CSI
The
tipping point for IP live investment has been reached but will
interoperability hamper adoption?
IP
networks were never intended for video. The brittle, time-sensitive
nature of video does not play well with the proven but lossy nature
of IP – even less so on shared and unmanaged networks like the
internet. The varying delay and constant packet loss of the internet
play havoc with every video stream traversing it unprotected.
Despite
this, the industry is entering a time where IP will be the video
transport technology of choice. “With the introduction of
multiscreen and personalised content there is a need for more agile
and flexible workflows,” affirms media transport solutions
provider, Net Insight. “This is driving the adoption of file-based
technologies, as well as flexible transport solutions.”
As
bitrates increase and equipment prices drop, IP-based communication
technologies are pushing more and more dedicated communication
systems into retirement.
“There
is not a new camera or production solution on the market today that
does not support IP,” informs Net Insight. “New and modernised
studios are often completely based on IP, and file-based workflows
naturally use IP to move files from point A to B.”
Potential
production of resolutions up to 8K (or beyond!) require
bandwidths previously unheard of and remote production further pushes
the need for speed. The ability to scale networks without a glass
ceiling and without
spending huge capex on replacement kit is arguably the most
compelling business argument for the move to IP.
To
get there however, broadcasters do have to make a decision to rip and
replace existing copper wire cabling. The cost and benefits are about
to reach a tipping point but just because COTS is tried and trusted
in almost every other industry, the broadcast community has
reservations about getting it right. This is especially pertinent of
live production and there are differences of opinion in the industry
about whether IP live is ready for primetime.
“The
great bright future is out there but as an industry we can't tell
them how it works with 100 percent confidence,” says Tim
Felstead, head of product marketing, Quantel Snell.
“In a live environment, when you have adverts to get to air,
people's jobs and reputations are on the line. You
have to prove to broadcast engineers that when a director says 'Go to
Camera 4 now' that it will happen.”
Mark
Hilton, VP
Infrastructure Products at Grass Valley,
agrees that there's an element of hyperbole about IP but that “it
is coming on quicker than we all thought. and we're seeing proof of
concepts being commissioned.”
With
its own brand of IP-enabled products from camera, production
switchers, servers and gateways to convert SDI to IP just launching,
Hilton believes small scale IP live production is possible just
around the corner.
Imagine
Communications is even more bullish. “Live IP has been possible for
years. It
is not about IP, it is about whether or not broadcasters should look
at operational changes,” says Brick
Eksten, vp, product strategy.
IC
points to the reference site it is building in New York with
Disney/ABC which includes full live production switching over COTS.
Other
first movers include Pac-12 Networks – the broadcast arm of the
conference of 12 west coast universities – which uses T-VIPS and
Nevion links to transmit talkback, telemetry and telemetric data to
and from sports venues as far as 2500 km away, apparently with less
than a second delay. ESPN's Digital Center 2 opened last year built
around a J2K–based Evertz EXE-X2 IP routing core with pockets of
baseband workflows. It is capable of routing more than 6,000 HD 1080p
streams and as much as 9 TeraBits per second.
“From
a master control operational perspective - hitting buttons on a panel
– should I, as a broadcaster or engineer, expect any difference
from SDI to switching video over IP or running video processing over
software?” poses Eksten. “The answer is no. Imagine is all about
transparency with IP. We are saying that the interaction feels the
same as it did when audio/video was run over SDI.”
Quantel's
Felstead is not so sure. “It
is much more difficult to see what is going on in IP. The control
systems don't exist [but being developed]. Where SDI routers were
very reliable with straightforward verification of what was
happening, IP systems are more opaque. This creates a lack of
confidence.”
Quantel
Snell's research indicates that 27
of industry stakeholders believe IP routers will replace SDI within a
decade. “While
[Quantel Snell] support 2022 we don't believe it is the right way to
go long term,” says Felstead. “The industry isn't able to
transpose IP into a live environment today.”
At
face value the Imagine and Quantel Snell stance appear at odds but in
fact they are voicing very similar concerns. SMPTE 2022-6 is the
first incarnation of realtime video over IP and deemed solid enough
to get the industry moving. It is the standard on which most
manufacturer's starter IP kit is based.
However,
2022 emulates the way base-band is used and does not have the
capability to send multiple data streams on the same wire. If you
want to freely mix and match different cameras or audio tracks, a
prime advantage that IP offers, then a new standard is required. This
could be SMPTE 2022-8/9/10 which the standard's body is working on.
The Video Services Forum, which has focussed on J2K, has another and
there will likely be demonstrations of both next year.
The
move to 4K complicates matters further. In a live environment do we
need fully pristine uncompressed 4K? Or will a mezzanine format be
good enough? Some form of compression will have to be good enough in
the early stages of 4K over IP since current 10 GbE connections do
not have the capacity to carry it uncompressed.
