Raw transcript

June 17, 2015.

1400 CET.

ITU.

(This text is provided as a realtime service and is not to be copied from any live event. Please contact Caption First if you would like to receive a properly formatted transcript of this event.)

>> Ladies and gentlemen, thanks for being back to session 3. Despite this very short lunch that we could grant you, I trust that more people are coming in, in a few minutes, but we have to start, otherwise we will run over too long.

So, we are approaching session 3, and the session Chair of session 3 is David Wood. You have already seen him in the first video. So David is chairing this session on technologies for television. David, please.

>> DAVID WOOD: Thank you, yes, hello, everyone. I have a different jacket now. So that is how you recognize me. I'm David Wood, I have the privilege to Chair working party 6C and our hope was this afternoon, we could outline the frontiers of system technology in Study Group 6 for you. By the end of the afternoon, by the end of this session, you should have a knowledge of where the frontiers are in these three areas, image technology, sound technology and integrated broadcast broadband. We have three speakers for you. I'm also joined at the table here, I think we have more up here than down there, but anyway, the colleagues who are giving demonstrations outside, so after the presentations, we will invite them to spend a few minutes to tell you what you might be missing if you don't have a look at the demonstrations outside.

With that, I'll turn to our great, in image technology, Paul Gardiner, tell us about the frontiers of video.

>> PAUL GARDINER: Good afternoon. Thank you very much for coming back, those of you who have come back. My job is to give you a introduction, overview to some of the challenges that we face in video terms. I'm going to briefly look back at how digital video has changed over the last more than 30 years amazingly. Some of the challenges we face in image technology going forward, and try to give you a glimpse of where we are going after HD TV, what is next on the horizon.

I have about 15 minutes to do this. Let's get moving. Except my screen has gone blank. Oh, dear.

It will come back in a moment hopefully.

>> DAVID WOOD: They have all gone blank. (chuckles) there we are.

>> PAUL GARDINER: I must have pressed the wrong button. Let's remind ourselves that the digital technology in the environment has long preceded the transmission to the home that we heard so much about this morning, that we have to go back to the early 1980s when we had the first worldwide discussions which eventually agreed to a famous worldwide standard, Rex 6011982. That was dealing with standard definition 525 and 625 line television in component digital form. It was a world first because we had a unique worldwide agreement on a common data rate for both 525 and 625 flying worlds, and we had scheme with 27 mega samples of video per second and a rather complicated connector which had a pin for every bit. But things moved on.

Before long, we had a serial interface, 270 megabits which made life more practical with a single cable, made it easier to interconnect stuff. That run at 270 megabits. Since then there have been a evolution to cover more high resolution formats. In the ITU R in working party 6C and its predecessors, there have been a number of important recommendations over the years, dealing with a range of image formats. I'm not going to list, describe every one.

But one of the key ones which came in the early '90s, '93, was HDTV studio format for production and international exchange. That was another milestone and it was a common image format, 1920 by 1080 pixels. Since then there have been other important recommendations which have been agreed within the ITU R.

I'll come back to talk about the UHDTV, req 2020, in a moment. But let's not completely forget free DVTV. One thing the ITU R did provide in the last year or two was something that will stand the test of time and that is a interesting report on the principles for comfortable viewing of stereoscopic free DVT, it turns into the psychophysical aspects of viewing and things you take into account for comfortable viewing. Maybe we will come back to that in a few years time when hopefully free DTV will have another resurrection.

Meanwhile back to the UHDTV world and req 2020. That took five years of discussion in the ITU R, but it built on the foundation of the common image formats, and went up in multiples of 1920 by 1080. We have two high resolution formats still, 3840 by    no interlace, that was a big breakthrough to at least say good bye to interlace and other features, frame rates, options go up to 120 hertz. Wider color gamuts and more bits per sample in the studio environment.

The trend has been to increase the spatial resolution, and to illustrate the advances since the world of standard definition, 525, 625, if you consider pixel array, you have a representation here of the standard definition format up to the HD format, 1920 by 1080 and along comes UHDTV at the lower level, 3840 by 2160. For a truly immersive environment, the ultimate is the HK format. That is something that is very suitable for very large screens and will give you a truly immersive environment to fill your field of view. Applications on small devices, not just large screens. But there is more to, to give you a better illustration of the advances in spatial resolution, looking on the left there, you can see an image extracted from a standard definition. On the bottom left you would hardly know it's a cyclist, and the one on the right, that is a extract of the roof showing how the roof tile is not very well, not much resolution when you blow it up. If you go to HD, you can now know it's a cyclist and that the roof tiles begin to look more like roof tiles.

