The traditional Sync Pulse Generator should stay as a standard even when SMPTE introduces its new version, argues Joe Ellis, product manager, Trilogy
have to distribute it; equipment manufacturers know that if they detect and react to it then their equipment can be plugged into a production system without problems. File-based infrastructures are
Joe Ellis: ‘Precision Time Protocol too inaccurate for HD’
One of the side effects of moving towards file-based video is that the basis of timing has changed. In a realtime video signal – analogue or digital – all devices are timed together with reference to a master Sync
Pulse Generator (SPG), so any source can be switched without picture jumps or glitches. The sync pulse – black and burst, or tri-level for HD – has become a universal standard. Systems designers know they
different: they need to use IT network timing protocols. SMPTE has been looking at this for five years now, and engineers from Trilogy have been monitoring progress and contributing to the debate. The solution for the timing issue looks to be precision time protocol (PTP), codified as IEEE 1588 in 2008. PTP is capable of sub-microsecond precision in the right circumstances, which is good for IT networks. Unfortunately it is not
accurate enough for HD, which expects accuracy in the nanosecond range. So SMPTE is continuing to develop the standard to achieve this level of precision. Once it is standardised it will involve a great deal of care in the
distribution of the timing reference. IEEE 1588 allows for multiple master clocks, with the network choosing the most stable, but in the broadcast environment where short lock- up times are vital, this decision-making will add a perceptible time overhead. The great advantage of PTP that it is carried over Ethernet, but there will be constraints on the number of hops from the grandmaster clock to the device, and it will probably only work over a local area network. It is possible that PTP-enabled routers will be manufactured in future with dedicated hardware to ensure minimum jitter in the time reference, but major manufacturers like Cisco have yet to announce them. As sometimes happens, the deliberations of the standardisation process mean that some manufacturers are pressing ahead with their own versions of ‘broadcast PTP’, hoping it will be subsequently
adopted as a de facto standard. This is always a dangerous situation.
What does this mean for the
future of timing? I see two issues. First, with no clear and agreed standard, the SPGs we have relied upon for so many years still have their place, and no system today can be designed without them. Second, when any new
standard is agreed, it will be unreasonable to expect manufacturers to update every device instantly. Completely greenfield installations are also so rare as to be very nearly unheard- of: legacy equipment almost always has to be integrated. So I do not see the need for the traditional SPG to go away even when SMPTE introduces its new standard. Certainly the SPG will need an ethernet port and the ability to translate PTP to black and burst, but our industry-specific, really realtime method of distributing timing information will still be around, still providing high precision synchronisation and timing for broadcast technology. 10A29
NO STRESS, WE MANAGE! Reliable IP broadcasting
IPGuard Seamless IP switch
VISIT US ON BOOTH
#2.A49
The IPGuard is the world’s first seamless IP switch with bypass mechanisms that provides safe and 100% delivery of any IP streams. It provides automatic 1+1 redundancy of encoders, multiplexers, DVB-T2 gateways, ASI to IP adapters,… Additionally, the IPGuard can align the same MPEG-2 TS or T2-MI stream carried over redundant IP-based networks to enable a seamless switch-over operation.