Coping with an increased digital data load
As demand increases on digital networks, it’s important to have the right infrastructure in place. David Meyer explains.
UltraHD (4K), or UHD for short, is barely here and there’s already chatter about 8K. Where will it end? Well… it won’t. Knowledge is power, and knowing the current and future limitations of what you’re installing for your clients is key to staying ahead.
In any installation, the hardest part to swap out at a later date is the cabling infrastructure. It’s also the bottleneck for not only current operation, but particularly for future development. It all comes down to the size of the pipe and the digital data load that you’re trying to push through it. The choice of devices and connectivity will lock a system into a maximum capability – freeze it in time, so to speak – or give it room to grow.
Whatever worked in past may not necessarily work in the future. Significant changes are afoot.
It is clear that we’ll continue to evolve to higher resolutions. This is a necessary step as screen sizes continue to increase, which has been eroding effective resolution over time. That is, where a 42” TV was regarded as normal several years ago, now 55” is normal, and 65”+ is regarded as large. Resolution hasn’t changed for years, so the pixels have been getting bigger. However viewing distances for the most part haven’t grown, so from a visual acuity point of view, effective resolution has been decreasing.
Enter UHD. An 84” 2160p TV has exactly the same pixel pitch as a 42” 1080p TV. With this technology we simply get back the effective resolution we used to enjoy. But all other things being equal, four times the pixel resolution means four times the bandwidth needed to deliver the signal, and therein lies the challenge.
QUALITY OVER QUANTITY
The key to selling UHD TV is in making sure it actually looks better and that doesn’t automatically come with higher resolution. For example, compare a direct rip of a 1080p movie from Blu-ray, resulting in a 30GB file size and approx
30Mbps raw data rate. The same movie can be compressed right down to a 1GB file, but it no longer looks anywhere
near as good on screen even though it is still 1080p resolution. There’s more to picture quality than number of pixels.
Similarly, Netflix have started streaming ‘4K’, but at less than 15Mbps (using H.265) it may not look any better than 1080p. In fact an equivalent bit rate 1080p stream would likely look better in visual terms.
PERFORMANCE VS THE DATA PIPE
Standards are being developed by bodies like the Society of Motion Picture and Television Experts (SMPTE) and the Consumer Electronics Association (CEA). There are others that then define the pipe to carry the ever-increasing data load, including HDMI Licensing and HDMI Forum, VESA (with DisplayPort) and HDBaseT Alliance, among others. There’s a collective move towards higher quality, uncompressed video resolution, a bigger colour gamut, higher frame rates and greater bandwidth to facilitate it all.
HDMI 1.4a defines UHD up to 30fps with either 8-bit 4:4:4 or up to 12-bit 4:2:2. This is at an aggregate data rate of 8.91Gbps, and bandwidth of 1.485GHz. HDMI 2.0 added in this same bandwidth an increase to 60fps, but at the cost of colour; limiting to 8-bit 4:2:0. Native High Speed HDMI will support all of these, whether passive or active, as will HDBaseT over regular CatX UTP/STP.
HDMI 2.0 also includes a new higher mode to enable UHD at 60fps with uncompressed 8-bit colour, or up to 12-bit 4:2:2. At this point the pipe is far more challenged as the bandwidth increases to a whopping 2.97GHz.
Even HDBaseT with its PAM-16 runs out of puff at that point. Cat 6 and even Cat 7 will die as a digital AV transmission line. Native HDMI cables can potentially still cover it, relying on the new 2.0 EQ profile in upcoming displays to boost shorter cables, and likely additional active augmentation for longer lengths.
Beyond this both NHK (Japan) and SMPTE have been talking about 7,680 x 4,320 resolutions (‘8K’), but state at that point 60fps is an absolute minimum, with 12-bit colour preferred. With current HDMI methodology this would theoretically compute to 71.28Gbps. They actually recommend 120fps to mitigate shudder (which increases in effect with resolution), so data rate would double to 142.56Gbps. All in real time, unbuffered. And then if we want to remove aliasing from the image (wagon wheel effect), we’d need to go up to 300fps. Whoa, let’s not go there!
Joe Kane Productions advises that 4320p requires a 24’ wide screen to capitalise on visual acuity, with anything smaller not delivering full benefit. This format is therefore really relevant to digital signage and public space video walls. In the residential AV space we’re likely to cap out at 3840×2160/60p 12-bit colour. HDMI 2.0 already has this covered for YCbCr (4:2:2), but RGB graphics applications miss out, being limited to 8-bit. To go to 12-bit 4:4:4 RGB requires an aggregate data rate of approx 26Gbps, higher than even DisplayPort 1.2 can deliver.
Back down to modern day earth, this makes the current data load of 4.45Gbps at 1080p/60 look downright easy. Yet it still provides day-to-day interoperability and connectivity challenges for the installer.
CHALLENGE TO THE AV INSTALLER
The custom AV installer’s job is to make it all work, get the job finished, get paid and be confident that the job will stay working. Any deviation from this erodes the real labour rate and profitability. There is also an expectation from clients
that they acquire a degree of future proofing, and that doesn’t mean that it should be good for at least six months. The difference between a system installed by the end user and one using the expertise of a custom installer is in who takes responsibility. With an installer, something issues during or after an installation is on their watch, something that influences the choice of products in the first instance to mitigate problems.
So what to do for ‘future proofing’? The answer to that is to provide the ability to swap out connectivity as and when required. Install conduits, route cables in a way to enable re-routing, etc. You may say it’s often not possible, but it’s the only way. Even optical fibre is a tough call, as although it’s inevitable we’ll go that way, we don’t know with absolute certainty which type/spec etc will be needed for as yet undefined future applications. HDBaseT have already begun to lay the foundations for this evolution, so watch this space.
We need UHD and with it a much larger digital data load that needs to be delivered in real time. The alternative is that we halt the growth of screen sizes, which is not going to happen. Cat 6 is currently a popular choice for digital AV extension applications, but has shown to be limited to the realm of HDMI 1.4, using PAM-16 to be able to support up to 1.5GHz bandwidth on cable that’s only specified to a fraction of that.
When it comes to cabling to support installations beyond 10Gbps, assume Cat 6 won’t do it. For HDMI 2.0 (and not just 8-bit 4:2:0, by the way), the best thing you can do is stick to high quality scalable native HDMI, and provide the ability to swap out cables at a later date if at all possible. The term ‘High Speed’ will never look the same again.