Time distortion

If packets are delayed by the network, some packets arrive in bursts with interpacket delays shorter than when they were transmitted, while others are delayed such that they arrive with greater delay between packets than when they were transmitted from the source (see figure below). This time difference between when a packet actually arrives and the expected arrival time is defined as packet jitter or time distortion.^[3]

Time distortion in Ethernet packet switched networks

A receiver displaying the video at its nominal rate must accommodate the varying input stream arrival times by buffering the data arriving early and assuring that there is enough already stored data to face the possible delays in the received data (because of this the buffer is filled before displaying).

Similarly, the network infrastructure (switches, routers,…) uses buffers at each node to avoid packet loss. These buffers must be sized appropriately to handle network congestion.

Packet delays can be caused by multiple facts, among which there are the way traffic is routed through the infrastructure and possible differences between link speeds in the infrastructure.

Moreover, some methods for delivering Quality of Service (QOS) using packet metering algorithms may intentionally hold back packets to meet the quality specifications in the transmission.^[4][5]

The effects of all these facts on the amount of packets received by a specific point in the network can be seen in the next graphics:

Jitter effects on number of packets received

Packet loss

Packets may be lost due to buffer overflows or environmental electrical noise that creates corrupted packets. Even small packet loss rates result in a poor video display.

Delay Factor (DF)

The Delay Factor is a temporal value given in milliseconds that indicates how much time is required to drain the virtual buffer at the concrete network node and at a specific time. In other words, it is a time value indicating how many milliseconds’ worth of data the buffers must be able to contain in order to eliminate time distortions (jitter).^[3]

It is computed as packets arrive at the node and is displayed/recorded at regular intervals (typically one second).

It is calculated as follows:

1. At every packet arrival, the difference between the bytes received and the bytes drained is calculated. This determines the MDI virtual buffer depth:

2. Over a time interval, the difference between the minimum and maximum values of Δ is taken and then divided by the media rate:

Maximum acceptable DF:^[5] 9–50 ms

Media Loss Rate (MLR)

The Media Loss Rate is the number of media packets lost over a certain time interval (typically one second).^[3]

It is computed by subtracting the number of media packets received during an interval from the number of media packets expected during that interval and scaling the value to the chosen time period (typically one second):

Maximum acceptable channel zapping MLR:^[5] 0

Maximum acceptable average MLR:

• SDTV: 0.004
• VOD: 0.004
• HDTV: 0.0005

It must be said that the maximum acceptable MLR depends on the implementation. For channel zapping, a channel is generally viewed for a brief period, so one would be bothered if any packet loss occurred. For this case the maximum acceptable MLR is 0, as stated before, because any greater a value would mean a loss of one or more packets in a small viewing timeframe (after the zap time).