Abstract: Developing a Compressed Video Test Strategy
Digital video broadcasts feature not only millions of bytes of digital data that must be error free, but also multiple video, audio, data and program guide transport streams that must be perfectly in sync. Small errors are commonplace and have the potential to derail a broadcast and cause costly blackouts. To make matters worse, these errors can surface intermittently, making them extremely difficult to identify, understand and solve. A sound test strategy is to first detect an error, prioritize it, judge its importance and then locate where in the chain the problem exists. Detection and resolution of these problems needs to be simplified to ensure time to insight is kept to an absolute minimum to avoid the end user being impacted.
The Challenge: Testing and Verifying Compressed Video Transport Streams
Digital video broadcasts feature not only millions of bytes of digital data that must be error free, but also multiple video, audio, data and program guide transport streams that must be perfectly in sync. Small errors are commonplace and have the potential to derail a broadcast and cause costly blackouts. To make matters worse, these errors can surface intermittently, making them extremely difficult to identify, understand and solve.
With the added complexity of today's digital video streams, substantial new test engineering issues become critical. These include standards-compliance, which is both assuring that equipment interoperates with that of other vendors, and the video and audio decodes correctly by the consumer's equipment. There is also an entirely new set of problems arising within transport streams carrying new compression systems such as MPEG4/H.264/Windows Media 9 and SMTE VC-1. Partial transport streams and signaling for mobile video standards such as DVB-H also present new test challenges.
Figure 1: Typical modern broadcast system See full size image
Modern Broadcast Systems
Traditionally conducted offline, analyzing and debugging modern broadcast systems and digital video transport stream files can be a time-intensive and complicated procedure. Traditional transport stream analyzers create a log file of collective errors, but users are forced to manually retrace the log to identify and evaluate each error individually. This often equates to sifting through lengthy reports with tens of thousands of logged errors to find a single, problematic fault. This process can take hours, days or months " or prove wholly unsuccessful.
The key to maintain reliable, high quality services over the different video distribution, compression and transmission systems is to focus on those factors that are most critical. These include service information (SI) and system timing in the form of Program Clock Reference (PCR) measurements, in addition to any additional services such as Internet protocol (IP) or Electronic Program Guide (EPG) data.
Fortunately, new tools are available that render the test and analysis of digital video broadcast streams, as well as the development and implementation of new compression standards a far easier and faster process, minimizing costs and maximizing revenues. With these tools, intermittent errors in a stream can be found in real time using triggers to pin down difficult problems. Recording triggers can be assigned to any of the tests, with pre-trigger buffers that can determine what led to the problem. Error logs can help with tracking frequency and occurrence.
These tools also provide a summary screen, offering an "at-a-glance" view of the most critical and important measurements on which to focus. Figure 2 shows a pie chart of the multiplex occupancy that allows users to quickly see if the stream is 'live' and decoding and determine whether the test probe and stream are operational.
Figure 2. MPEG layer summary screen on the Tektronix MTS400
See full size image
Developing a Test Strategy
A sound test strategy is to first detect an error, prioritize it, judge its importance and then locate where in the chain the problem exists.
How to do this?
In an instrument like the MTS400, users connect a transport stream feed (from ASI or Ethernet interface typically) then look at the programs tab for red error LEDs, in order to determine which service is contains an error. Critical errors are displayed, which users can correlate with error logs to further trace the problem.
How To Measure PCRs (PROGRAM CLOCK REFERENCE)
What is a PCR?
PCRs enable the MPEG decoder to synchronize to the encoded video and lock the 'system time clocks' together. A 42 bit sample of the encoder's System Time Clock (STC) indicates to the demultiplexer what the STC time should be at the decoder when each PCR is received, as described by the ISO 13818-1 Annex D3(1) MPEG-2 standard. Synchronization errors arise if the PCR value generated by the multiplexer is inaccurate or if it is received late because of network delays. Variations in these delays are described as PCR jitter.
From where do PCR errors originate?
Inaccuracies can come from faulty encoder PCR circuitry, faulty re-multiplexer PCR circuitry or a failure to seamlessly loop a Transport Stream. They can also arise from unstable RF modulation or demodulation, unstable fiber demultiplexing or from ATM network packet jitter and IP packet delay variations. The buffers in set top boxes (STB) are able to address many PCR jitter situations, but there may be problems, particularly with large PCR variations that cannot be resolved at the STB.
Presentation (PTS) and decode (DTS) timestamps are derived from the PCR in the decoder to re-order bi-directionally coded frames to the correct display order.
Which Are the Critical PCR Measurements?
