The Language of AV/IT is IP: Q&A with Matrox Graphics
When we talk about AV/IT convergence, we often focus on the IT industry’s impact on the AV industry. But the improvement in audio and video quality has undeniably influenced the IT industry as well. In many ways like Thomas Edison and Nikola Tesla, the two industries have been working independently (and together) on new solutions and standards for transmitting audio and video signals over the network. Whose idea is simpler? More efficient? More disruptive?
When it comes to 4K, the main design goal is to reproduce the video signal over long distances faithfully and without latency. As we’ve shared in our recent white paper, 4K and Beyond: Technical Challenges and Opportunities, AV integrators struggle with bandwidth constraints and delivering high-bit content at high frame rates. A lot of faith is being put in encoders and decoders to push massive video files through a relatively small pipe error free. Could Internet protocols provide a simpler and more efficient video transmission in the pro-AV world?
Matrox Graphics has been in business of delivering pixels for quite a while now. When the graphical user interface (GUI) replaced the character prompt in computing, Matrox was there with the market’s first 64-bit graphics processing unit (GPU) and mainstream Windows graphics accelerator. At the U.S. Display Consortium in 2003, they presented their first 4K graphics board that could run 4K at full frame rates on the IBM T221 3840×2400 monitor from a single GPU. They’re a pioneer with an interesting history, so we met with their Vice President of Research and Development David Chiappini and Sales Director Samuel Recine to talk about 4K over IP from the computer science side.
NM: How are the Internet protocols influencing video transmission in the pro-AV world?
Recine: In a total way. Unlike HD, where full-quality encoding came much later than baseband extension and switching products, 4K is entirely different and the encoders that are capable of full quality low bit rate encoding are available at the beginning of the move from HD to 4K AV infrastructures. In some facilities, the move to standards-based IP encoders will be the first 4K extension and switching infrastructure. Since the ability to use standard deployed Category cable exists already, it’s hard to compete with 4K encoders that can leverage cables and infrastructures already in place.
NM: What current encoding/decoding techniques do you use?
Chiappini: There are many, but if I had to pick one, then by far H.264 has made the largest impact. Why is that the primary one is because it serves 99 percent of markets with highest efficiency. Looking forward in that space are improvements with HEVC [high-efficiency video coding or H.265], but I think that as far as maturing and growth goes, infrastructure isn’t there. I know Netflix does it, but you are going to need custom endpoints. But, with H.264, I can stream to almost anything. My brother has a car that decodes H.264.
NM: How far do you think we are from streaming uncompressed 4K?
Chiappini: I don’t think that streaming 4K uncompressed really makes sense. I’d say that if you’re talking about streaming 4K inside a facility, that may happen, but keep in mind that with uncompressed 4K, you break the 10-gigabit barrier and all of a sudden you’re in a much more expensive realm of distribution. So, I don’t think that’s really going to happen. Not in a big way without compression. And then, when you start to bring the factor of compression, it’s well, what type of compression, what’s your use case.
NM: Where do you see IP-based capabilities going in the next five years?
Chiappini: Today, chips that can encode and stream 4K are available. So, the world is not going to revolve around entrepreneurial, ad-hoc solutions like it used to. It took twenty years for MPEG 2 to get into full stride. It took seven years for H.264 to get into full stride. It’s going to take probably three to four years for H.265 to get into full stride. If I have to invest in the infrastructure for my company using standard baseband switching technology like matrix switchers and so on, that’s expensive and they’re typically tied to a resolution. We already have commodity displays from Dell and LG that are bigger than 4K. So another big advantage of IP is the scalability. The minute I can jump on IT, and the minute I can compress the signal to a reasonable amount, I can distribute it to mobile devices, I can record it, I can send it to multiple end points — the amount of flexibility gets opened up and the options available to me go through the roof. The whole story of IT/AV convergence and IT/AV communications is here because IT has been gaining so much in terms of enablement and availability and the ability to have enough horsepower to use it the way it was meant to be used. And the language of this communication is IP.
NM: As more consumers begin using 4K, won’t that overwhelm the current bandwidths? Won’t such a development require a big change in the infrastructure?
Recine: We can stream high-quality 4K at 15 megabits per second. Codecs continue to improve their efficiency and continue to deliver as much quality at even lower bit rates. So, if we can do 4Kp60 at 15 megabits per second today, you can be sure that a year from now that bit rate will decrease further, and years from now, it’ll be a fraction of that.
Chiappini: Distribution methods are also increasing in intelligence to support that. If I’m streaming on my phone, I really don’t need to have 4K to do that. When that 4K content goes out to a large display wall for visualization, then you want to see it in full res. When it comes to monitoring applications and any other mission-critical applications, that’s where more advanced transport mediums come in, like HLS [HTTP Live Streaming] and MPEG-DASH [Dynamic Adaptive Streaming over HTTP]. What they do is stream multiple bit rates for the same resolution, autonegotiate the whole protocol, allow you to optimize the link based on the amount of errors that occur, the amount of decode power and the general quality preferences of the user. So, even if you don’t have fiber-to-the-home, these techniques will stream at lower resolutions. It’s a proven fact that people prefer smooth playback than choppy quality playback.
NM: What can the AV industry do about improving 4K transmission?
Recine: Become experts in benchmarking the best 4K low bit-rate encoders. Some applications still depend on the extension and switching of video. And a subset of those applications require low to no latency. Latency is probably the dividing line on a paradigm shift as 4K mushrooms at breakneck commoditization levels — driven largely by display vendors. The paradigm shift is towards encoders. Every frontier has been breached. 4K encoding is not only in addition to 4K AV infrastructure for extension and switching. 4K encoding can now often be instead of 4K AV infrastructure for extension and switching. Encoding 4K can achieve an indistinguishable-from-source quality level at low bit rate over standard IP. This enables cost and capabilities that dwarf legacy approaches and provides many-to-many and any-to-any capabilities on open standards that proprietary vendor solutions cannot match. Still, for applications where IT-based solutions are not understood or welcomed, as well as for applications where sub-50 millisecond latency is a hard requirement, the standard fare of products that support uncompressed extension and switching will continue to yield new products aimed at 4K infrastructures in the legacy AV model. Our only advice in this case is to try to avoid proprietary products and, as much as possible, stick to stuff on open standards to avoid unwanted costly infrastructure updates in the near future.
What say you, AV pros? What are your thoughts on proliferation of 4K over IP? Weigh in with your comments.