Codec
contenders include IntoPix' Tico alliance backed by Grass Valley;
J2K; VC2 (backed by Quantel) and Sony's Low Latency Video Codec.
London-based V-Nova claims its codec can deliver 4K picture quality
at half current rates (just 7-8Mbps) with hints that this could be
applied to production.
“One
of the great big wins of IP infrastructure is leveraging the cost
savings by using commodified IT kit,” says Felstead. “To do that
we need to be able to ensure interoperability, but there is no
standard common between enough manufacturers to ensure this.”
Quantel
flags that economics of multiple proprietary codecs would negate much
of IP's supposed cost-savings. “When you compare the efficiency of
SDI routers and IT routers in handling different encoding standards
and you add up the core devices and peripheral devices you come to
the conclusion that the driver is going to be the edge devices,”
says Felstead. Vendors
like Quantel Snell are hedging their bets, incorporating a variety of
IP standards into switching and routing gear.
“If we have lots of pieces from different manufacturers and every
link has a encode and decode stage with a royalty fee it will run
counter to the very principal of COTS driving infrastructure costs
down.”
Hilton's
concern is that some technologies require different types of
hardware; “The Sony LLVC needs to be designed into the hardware,
J2K has quite a long latency right now and is fairly computationally
intensive and Tico, while optimised for this light compression, is
not a good enough standard yet.”
Sony
played a major role in developing the original SDI as a universal
interface is trying to do the same with its own IP connection. It has
the support of a number of manufacturers but is unlikely to receive
the blessing of rivals like Panasonic.
“For
a real successful implementation of IP it's very important that one
standard is accepted and adopted to allow interoperability between
systems just as SDI currently provides,” says Peter
IJkhout, CTO, VidiGo.
“At present, several organizations or companies are developing
competing protocols and we have to wait and see how this will
progress.
“Adoption
will depend on acceptance of compression in the production chain as
well as unavoidable longer latency compared to traditional SDI,” he
stresses. “Without
a well-designed protocol that can be shared seamlessly between
vendors and equipment there will be reluctance to invest in IP as SDI
replacement.”
Ericsson
Broadcast & Media Services, CTO Steve Plunkett, is more
conciliatory. “The
proprietary implementations prove the technical viability of IP as a
transport medium,” he says. “They are providing real world
experience that in turn feeds into a general body of knowledge of
professional media over IP and they offer short term solutions to
organisations who need to implement now. However, they are not viable
in the long term. The industry needs scale to reduce costs and that
will not be achieved with closed vendor specific solutions.”
According
to Imagine's Eksten
the major broadcast kit players are talking 2022 interop mainly about
their own equipment. By contrast, “We are targeting expansion of
our universe of 2022 interop with other companies, which is a huge
step forward required by the industry.”
In
any case, the issue may soon be redundant. With technologies of 40GbE
and 100GbE already out of the labs (and in place at Disney/ABC) the
velocity of advance in IT should iron out temporary capacity
restrictions though not necessarily cost.
“SDI
routers are based on a price per port while IP is typically on amount
of bandwidth,” explains Felstead. “If you put video over IP
unconstrained in bandwidth you may have a problem cost-wise. If you
use uncompressed SD 200Mbps you've got less of a problem but if you
use HD, 3G or 4K bandwidth consumption quickly becomes an issue.”
More
significantly for some is the human factor. IP requites not just a
change in technology but a change in the way people do things. “The
required network systems for 4K over IP are complex and expensive,”
says IJkhout.
“Traditional engineers at broadcasters are very video oriented and
it will take time, and being honest often different people, to make
the transition into IP engineering.”
“If
you've been working in SDI for 30 years and all of a sudden it's
based on servers this requires different skills sets,” says Adam
Cox, head of broadcast equipment, Futuresource Consulting. “You can
retrain them, but they will still think like engineers. The lack of
skillsets are a big barrier to IP live.”
Quantel's
pitch is don't hold off on IP plant infrastructure but do so with the
confidence that the investment is going to be used for the lifecycle
of the equipment and won't block you out of future standards. “We've
engineered IP interfaces into live SDI product like Kahuna and Sirius
so we can offer an immediate hybrid approach,” says Felstead.
IP
has already swept through contribution and distribution and will
inevitably become the defacto signal route for live. The
opportunities are simply too compelling.
“Competition
from internet and cable is intense on broadcasters which have
incredible opportunities to interject advertising into their
programming,” argues Eksten. “That is not the case in broadcast
but it can be and needs to be.
“IP
mean not having to capex a bunch of equipment every time you want to
launch a channel but by using virtualised networks and compute
resource to spin-up a channel in hours rather than month. And then it
turn off again as needs be.
“When
you start to see 2022 capability in pure software, the ability to
scale and change network routers to adapt to new business parameters
is just phenomenal.”
