In Broadcast - April 2015 - 34

34

www.inbroadcast.com | Vol: 5 - Issue 4 | April 2015

InForm

Long-Duration Testing In
The IP/Internet Domain
By Adam Schadle, Vice President of Video Clarity

A

s broadcasting and content
distribution move ever farther
into the Internet domain, satellite,
broadcast, cable, and IPTV network
technologists, system architects, and
engineers must continually test new
codec technologies to ensure that
distribution setups are primed to
deliver the best possible video quality
at multiple bit rates, frame rates, and
resolutions. That quality is ultimately
determined by the viewers and thus is
not easily quantiﬁable.
Testing Duration
Reference-based
video
quality
testing - whereby the quality of
the output is measured against the
source content and then compared to
accepted scales of human perception
- is still the most accurate way to
determine how viewers will see your

video and the level of quality they will
perceive.
Testing is important because many
things can go wrong when transmitting
video over IP, such as:
*		 The	video	encoder	can	badly		
compress the video;
*		 The	video	encoder	can	package		
the data incorrectly, causing
a poor transmission;
*		 The	transmission	media	can	add,		
delay, corrupt or miss packets;
*		 The	video	decoder	can	fail	to		
read the transmitted stream;
*		 The	video	decoder	can	badly		
decompress the video.
Any one of these things can
negatively affect what the viewer sees.
Often this type of testing is done
on short clips of video - perhaps a
couple of minutes' worth - to test and
compare performance over a series of

Video Clarity's RTM real-time audio and
video monitoring solution

frames. This type of testing is shortduration testing.
But what would happen if the video
processing units did not produce an
error for several hours or days? What
if a particular set of incoming data
sent at just the wrong time was what
it took to trigger a problem? Those
scenarios are very difﬁcult to replicate,
but they will be the ﬁrst things your
customers notice should they occur
after you transmit the content. That's
why it's also important to monitor
video over a long period of time for
video quality, audio quality, lip sync,
and loudness. Long-duration testing
not only gives you an understanding
of perceived video quality from
the viewer's perspective, but also
helps you pinpoint signiﬁcant errors
and problems with uptime (such as
breakups, black, artefacts or skipped
frames).
Delivery Results
When performing reference-based
testing in the IP domain, the source
material comes in via IP or SDI. In
order to accommodate the different
proﬁles required for IP delivery, you
ﬁrst have to reduce the resolution/
frame rate of the source material
in order to create an acceptable
reference - and do so using the target
transcoding device so as to minimise
the artefacts and/or degradations that
result.
You must create a new version of the
reference for each stream that needs
testing and for each proﬁle or device
under test. Only then does it make
sense to compare it to the encoded/
transcoded material that the end user
will see.
Once you create the acceptable
references, what's the best way to go
about long-duration testing of your
IP-bound content? The answer is Video
Clarity's RTM real-time audio and video
monitoring solution (RTM).

RTM continuously monitors the video
quality, detects the error event, and
saves the video and audio event and
surrounding content before and after
the event, so that you can diagnose the
problem and devise a ﬁx.
When comparing the reference to
the downstream output, RTM aligns
the audio and video in the two streams
and looks for areas with temporal
disturbance, such as scene changes,
fades to black, fast motion changes,
talking to silence, talking to music,
or music to silence. If RTM detects
any dropped frames, it dynamically
realigns them.
RTM analyses the reference and
processed audio and video streams
separately, calculates an offset for
each, and creates text and graphical
test data for video quality, audio
quality, lip sync, and loudness. The
offset depicted is the delay caused by
processing. In theory the audio and
video delay should be equal. If they are
not, then the difference could create a
lip-sync problem.
RTM can run via SNMP, socket
commands, or from its own intuitive
interface, shown below. RTM logs all of
the data and generates alerts based on
user-deﬁned proﬁles when:
*	 Video	quality	drops;
*	 Audio	quality	drops;
*	 VANC	data	is	not	complete;
*	 A/V	delay	is	incorrect
(indicating a lip-sync problem).
RTM is an easy-to-use quality
measurement solution based on one
of the most highly accurate videoquality measurements available. It
performs several tests that are critical
to determining quality of delivery in
an IP or baseband environment. With
RTM, you can set up a test and leave
knowing you will see alerts and logs
if an error occurs. And the ultimate
payoff is a keen understanding of the
performance of any potential delivery
method you are considering.

Figure 1: Real-Time Quality
Measurement and Monitoring of New HEVC Codecs and Services
http://www.inbroadcast.com http://videoclarity.com/rtm/ http://www.truelens.co.uk http://videoclarity.com/rtm/ http://nab15.mapyourshow.com/6_0/exhibitor/exhibitor-details.cfm?ExhID=1315 http://www.truelens.co.uk
In Broadcast - April 2015

Table of Contents for the Digital Edition of In Broadcast - April 2015
