Pushing packets
The previous “Transition to Digital” tutorial briefly addressed Quality of Service (QoS) and its relationship with Quality of Experience (QoE) in a digital stream. Both these terms cover a wide variety of digital signal testing, evaluation and viewer experience (QoE) measurements. This tutorial will focus on packet QoS.
QoS measures the ability of a packet-switched digital network, typically IP, to move discrete packets of data from source to destination by controlling the flow of the packets through routers, switches and the transport layer with minimal errors. Not only is QoS a measurement, it is also an integral feature found in many modern network devices.
ATSC transmits MPEG data packets somewhat similar to IP packets. Other than technical differences in how the packets are defined and constructed, the most important difference is that over-the-air (OTA) broadcasters have complete control of the performance of the packet delivery system as defined by the ATSC and the FCC. This is not so much the case over an IP network.
When DTV packet data is sent over an IP network, packets are sent at a higher rate than the source baseband signal. This accounts for much of the original delay because signals are converted from a dedicated SDI stream to MPEG packets, and allows for much more information to be transported across a network structure than a single AV stream. There are some pros and cons to the AV packet data is moving at a higher rate of speed than its baseband speed. The good news is that packets are less susceptible to impulse and noise spikes; the bad news is that the loss of a single DTV packet can have a significant negative impact on picture and/or audio quality, which can last several frames.
An IP packet can contain up to 65Kb of data, which could define more than 2000 RGB samples in a video image. A typical DTV packet consists of 188 bits. Time code, if necessary, will add another 4 bits, resulting in a 192-bit packet. The ATSC 8-VSB DTV packet specification is 208 bits, regardless of whether the video content is SD or HD. Those 16 extra bits carry Reed-Solomon Forward Error Correction (RS FEC) data. One of the cornerstones to the stability and reliability of an ATSC stream is that it requires a constant bit rate. Depending on the complexity and movement of a scene, a multiplexer might insert additional null packets, which may not contain any relevant data, to maintain the constant bit rate. The null packet’s function is to instruct the receiver to ignore its contents or lack of contents.
A DTV packet contains a header that starts with an 8-bit sync byte, followed by three flags that reveal a transport error indicator (TEI), a payload unit start indicator and transport priority flag. These flags are followed by a 13-bit packet identifier (PID). The PID enables the demultiplexer to identify all packets that contain the same PID. Typically, time-division multiplexing is employed to determine the rate the same PID appears in the transport stream. One bit in the PID is reserved to signal a null packet.
The next four bits that follow the PID are a continuity counter, which tells the receiver if a payload exists. The rest of the packet is the payload, which, in the case of DTV, is a tiny slice of video and audio data.
Get the TV Tech Newsletter
The professional video industry's #1 source for news, trends and product and tech information. Sign up below.
The transport priority flag network identifies packets that have priority over others to routers and other devices on the network. When a network becomes congested, the decision to drop or discard data packets is based on priorities. A Type of Service (ToS) field in an IPv4 data stream is an 8-bit field, starting with a three-bit explicit congestion notification (precedence) message that takes priority identification to the next seven levels. The three precedence bits define a range from zero to seven, with seven being the highest priority. Higher numbers are frequently used in live streams. The next five bits are the Differentiated Services Code Point (DSCP), which can set routers to normal or minimum delay, normal or maximum throughput and/or normal or high reliability.
The ToS field is set at the location where the IP packet is originally created. Sometimes, the ToS field can be ignored as the stream is transported because not all routers follow the latest QoS protocol. Other times, some routers may modify the ToS field to conform to the priority it has been preassigned. From that point forward, it will pass through other routers and switches with that new priority. For example, a video server’s MAC address could be programmed into the rules of a router to always have the highest priority to guarantee continuous delivery.
A router remembers the ToS data for each route in its routing table. Data or equipment using a protocol that doesn’t support ToS will usually cause a router or switch to default to a ToS of zero. Within a broadcast facility, engineers should verify that all equipment in a dedicated AV network is set to the highest priority to ensure the highest QoS across the network.
If the transport layer of a live AV stream on the network drops or discards a packet, the result will be a degraded image. If the packet happens to contain an I-frame, the next few P-frames and B-frames will degrade further, or be lost altogether. To quickly review, an I-frame is an “intra-coded,” fully specified, stand-alone video frame similar to a JPEG still image. P-frames are “predicted” picture frames, which describe only the changes in the image from the previous I- or P-frame. B-frames are “bipredictive” picture frames, which use the differences between the current frame, the preceding and following frames to determine its content.
DTV transmission, reception and viewing problems, whether IP, ATSC, MPEG, QAM or streams, are usually caused by dropped packets, metadata errors, buffer overflow/underflow, jitter or overall system health. Errors at the user’s end often appear as macro blocking, lip-sync inconsistencies, erratic PSIP or other metadata information, or frozen or lost signals.
The lesson to be learned is that to maximize performance and minimize AV network problems, careful study of the manuals of all the gear that touches your network is a valuable time investment. It only takes one bit to ruin your day.