There seems to be a blurring of the line between engineering fact and marketing hype emerging within our industry. We must push back to maintain high technical standards.
One of the interesting aspects of my role as editor is reviewing the large volume of press releases I receive each day. The Broadcast Bridge has very high editorial standards and our focus is on technology, product announcements, and applications. But recently, I’ve noticed several press releases making some interesting claims. Assertions I believe we need to question and even push back on.
One press release that sticks in my mind made the claim that a video to IP converter has zero latency. Well if that’s been achieved then congratulations to the engineer that has just made electrical signals travel faster than the speed of light.
The best definition of broadcast specific latency I found is from The Farlex Dictionary and states “latency is the time it takes for a packet to cross a network connection, from sender to receiver”. Even if you don’t agree with the specifics of this definition, I think we can all agree that latency implies some time delay. Quite often, I find latency and delay are used interchangeably.
And zero, means zero, not nearly zero! Even if it did, “nearly” needs qualification. As far as I’m concerned, “zero latency” suggests the output packet of data is available at the instant the pixel of video is presented to the input of the device. What about propagation delays in the electronics? Yes, they’re small, but they exist. And what about the buffer time it takes to build the ethernet frame and IP packet? Even if the whole frame isn’t buffered, there will still need to be at least 32 clock cycles of delay so the cyclic redundancy check can be assembled for the ethernet frame. Then there’s the ethernet preamble and the interpacket gap etc.
I can hear a whole industry of marketing people shouting at the screen and accusing me of being pedantic, anally retentive, and even too boring! But as an engineer, as with many engineers, I need to understand exactly what is going on under the hood to build the most optimal system possible. If I assumed the zero claim was valid then I would have a design with confused timing planes.
I accept, in this instance, the latency may be very low, but I cannot accept it’s zero – and to me that matters.
I think, as an industry, when we see claims such as these, not only should we throw in a healthy dose of cynicism, but we should also push back and ask the vendor to substantiate and validate such assertions. We must do everything we can to maintain the highest technical standards and drive forward core engineering principles.