The most critical PCR measurements are those that directly impact the presentation and decode timestamps and are listed below: (see Fig 3)
PCR Arrival Interval (PCR_AI): Limited by MPEG-2 to 100mS and by DVB to 40 ms
PCR Accuracy (PCR_AC): Limited to +/- 500 ns based upon a stable reference
PCR Overall Jitter (PCR_OJ ): Algorithms are defined by the TR 101 290 standard (MGF1, 2, and 3) as filter functions using: 0.01, 0.1, or 1 Hz. There are no limits defined by this standard, but PCR-OJ values can be very large in the order of micro or milliseconds. PCR_OJ is inter-related to jitter in the PCR Arrival Intervals and PCR accuracy (see Figure 3 below). These are real-time effects and the transport stream packets must be accurately time-stamped to a stable reference clock at the interface to achieve accurate measurements. The Tektronix MTS400, for example, allows users to define their own PCR-OJ limits to accommodate acceptable known performance on a given system. Real-time PCR measurements made on a real Transport Stream are shown in Figure 3 below.
Figure 3. PCR Jitter Relationship
How to make a PCR measurement
Users can make PCR measurements either in real time directly from live transport streams, or offline from a previously recorded transport stream file. If the test equipment time-stamps incoming packets, both arrival interval jitter and overall jitter can be measured and assessed.
Specifically with the MTS400, users click on a program and then look for the clock symbol icon which represents a PID containing PCRs and click on it. The instrument then presents users with a 'PCR graphs tab' that enables them to open and view all the available graphs and measurements. Clicking on the "limits icon", will display red "out of limits" areas above and below the graph, enabling users to see, at a glance, if a PCR error exists. Errors are also listed in tests (2.3, 2.4) and error logs, documenting any errors that are found.
Another icon allows users to zoom in time and amplitude to examine the nature of the PCR error (drift or sine wave distortion, for example).
Figure 4: PCR Graphs
See full size image
How to detect & fix MPEG4/AVC/VC-1 problems
Next generation compression systems, such as H.264, Windows Media 9 and VC-1 require other specialist Elementary Stream test tools. These tools are able to analyze the various macroblock types, motion vectors and improved motion estimation supported by the new compression standards.
Figure 5. Elementary Stream Analyzer display
See full size image
Better Approaches to Analysis and Monitoring
As complexity increases with these new compression standards and new broadcast systems such as delivery of Video over IP networks are implemented, new approaches are needed to minimize time to insight. Test systems that provide displays with processed information, not simply raw data are essential to quickly isolate critical problems. Additionally, analyzers that provide simple error color codes to help pinpoint issues to service/programs, and are equipped with user interfaces that require minimal navigation to focus down to the root of a problem are helpful in solving problems before they impact the end user customer.
Figure 6. Processed information instead of raw data simplifies and speeds troubleshooting
How to detect Electronic Program Guide (EPG) Problems
Real time and offline EPG displays can help locate errors in the EIT and ETT tables which are used by set top boxes to generate the EPG.
Figure 7. EIT table, part of EPG
See full size image
How to Detect IP problems
The deployment of Video over IP means the transport stream will be transmitted using 10/100 or Gigabit Ethernet interfaces rather than ASI, so monitoring and capture from these interfaces are required to analyze and debug these new transmission networks.
Figure 8. Real-Time MPEG Over IP Analysis
See full size image
How to capture intermittent problems
Finally, help is needed on the most difficult problems - particularly intermittent ones, where a fault can cause a decoder to crash or cause errors with processing equipment in the broadcast chain. Methods have now been developed to identify and capture intermittent faults in real time or from stored files, allowing the user to debug and locate the root cause.
For example, MTS400 software uses a sophisticated trigger scenario to capture the error, halting the analysis exactly at the point of error. This enables faster and easier debugging of intermittent problems that would otherwise be missed.
Figure 9. Tektronix MTS400 with CapturVu halts the analysis exactly at the point of error
Digital TV transmission systems are very complex and new compression standards are bringing new challenges for equipment manufacturers and network operators. Only by thorough testing with the most effective tools in early development, commissioning and roll-out stages, can users ensure their newly launched services will work cleanly. The benefits of new technologies are great in reducing bit-rate and maintaining picture fidelity, but they bring new challenges which need new approaches to testing.
Test systems must be designed such that lower-skilled engineers and operators can carry out the most critical tests easily and in minimal time. When a problem of interoperability occurs, it's essential to be able to pinpoint the cause quickly and determine where the problem resides.
The overall aim is to simplify these difficult test and monitoring tasks into just those critical ones that truly impact the end customer. The equipment must enable speedy diagnosis of the problem at the appropriate layer and location. Only then can remedial action be taken to minimize down-time and maintain the highest quality of service, at minimal cost overhead.
Chris Purdy is an Applications Engineer at Tektronix, Inc.