Of course once we come to UHDTV, things improve again. There will be another step up in resolution from the high resolution to HD format. One advantage is extract HD formats from that and provide a good service to HD audience.

Resolution is one factor. But for image quality, there are other important factors as well. In the studio environment, we have a trend to code the samples with higher bits with more bits, 8, 10 and now 12 bits for UHDTV is available. Frame rate options have been increasing. We now have up to 120 or a hundred or 120 frames per second, in UHDTV. You see a demonstration of the improvements in frame rate at the end of the corridor here.

Not everybody thinks that frame rates must go up. Some people in the artistic community think 24 hertz is the right way to do things. It is a interesting debate about frame rate but nevertheless, the technology allows a choice of higher frame rates.

We had a improvement, much wider color gamut in req 2020, to show you a illustration of the sort of improvement. The small triangle represents the conventional HDTV color gamut. The larger triangle represents 2020. The range of colors that 2020 can now display is pretty well matched to the natural world. There is not much that you can't see, and you can see certain colors, there is Siams and greens with UHTV and req 2020 you can't see. It is much more film like, natural in appearance.

But there are other factors as well. The dynamic range of the signal and the display comes into play. In fact, the original cathode ray tube had a range between, greater than most displays have today, 40,000 to 1. But in the future we will have displays which can have remarkable dynamic ranges, maybe a million to one or thereabouts. True blacks and quite high peak whites.

But generally, the displays are now much brighter than they were a few years ago. The brightness originally were restricted by what cathode ray tubes could do. But now with new technology the peak display brightness is able to be very much higher.

One of the basic issues with the existing television system, including HDTV is that it was designed for reference viewing environment where displays had a peak brightness of 100 per square meter. That has been with us for a number of years. Meanwhile even the best TVs have typical brightness of two, three or maybe more than 300 per square meter. The way the camera captures the image is designed around the original reference display, 100.

What happens is that very bright parts of the picture become clipped, because you cannot convey the full range of brightnesses in a natural scene, with the existing television systems. What this means is taking an example of an image here, they typically have a football or rugby stadium or sporting event where part of the scene is in daylight, part is in shadow. While the camera can capture a wide dynamic range, would have capability of capturing 13 or 14 stops, you cannot convey that to the home directly with the current system. You end up clipping the whites, losing the detail.

If you were to try to expand the brightness, you would end up with other problems as well. The high dynamic range system would aim to deliver much more detail in the high brightness areas of the picture. That is what we call HDR or in the ITU R terminology, extended image dynamic range television.

I can't demonstrate that with the limitations of displays we have in the room on the projector here, because we don't have the displays in this room with the high brightness that we will have in the home hopefully within a very short time.

But this is an example of an image which has been nicely graded for conventional displays. Notice the sky. It is blue, yes, but there is no    you wouldn't think there was a cloud in the sky. It looks like a gray day. But the high dynamic range image of that would be able to convey the clouds, the detail, the blueness properly. It will be much brighter picture than you see with this display. Compare the standard dynamic range and the high dynamic range picture.

That is a simulation. But the real thing would contain more sparkle, be more life like. That is an interesting new technology which is now under study.

To make that work, we do have to make some changes to existing television system, because the original way we capture the image does mean that we are limited in how much of the brightness range we can capture. If we were to try to increase the range of brightness using existing system, we would end up with banding in the high brightness parts of the picture.

We need to find a slightly modified way of capturing that in the camera in order to get it back to the home. We could do it. We have many more bits in the signal. But that would increase the studio bit rates which increase delivery bit rate for home. We are talking about starting a new way of transferring the image that would be more perceptually based to capture detail in the white areas, high brightness areas as well as dark areas of the picture.

That is where we are heading at the moment. We certainly do have ultra high definition television. It is coming. Especially in Japan, Republic of Korea and other places are seriously talking about delivering it to the home. But as well as the extra spatial resolution of new HDTV we have a new feature in the active study at the moment, which we sometimes call HDR, high dynamic range or in the ITU R we have been calling it extended image dynamic range television. That is on its way. We have to take into account the number of issues to do with compatibility, with existing studio operations, and how to convert the color gamut between the wider color gamut of the UHDTV formats and existing HDTV formats and understand more about technology bits. It is a interesting area of study. But it does promise much brighter highlights and more realistic pictures.

