File based, SDI and AES signals have well-defined quantization levels delivering predictable signals. Bars and Tone are an historic relic from the analog VTR days and the 2T pulse lost any relevance when we moved from analog contribution lines. So, do we still need monitoring?
There is an argument to suggest monitoring a signal throughout the broadcast chain is no longer required as component drift rarely affects the frequency response and level of a digital video or audio signal. However, the signals have to be right in the first place and must be in spec. The problems caused to downstream equipment of having problems such as out of gamut errors or clipped audio will have major implications for the viewing experience, especially when we consider the massive amount of compression we now use.
With modern compute infrastructures we can easily apply so many processes to a video and audio signal it’s almost frightening. But just because we can boost the video level by 9dB at the touch of a button, or remove LF from an audio signal, doesn’t mean we should. Regardless of the medium, noise is still one of our greatest enemies. The highest quality video and audio must be right at the point of acquisition and recording and be maintained throughout the whole broadcast chain.
More fundamentally, the move to IP is changing what we monitor and how. SDI and AES tie the transport stream to the media signal to deliver predictable timing with latency in the order of a few microseconds. But IP, especially with ST2110 and NDI, have abstracted the media from the underlying transport stream. This is one of the greatest challenges of IP and simultaneously its greatest strength. Through this abstraction we have flexibility, scalability, and resilience.
Consequently, IP is providing broadcast engineers with new challenges, mainly due to the influence of the asynchronous network. We need to understand where packet loss is occurring, how, and why, as even the highest quality network is going to suffer some packet loss due to the nature of asynchronous distribution. Modern ethernet and IP networks are awash with buffers to try and mitigate packet loss caused by congestion. Although this may be effective, the net result is to not only increase latency, but also make it variable. Broadcast engineers are now playing a balancing act between guaranteeing packet delivery and keeping latency predictable and low.
Furthermore, as broadcast chains become more dynamic, that is cloud type servers are created and destroyed to meet the demands of the business, understanding when and how to increase and decrease the resource pool is becoming critical. This makes further demands on our need to understand the intricacies of the workflows leading to the adoption of highly advanced logging systems.
It’s certainly my view that not only do we still need monitoring, but we must expand the systems within a broadcast facility that we monitor. We’re no longer restricted to the media signal and must now understand the whole transport distribution system to keep packet delivery reliable while making sure latency stays low and predictable. Also, to make the most of scalability, we must understand the broadcast facilities resource allocation and even better, predict when changes are needed before they are needed. Thus, leading to a truly dynamic and proactive broadcast infrastructure.
We need monitoring and logging more now than ever to build truly dynamic, scalable, flexible, and resilient IP infrastructures.