Steve Lampen, multimedia technology manager of Belden sheds light on what's next after HD-SDI.
There's a famous quote from Charles H. Duell, who, in 1899 was Patents Commissioner in the United States.
He said, "Everything that can be invented has been invented!"
Of course, we laugh at that now, because our entire technological world has appeared since then. And it doesn't look like our technological rush will be slowing down any time soon!
When it comes to video for instance, don't think for a minute that HD-SDI is the 'end of the line'.
Next up is 1080p/50, a format with even higher resolution. 'Regular' HD-SDI (720p or 1080i) has a bandwidth of 1.5 Gigabits-per-second (Gbps), and a clock of half that number (750 MHz).
Since we are working with data here, and data is square waves (or is supposed to be square waves), we know that a perfect square wave is a fundamental (clock) frequency with an infinite number of harmonics added in.
Since it is impossible to make cable, or anything else, that can handle infinite harmonics, we must instead choose how many harmonics to check.
The rule of thumb has always been to check video data streams to the third harmonic.
That would mean a standard HD-SDI signal (750 MHz clock) should be tested to the third harmonic (750 MHz x 3 = 2.25 GHz).
This simply means that anything intended to carry HD-SDI should be tested to 2.25 GHz. This would include cable, connectors, patch cords, patch panels and pretty much anything.
The only things that you don't have to worry about are those items that are less than a quarter of a wavelength at 2.25 GHz.
The wavelength of 2.25 GHz is slightly more than 5.25" so a quarter-wavelength is slightly more than 1.3 inches. What's that small? Maybe one connector.
This means that you could have one connector that is bad, or installed wrong, and everything would still work. But that's about it. Two connectors would be over that distance, and everything else, patch cords, patch cable, and certainly the cable in the run itself, has to be as close to 75 ohms as possible.
So what about 1080p/50? This signal is obviously double the bandwidth of 1080i. That 'i' means 'interlaced' where every other line is drawn in each field.
In contrast, the 'p' means progressive, where every line is drawn. So where regular HD-SDI is 1.5 Gbps, 1080p/60 is 3 Gbps. The clock is half that, 1.5 GHz, and the third harmonic is a whopping 4.5 GHz.
So, if you're buying anything to run 1080p, your first question should be "Has this been tested to 4.5 GHz?" I'll wager you that you will get a few angry responses.
And some of them will tell you that it is 'not possible' to test anything at 75 ohms to 4.5 GHz.
This is no longer true. At least one cable manufacturer and soon one connector manufacturer, are already testing their products to 4.5 GHz. The truth is, other manufacturers are unwilling to spend the money to buy the custom test gear required. Most current analysers go out to 3 GHz, fine for standard HD-SDI, but not enough for 1080p/50.
And what do they need to measure way out there? Simple. Is this component truly 75 ohms from low to high frequency?
If it's not stable, if the value of the impedance wanders around, it means that some of your signal will be reflected by the off-value impedance and there will be less signal at the other end of the cable.
This impedance variation measurement is called 'return loss' (RL).
With all passive pieces lined up, the maximum return loss allowed under SMPTE 292M is -15 dB. That translates to a reflection of just over 3%.
That doesn't sound like a lot, but think about how many cars have to be going the wrong way on the freeway to cause an accident?
One would be a reasonable number. It doesn't take a lot of signal going the wrong way to ruin your HD-SDI delivery.
So maybe you should aim for something like -20 dB RL. That's a 1% reflection.
You could also ask for -30 dB RL, but that is 0.1% reflection, and no cable ever made could maintain that consistently to 4.5 GHz, and not even to 3 GHz. So, -20 dB RL up to 4.5 GHz, sounds pretty reasonable.
If they can't measure that high, then -20 dB RL (or better) up to 3 GHz, might work running 1080p/60. But you see the problem. There's a whole swath of bandwidth that you simply don't know about if you stop at 3 GHz.
The most popular cable used for 1080p/60 has a return loss spec of -23 dB from 5 MHz to 1.6 GHz and -21 dB from 1.6 GHz to 4.5 GHz.
So who is asking about 1080p/50 and why would you put that in? This is especially pertinent if you have priced any 1080p stuff. It is way expensive! Of course, like any new technology, 1080p is expensive to start with but will be dropping as the technology improves. But there are three compelling reasons why you should consider 1080p now.
First is that 1080p/50 home displays sales are increasing. Plasma, LCD, L-COS, OLED, and especially projectors can come in a 1080p format. So the only thing you need is a source.
In the United States, the ATSC on what HD is, will be considering ways to add 1080p to the over-the-air standard. In the meantime, you can get 1080p from a BluRay player (now that HD-DVD is dead), or from a Sony Playstation.
In other words, not much yet! But, sooner than you think, there will be lots of products with the 1080p/50 format. By shooting in 1080p/50 right now, you can capture that 'future market' that those shooting 1080i or 720p material will have to up-convert to or abandon the sales after all.
Secondly, 1080p/50 material will have a longer shelf-life and will be accepted and played more often than 1080i/720p material. And finally, 1080p simply looks amazing and provides a nearly theatrical film experience. It always helps to be ahead of the pack.