Almost 3D like, without the glasses, they really have much more impact. There are issues over display brightness and power consumption. We have to consider how much more power might be needed to deliver this and whether we have to think green at the same time and right balances. There are interesting issues to consider for viewing comfort, if we have pictures with bright highlights.

All this is under study now within working party 6C. We are working towards maybe a primary draft recommendation, draft recommendation and it's a big topic of considerable interest. It will be one of the main topics on the agenda in our July meeting.

Finally, the audio has a big impact on how we perceive the quality of pictures. That is the next speaker to describe. The future of video is interesting and definitely has a bright future.

(applause).

>> DAVID WOOD: Thanks, Paul. You will have his presentation on the website. You will be able to look at those things in detail, and impress the people at your next cocktail party.

Let's have a look at the frontiers of audio or is it frontiers of sound? I'm never sure. But Christophe Chabanne, you have the floor.

>> CHRISTOPHE CHABANNE: Hi, everyone. I'm going to talk about next generation of audio, and where we are going beyond surround sound and what we have had over the last ten to 20 years.

First of all, I want to cover immersive audio. Making things more immersive than they have been so far with surround sound, and there has been reasons behind this effort that started about four, five years ago. It came from the cinema side, where 3D movies were successful. It was bringing people back in movie theaters, and the cinema industry realized that the technology has tremendous value, because it brings people in theaters and they can have more revenue.

In our conversations with them, it was very clear that the industry was ready for new technologies. They were thirsty, and they wanted to have some more technologies.

The first thing that we spoke about was 7.1, that is what they requested. That was the higher end of the consumer side and always been a bit of a battle between the home theater world and cinema world.

7.1 was the higher end. They wanted to do 7.1 and we have done that. It's been released, movies have been released, but it has been extremely difficult because 7.1 and 5.1 you had two separate hard drives in order to distribute and they were going to the wrong theaters and they were missing channels.

There was not a right solution. Plus the incremental benefits of 7.1 over 5.1 is noticeable, certainly not significant, sometimes did not play very well.

As a whole, we all had to start thinking on what we were going to do next for surround sound. After many discussions with content creators, discussions with exhibitors and trying to keep it in a realistic frame, we ended up, we as industry, not just the company I work for, we worked on a new dimension, so height elevation was the new dimension, being able to bring things up and down in the room and creating that spherical experience.

Enabling new artistic possibilities also, that was a major driver of the whole next generation audio work. Having discussions with mixers, with movie directors and understanding what they were missing, what was really a frustration for them when it came to authoring and the way they wanted to express their art.

There was many legacy issues or legacy limitations in movie theaters, that we had to overcome in order to bring the experience to another level.

So making it more flexible and simplified, 7.1 attempt was also very key. Once we had this package in place and there has been multiple forms of this package, it became basically fully justifiable. The industry was excited, and they said let's move forward, let's try to mix experiments and eventually release movies.

To understand the next generation immersive audio, I want to make sure that we have background around channels. The definition of a channel is a signal that is assigned to a predefined speaker. I know my left speaker is going to be there, my right speaker is going to be there, center, etcetera. I mix all my tracks in my movies. I bring them into that specific speaker, in the case of movie theaters you have arrays of speakers. You don't have single speaker per channels.

When you go to objects, because that is the new buzzword around next generation audio, object audio is a big thing. An object as opposed to have multiple tracks that are mixed down into a speaker feed, you now keep these tracks independent. You assign them positions, so they move around the room. They have XYZ coordinates. These coordinates and this unique track is sent over to the playback system and the play back system can now know exactly where the sound should be in the room. Depending where the speakers are, it will optimize the room and maximize the room.

That is the benefits of objects audio to be able to describe things, to be able to adapt things to any environment and playback system.

Here is an example of what happened in cinema. You can see there are many configurations, cinema configurations but that is one of them. You see the overhead speakers that have been placed. Each of the speakers are addressable individually. They are no longer arrays. You can have an object through all of them. When you have cinema with 16 speakers in the past with 8 of them being coupled, you now have a proper full 16.1, if you want, system instantly.

In cinema so far, immersive audio has come in two major formats. One is a hybrid channel and object based system. Channels have plenty of benefits. They are very efficient in carrying information. When you have a scene for example with a stadium, and everybody clapping, you don't want to have 10,000 objects of people clapping in order to rebuild the stadium. You want to put that into channels. Objects have benefits as well in being more accurate and more controllable.

The other system is also channel based. It's only channel based, it brings different, slightly different configuration with different layers. I think there are interesting things in this format as well.

