In this edition of the Matrox Podcast, Sound & Video Contractor Magazine contributor Phil Hippensteel talks with Ron Berty, Business Development Manager at Matrox and Dave Chiappini, who is the Vice President of Research and Development for AV-over-IP Solutions. They discuss the transmission standards used by AV over IP systems including H.264, JPEG 2000 and MPEG-TS. They also explore the pros and cons of open vs. proprietary standards and where Matrox products fit into the larger picture.
Links of Interest
Download Podcast Here:
Hello. Welcome to today’s podcast sponsored by Matrox, a partner to Sound and Video Contractor. My name is Phil Hippensteel and I’ll be your moderator. I’m a writer and a researcher about all things IP and an adjunct instructor at Harrisburg campus at Penn State University. With me today are Ron Berty, Business Development Manager at Matrox and Dave Chiappini, who is the Vice President of Research and Development for AV-over-IP Solutions. We’re going to be discussing the AV-over-IP paradigm and some issues that seem to keep popping up when this evolution is discussed. So we’ll dive right into the questions. Let’s start with an overview of AV-over-IP. It’s becoming something of an industry buzzword, but the paradigm has been developing for quite some time, hasn’t it been?
Ron: Yeah, it definitely has. This is Ron, here. And you know, if we look back, Google acquired YouTube about 10 years ago now. And that was for about $1.5 or $1.6 billion or some vast sum at that time. But from 2005 where YouTube had almost no viewers to one year later in 2006 where YouTube had reached almost $20 million audited visitors per month to their site – and that’s all about AV and video-on-demand over IT – several IT companies like Microsoft and Google realized that the era of mainstream video-over-IP had arrived. Everything that has followed that point is natural. The emergence of like global video streaming at low bit rate and H.264, the emergence of hardware-accelerated H.264 devices at low cost, low power like laptops, PC’s, phones, tablets, media players, and even the emergence of nichier products that use the same base technology, all of that has come from that point. So it’s really been going on for like 10 years and the real reason that people kind of moved to AV-over-IP is not because they weren’t, you know, back then they were going well, you know, we were dissatisfied with the products that already existed that did sort of AV and baseband or traditional mode. But it was to tap into this new era of development and era of AV-over-IP that was growing at a very rapid rate. And people like corporations even started moving towards YouTube. And it wasn’t just broadcast or small video guys that were using YouTube as a communication channel. Corporations even started using it and making it part of their marketing strategy. And so from there it continued to grow and build out to its own industry and its own – basically a paradigm shift in terms of how AV-over-IP or video was transported. [Timestamp: 2:46]
I’ve noticed in the research that I do when I gather up materials to get ready to write an article or to write a white paper or something like that my resources are changing. I used to use books and then I started to switch to white papers. And now I use a lot of things that are videos that are from YouTube and from manufacturers as resources. So it’s clear that our behavior is being modified by the change that you’re talking about, Ron.
Ron: Yeah, I think no doubt about that.
What do you think the keys to AV-over-IP, the keys to this paradigm are? What are the must-haves to make this work in enterprises of all different sizes?
