Case Study:  Anue Network Simulator provides network impairments to streaming audio/video traffic

Tests on FUB/ISCOM test plant with the Anue Network Simulator

Summary

The Italisn Ministry of Communications needed to understand the impact of signal corruption, packet loss and signal impairments.

Project Implementation

The Anue platform supplied by Federal Trade S.p.a. was introduced on Fondazione Ugo Bordoni / Istituto Superiore delle Tecnologie e dell’Informazione (FUB/ISCOM) test plant (see the diagram below) and through its emulation and network parameters alteration, the degradation on the audio/video streaming in transit has been evaluated.

In particular, the following parameters have been considered:

ADD DROP probability

Above parameters have been altered according to the specs.

ITU-T Y1541  QoS  network performance

ITU_T G1010 QoS  end user

ITU-T WT-126 Quality of experience - requirements and mechanisms

During this test session, made by Eng. Sergio Pompei and Cristiano Zema, and coordinated by Eng. Luca Rea from FUB and Mr. Elio Binnella from ISCOM, above alterations have been set based on a deterministic way, i.e., by applying them in a specific segment of the audio/video plot and for a defined time slot.

“The Anue Simulator has proved itself in supporting our audio and video streaming test methodology based on introducing elements of signal corruption.”   

-Luca Rea (FUB) & Elio Binnella (ISCOM)

HD/SD-TV Video transmission tests on IP

              Test Streaming:

• SDTV: MPEG2 720x576 16/9, bitrate 9,29 Mbit/s of average, 29.97 fps progressive
• HDTV: MPEG2 1280x720 16/9, bitrate 24 Mbit/s of average, 60 fps progressive

MPEG2 1920x1080 16/9, bitrate 24 Mbit/s of average, 29.97 fps interlaced

• AUDIO: In both cases MP2 192 Kbit/s and 48 kHz was used

DSL Forum WT-126 Standard on IPTV:

• 5 pks lost every 4 seconds, SDTV
• 5 pks lost every 4 seconds, 1280x720p
• 5 pks lost every 4 seconds, 1920x1080i

The above standard, still in a debate phase, provides for standard SDTV a maximum of 5 IP packets lost consecutively every 30 minutes and as for HDTV a maximum of 5 IP packets lost consecutively every 4 hours.

Lost packets, periodic distribution:

• 80 pks lost every 3500 pks (80 pks about every 4 seconds), SDTV.
• 80 pks lost every 6400 pks (80 pks about every 4 seconds), 720p.
• 80 pks lost every 8857 pks (80 pks about every 4 seconds), 1080i.

This kind of test has been useful to evaluate the high level of uncertainty of the impact on video generated by codec, even though packet loss is constant

 

Simulation of a 50 ms gap in transmission

• 44 pks lost every 3500 pks, SDTV.
• 80 pks lost every 6400 pks, 720p.
• 111 pks lost every 8857 pks, 1080i.

This case was designed to reproduce the damage caused by an interruption of a connection due to link failure (the changeable number of packets depends on the number of the total packets in each stream and the relevant bitrate).

Conclusions

The packet loss with periodic distribution has outlined a better video response in high resolution than in standard resolution.  This is due to the fact that the percentage of lost packets compared to the total is lower in HDTV than in SDTV (the damage is however evident and unacceptable if frequent).  The stream in SDTV contains 24500 packets compared to more than 62000 of HDTV at 1080i.  Also the weight in bytes is obviously different.  It goes from about 30 MB of SDTV to the about 80 MB of the HDTV at 1080i for 28 seconds of film.

As already mentioned in several occasions, the tests have been made according to state-of-art terms not yet defined by a reference standard.