The choice of channel is efficient. It is reassuring to the industry who is slowly moving towards object audio but might want to keep going longer with channel based yet still enjoy next generation audio.

Look at what happened in cinema. We today in the cinema industry have 1500 screens. If you combine the two formats and their supportive screen it's 1500 screens worldwide and 300 movies. The next generation audio started about three years ago. It went pretty fast, very quickly these movies started being released and pretty much all the top movies are being released in next generation audio.

In the home, having this vast amount of content coming from the cinema world has created a big pull in the home. People want to bring technologies to the home, bring the same experience to the home.

Blu ray disks have already been released in various formats. Video games are jumping in. So you can have video games in next generation audio with fully immersive environments. The initial next generation audio system is leveraging existing code ex, deploying new code takes a while. It takes a long time. You have to author. You have to distribute and transmit and decode. It takes a while to get established. As the initial step, these next generation audio and object audio have been able to travel through existing codex which enabled instant release of new types of content.

Another thing that was a big thing is practicality. When you saw that picture of the movie theater, with speakers all over the place, it is difficult to bring that to the home. I have a understanding wife and she is not happy about what I'm bolting on the ceiling and everything these days.

But it is just not convenient. It is not realistic to push things very far, even if we look back in the past, 7.1 versus 5.1 adoption was dropping like a rock. Very little 7.1. Fair amount of 5.1 in the household.

Some people have been looking into the development of new types of speakers that enable overhead speakers without having them there. A interesting thing of next generation audio in cinema and what it has done to the home, not even in object audio for the home but 5.1 is having free the cinema world to author new content without limitations, and then you have the new content that is being next to 5.1 for backward compatibility or DVD release. You realize that in the past they were afraid of pulling elements in the surround sounds. Today they are able to pull that in cinema and when that exact same release in 5.1 comes to the home, it is more intense. The energy is more distributed around the room and less screen centric overall.

It has been an interesting turn of event of 5.1 standard 5.1 benefiting of new generation of audio system, although it's not object based.

All these early adoption of object audio and next generation audio, they are small in terms of size. But they are very important in solving the chicken and egg problems. If we want to push these technologies further and have more people enjoying next generation audio, you have to have these basic install base that will allow you to actually start looking to broadcasting, looking into streaming and all these things.

Another thing beyond cinematic content, that is what we have been talking about so far, cinema, but there is more types of content. One of them being live events, and that is a big thing in broadcast, when you look into sports or live entertainments. There are many things you could do in there with objects audio. It is a challenging environment because in the cinematic content you are in post production environment. You have full control of your tracks.

They are clean. They are well recorded in the studio. When you start placing a microphone in the stadium, it is very difficult to capture just the commentator. You will capture the crowd and about everything around it. It is an additional challenge that occurs in sports.

The other challenge also with sports when it comes to immersive is every sport is different. Stadium is one thing, you get reflections from the stadium. When it comes to golf it is completely different. Tennis or formula one. It is a new exploration every time in sports as opposed to cinematic content where you are able to create something flexible enough for about every movie, because the production process was the same.

When it comes to end point, we have many types of end points. But multi channel is what we have been talking so far about. But there are two other categories. At first I wrote stereo devices, but it's not the right description. If I wrote it as devices, with self contained speakers, it is a device that has speakers in it. It is more broad. That would include potentially mono phones or tablets or anything like that.

The other kind is headphones. That is also something very interesting when it comes to objects audio, because headphone rendering using the way to bring perceived sound spatially using time delays and HRTS has been locked to the channel world, where you have fixed position and you have to deal with only five positions in order to attempt to hear something.

When you deal with object audio you have coordinates. You can recompute things on the fly. You can make much more realistic headphone experience, so that is an interesting area that the industry is looking at.

I spoke about immersive audio but there is another one which is very interesting, personalized audio. It is a interesting use case, because you are not addressing the multi channel crowd. You are not addressing the people who are willing to put wires and speakers everywhere. There is a cost to that.

Personalized audio, here is the difference. Immersive is about dynamic spatial information. You are moving an object with its spatial information and you use the speakers that you have around to maximize its rendering. The objects come and go in movies, there is something flying by, airplane flying by and it's gone. The object goes away.

The consumer is experiencing what the content creator did and has no control over the movie. In the personalized world, it is quite different, because you now have as a consumer control over what you are listening to. You have different ways of managing the content. You can choose what you want to hear. It becomes quite a different thing. It is also not dependent on the complex multi channel system, mono device or stereo device will work just fine.