many companies offer  modular systems that allow mixing and matching of input and output types, additional components such as back up control cards and redundant power supplies and extended features, including USB interfaces for data transmission to allow the addition of keyboards and mice to the end terminals.

Individual selection of  the most appropriate device depends upon its intended use and  the features required for the particular installation.

“There is a long list of questions the integrator should address,” says Dave Schorah, WyreStorm product application specialist.  “Start with the desired transmission distance and consider how many source and display devices are required, what type  of content is to be transmitted, what does the end-user  want to do with that transmission and what control features are necessary.  That leads on to detailed system
requirements:  decision on video resolution, formats, control interface, whether  Ethernet needs to be passed, and how remote terminals are to be powered. Only then can the actual device be selected, to meet the expected number of inputs and outputs expected; with allowance, of course, for future expansion.”

Since its introduction  in  2011,  HDBaseT  has had  a  major  effect  on  the  video  matrix  switch market.  It has become the de-facto  standard for long distance HD video content distribution when HDMI fails at around 15 metres. HDBaseT devices can transmit video up  to 4K  UHD resolution (although not at 60Hz, 4:4:4 colour space) and the newly released HDBaseT 2.0 will add  additional  features  to  add  multipoint  and multistream capability as well as USB support for KVM applications.

At  the  same  time, a new version of HDMI is being launched.  HDMI 2.0 caters for 4K transmission, multiple video streams and users, and adds extra audio channels and  resolution.

However, as Mawer points out: “The HDMI 2.0 spec exists, but display and source manufacturers are not rushing  to meet it yet.”  David Shamir, Atlona’s  senior product  manager for IP Audio/Video, points out, along with an interesting twist on HDMI marketing: “Manufacturers have been asked to specify product using the specifics of the features rather than the HDMI version number.”

The  net  effect is  that matrix switchers are keeping pace with the rapid evolution in display technology,  even if their full capabilities are not yet being utilised; a characteristic that will maintain  their  leading  position  and  longevity in an industry that is constantly being asked to provide the latest-and-greatest products.

Whilst 4K is receiving a lot of attention and many  UHD  monitors and TV sets are  already available,  there  are  relatively  few  devices  that are capable of producing  native  4K  images.

This  will  change.  “Given how fast prices of screens are falling and the inevitability of  the technology  as  an  evolution  of  HD,  4K screens are  certainly  a  necessity  for  new  installations,” believes James Meredith, WyreStorm NetworkHD and Enado  product manager.  “Laptops,  tablets and even some phones are capable of generating 4K content, so anyone specifying an HD screen now is really not looking very far down the road unless they are happy to scale content down to HD.  However, in terms of  whether  a  client  or end  user  is  looking  for  4K@30Hz or 60Hz, as long as its 4K  they don’t really know, or care, about the difference as long if they can perceive improvement over their old HD screen.

“The reality is that even with no noticeable improvement, most users will probably presume they are getting 4K, a problem further compounded by screens offering an HD  to  4K  upscaling  function  that  might  offer slightly higher picture quality that still  falls short of true 4K.”

Ooms points to another consideration for mixed resolution systems: “Many devices in use now are referred to as ‘tweeners’: they fall between 1080p resolution and 4K. These native resolutions are far greater than HD distribution systems can handle.

Many organisations still use a 1080p switching environment so a presentation produced at the native  high  resolution  of  a  laptop,  might  lose some  of  the  imagery  and  text  impact  when  it is  scaled  down  to  1080.  Also,  many  industries, such as medical, government and military require the highest resolution possible. Those who have invested in new 4K displays expect 4K resolution, rather than downscaled 1080p.”

Matrix switchers based on HDBaseT distribute uncompressed video and therefore do not suffer image artefacts brought about by compression, nor do they suffer transmission delay or latency from the packetisation  process  employed  in video-over-IP network transmission. The degree of acceptance of compression and delay depends largely  on  the  application.  Live  content  and videoconferencing  should  not  be  delayed  by ore than two frames, although in large event staging  the  delay  may  be  nullifi ed  by  audio delay  across  a  stadium.  As  long  as  screens  are synchronised, signage and playback of recorded content is not usually important.

Every  application  can  accept  an  amount  of compression:  presentations and  conference rooms  have  one  level,  live  events  another. Command and control and medical applications cannot  tolerate  any  compression  artefacts  at all.  There are  many  different  and  alternative compression  schemes  although  JPEG2000  and H.264 are the two most commonly used in AV network video transmission devices.
JPEG2000 is a wavelet-based technique that tends to make a higher quality image then macro block based compression algorithms. It is typically used  in  mission  critical  applications  such  as military,  government,  legal  or  healthcare  where the  visually  lossless  nature  of  the  compression ensures the image at one end is the same as at the other. H.264 is efficient and highly selective but is not desirable in applications that require high quality, as it can discard important elements of the image.

According  to  Paul  Harris,  ceo  of  Aurora:  “A compression ratio of 8:1 can be used, however quality  depends  not  just  on  the  level  but  the technique  and  choice.  Many  of  the  higher compression ratio devices will convert 4:4:4 to 4:2:2 making a PC image softer-looking and the colour  palette  will  no  longer  be  correct.  Users understand  the  term  ‘lossless compression’, however  the  issue  the  industry  has  is  there is  no  official  testing  to  confirm  an  industry-standard  minimum  requirement.  Most  visually lossless systems are not. JPEG2000 products are typically 9:1 and you can see artefacts with some content.”

The  bottom  line  is  that  we  are  all  heavy consumers of visual lossy compression. Lossless video  is  essential  for  some  highly  demanding applications but high quality lossy compression is acceptable for most others.

“There  are  many  variables  that  influence whether  an  existing  data  network  can  support AV  streaming,”  explains  Shamir.  “Network load  pattern;  how  busy  the  network  is,  and  at what  time  of  the  day  it  gets  the  busiest.  How much  bandwidth  can  be  allocated  to  the  AV component;  achievable  image  quality  and latency when operating within the allocated bit rate; available bit rate between network switches; traffic  forecast  and  whether  the  network  can sustain that traffic at a sufficiently high quality of service (QoS).”

Both  technologies  have  their  advantages and position in the  market.  Harman’s  recent acquisition  of  network  AV  provider  SVSi indicates  its commitment  to  that  technology and  complements  the  AMX  Enova  matrix switch product range.  “In  the  long  term the ability  to  move  large  amounts of data anywhere is most  likely  to  be  based  on Ethernet, as network bandwidth increase,” says  Brian  Davies,  European  technical  director at  AMX.  “We  can  already  distribute  4K/30Hz over  10Gb  fibre  networks  and  with  new encoder  firmware  that  will  soon  be  possible over 1Gb too. As the industry moves upwards to 40Gb  and  then  100Gb,  it  will  make  video distribution even more efficient, reduce latency and  enhance  picture  quality,  opening  up opportunities even further.”