Ron: Well, I think there’s a lot of different keys that can be used and different requirements that are there. But at the same time you can sort of generalize to a degree and different users place different emphasis on key points. So if we look at a couple of scenarios. For example when leveraging AV-over-IP to address new applications of AV in a significantly expanded AV reach. So you’re really trying to get beyond the walls or beyond the limitations of traditional AV and that can mean things like global distribution or distribution to wireless and handheld devices. Distribution to massive concurrent viewing video-on-demand, etc. And the keys technologically to AV-over-IP tend to hinge on factors like standards based interoperability and quality audio visual transmission at ultra or at least reasonably low bit rates. A strong ingredient to ever-improving AV and IT cooperation is safe, controllable use of networking. This includes the option to phase in slowly and use alternative networks for AV only and to eventually bridge applications to leverage the main network for converged data, communications and even AV-over-IP. Maybe not completely, but at least at a minimum partially. On the other hand when leveraging AV-over-IP performance applications that strain key elements like resolution, refresh rate, image quality and video processing, the main impetus for IP is standards-based switching. The idea is just use the network because the network is a standard and now you’re really using a standard-based way to move the video from location to location and you’re not dictated to start from one place and go to another and just stay from source to destination and that’s all you can do. You can move your destinations and you can move your sources. So the idea is delivering that low bit rate on the network and significantly increasing the density, the density that you can have there for both sources and destinations, and you can do that with highly efficiently-compressed video or AV. You can do it with lightly-compressed video or AV. And here there’s less standards, but codecs like JPEG 2000 do a good job of balancing and working on standard IP with maintaining quality and relatively low latency. And then you really have something sort of relatively new, which is fully uncompressed. And of course you might be on a 10-gig network to do that, but you can actually do fully uncompressed HD at 60 Hz and get that from one place to the other. And 10-gig networks are pretty common today and it’s moving to a much greater standard or a bigger standard of 25 and even 40-gig networks in the near future. And those prices are coming down so it makes it more commercially viable to possibly do that. So, you know, in all cases of AV-over-IP, the nontechnical requirements are that either the AV team has a computer-based networking skillset and they control the movement of data across the network or the AV team and the skilled – their skillset cooperates with the IT people who tend to own the decision-making authority on the network and the security of that in the facilities. But in either case these are – I guess not in either case, but in all three of those examples, that’s really where we see the users or typically the impetus for putting the emphasis on different key points. So depending if it’s reach, performance type things like strain resolution, refresh rates, image quality, video processing, size of video and that sort of thing or really just managing who’s controlling the network one way or the other, whether it’s the IT department or the AV guys, that’s kind of the way we see that emphasis and different key points that people are managing in a general sense. [Timestamp: 7:19]
So if I understand you correctly, you’re pretty much saying that the IT side of the world has been very much standardized and as far as in the AV world it’s AV-over-IP that’s bringing some standardization, what I would call relatively consolidated standardization, that allows for high interoperability and flexibility at the same time.
Ron: That’s exactly right. You said it, especially with that last word. So the standardization is there and then the interoperability is key because that’s what the standardization brings. And that’s absolutely right, Phil. [Timestamp: 7:53]
So I think that kind of leads to what I had written down as my next question, which was to talk about codecs. What we’re really talking about here is that we’re getting this capability from H.264.
Dave: Well, yes. H.264 is one of many codecs that enables this. And in general H.264 is the broadest supported codec standard out there today. And when Matrox was investigating that market space we wanted to provide a solution that would have broad ranging support, which means it would have to be open-standard. It would have broad spectrum market option, the ability to deliver on density bit rate, reach, quality, latency. And after reviewing the options H.264 is the one in our opinion that met all of those criteria. Interoperability hardware and software-wise, H.264 is an open standard. It meets all the high quality of audio and video – well, it doesn’t do the audio, but it supports the transport protocols that do. You know, all at low bit rate H.264 is supported by all the popular browsers, laptops, media players, smart phones, TV’s. They all support in hardware natively H.264. So H.264 today can offer every resolution from SIF up to 4K/60. Even YouTube recently actually upgraded their support now for H.264 on RTMP to use that as a 4K transport into YouTube as a CDN or video-on-demand repository. H.264 is such a mature standard. It supports everything from video distribution quality at YUV4:2:0, broadcast quality 4:2:2, desktop mode quality at 4:4:4 so you get all the fine details of your applications reproduced perfectly. This can all be done at very aggressive bit rates without compromising your quality. You know as an example, H.264 typically runs 800-1 compression ratio compared to JPEG 2000 at 25-1. So it’s a massive difference when you’re talking about the total bit rate so there’s no need for special network equipment or 10-gig equipment. You can get low latency. Matrox provides better than 50 millisecond glass-to-glass latency on solutions. Clear and accepted licensing structure for H.264 as opposed to some of the newcomers. It makes it much easier for people to adopt and develop on. Of course, you know, overall there will be new standards that come out in the future. They’re working on a new HEVC, EV1. Google is talking about their VP-9, VP-10 series. All viable solutions later on when maturity is there. And Matrox will keep up with the latest and greatest in the market when the world is ready for those, but if you ask me H.264 is at a sweet spot today and is running with the most broad-spectrum support. That’s where all of our bets have been placed because this is the one that delivers on all of the things I described. [Timestamp: 11:11]
Well, it seems to me you’re right about H.264. The other day I noticed when I – when my computer did an automatic update Windows decided there was a couple of files that were part of the operating system it didn’t want to update because I had some video and audio files on my computer which would no longer have digital rights management support and that I might want to remove those from my computer before I did the update. And I sat there and thought for a while and I realized what they were talking about were two things: QuickTime and .wmv and .wma. And I thought that’s what happens when you’re using proprietary standards. So I don’t think I’ll ever get a message like that about H.264.
Dave: Yeah, I would agree. I mean there’s a place for the proprietary standards. When there isn’t an open standard yet available because it’s either something too new or a new type of problem that’s emerged then proprietary solutions tend to respond quicker. But the minute, and usually very quickly, the open standard comes along all of the adoption will be on that and those open standards are what really protects your long-term investment, the ability to have the protection of multiple suppliers on the same type of technology. It drives costs. It drives quality. It drives everything. Open standards are very important. [Timestamp: 12:33]
So there’s also a message in the fact, some people say well how can we depend on that fact that H.264 will be around for a while? When you stop and think about the fact that MPEG-2 probably, and it’s still widely used, has been around for close to 20 years, there’s no reason to believe that H.264 won’t be around for at least another decade.
Dave: Absolutely. All these standards start out that way and you picked the perfect example. There’s been a lot of other codecs in the meantime, but you know, MPEG-2 is probably the last what I like to call super codec. You know, massive, broad-spectrum support for the day. Obviously today H.264, I mean there’s a refrigerator at the local furniture store that plays back H.264 video clips. I mean everything supports H.264 today natively. So the fact is that that will be around for another 10 years easy if not longer. And sure, there will be – like I always said, there will always be something new, but before another super codec comes along MPEG-2 and H.264 are the only two of them. It will take time. Maturity has to build out. Infrastructure has to build out. So it will come, but it’s not the time right now. [Timestamp: 13:43]
There’s a term that I always like to get clarified when I’ve been at tradeshows and even when I’ve talked to people I’m writing papers for or doing presentations on video-over-IP related topics. I bring up the issue of MPEG transport and people have it confused with MPEG compression. Matrix and coders create an MPEG transport stream. While it shouldn’t be confused with MPEG compression, can you explain why that format is so important to the industry, the MPEG transport stream format?
Dave: Yeah. So first off you’re absolutely right. It’s very common to confuse that the – you know, sometimes called the MPEG transport or MPEG-2 transport stream and that’s often confused with the video that it can handle. So MPEG-2 transport stream can carry H.264. Yeah. So the name of this or what a transport stream is, is a way of formatting data so that it can be put over a network and it’s a standard. So it describes how all different options can be put into something so that it becomes interoperable. Why is MPEG-TS so valuable is because it is such a strong standard that was pushed by the broadcast industry originally and has gotten so much adoption out there, you know, because it first started with set-top boxes and then moved its way into EV that everything has MPEG-TS support in it. And so if you want a strong legacy codec that has good feature set and lots of interoperability, MPEG-TS is a pretty safe bet in many cases to make sure that you’re going to find multi-vendor support on that platform. Not to mention as a standard it’s very feature-rich. It can carry multiple videos, handle synchronization, optional reliable transport over noisy networks. And like I said its big selling point to me is that it’s supported in so many different units that it’s a pretty safe bet that you’ll be able to find multiple third party vendors. Of course it is an older transport protocol and there are newer technologies today. But I think MPEG-TS is likely to be around for a very long time. So you know it’s a safe bet and you should make sure that if you’re trying to find something with a conservative, long-life approach it’s a good way to go. [Timestamp: 16:00]
I always like to point out, too, that I did a lot of work with the companies that built test equipment to monitor and troubleshoot IP networks and you’ll find that MPEG transport streams are almost always decoded by the major products and by even free tools like Wireshark. But when you get into other transport stream formats, most often the protocol analyzers, the sniffers and products like that that do network monitoring that try to help you out, they don’t even recognize that it’s video very often.
Dave: Yep. Very good comment.
I want to thank you, Dave and Ron – Dave Chiappini from Matrox and Ron Berty from Matrox – for spending this time with me. And I hope the discussion we’ve had is useful to our listener. Thanks very much for joining us.