Okay, ladies and gentlemen, thanks for attending the session today. My name is Benjamin Feinman. I'm a senior product manager in the QuickTime group, and I focus on mobile media delivery solutions. Today, we're going to split our time into two presentations. One from Erickson, Eugene Sarmiento. He's going to take you through in a little bit more detail the components of their mobile delivery solution and talk a little bit about how Apple integrates into that. And then we have Dr.

David Guyard from Buick Telecom, which is a major European operator that we've been working with on video services, and he's going to take you through what it takes to create and deliver content from the operator's perspective. And at the end, we'll do just a little bit of Q&A. So I'd like to invite up Eugene. The clicker is now yours.

Good afternoon everybody. My name is Eugene Sarmiento as Ben mentioned earlier. I'm a product marketing manager as well, but mainly focusing on Ericsson's content download and streaming solution. How many of you here were in yesterday's presentation?

And in terms of opportunities that we see in this space, okay, according to Business Wire, right, this is on the fixed side, there was 7.87 billion streams in 2003. That's 100% increase from 2002 or the year earlier.

And we see that same expansion or growth going, this is also going to happen in the wireless space. And just also from another analyst, it's strategy analyst, they see that in 2008, over 150 mobile users will be viewing video content and video services generating $4.6 billion. So this is one of the analysts. Some are saying 6 billion, some are saying a bit higher, some are saying a bit lower, but there's basically a lot of revenue to be made in the wireless space.

Okay, now to give you an example of what or which operators are actually mature in wireless media, I have two examples, one from Japan where they have launched a lot of 3G handsets. They're going to increase their subscriber base by six-fold. And one of the operators there, which is NTT DoCoMo, they launched a service called M Stage Visual, which subscribers pay for on a monthly fee, and they also charge for a transmission charge. So the monthly fee is $1.8 to access all the content in their portals, and the transmission charge is roughly 14 cents per minute of data.

Also in the same region, we have South Korea Telecom, who's also earning a lot of money on mobile media. They're earning an average of 7.5 US dollars on top of their existing revenue because of the streaming and download capabilities in their network. Benjamin Feinman, Eugene Sarmiento, David Guyard Also in the same region, we have South Korea Telecom, who's also earning an average of 7.5 US dollars on top of their existing revenue because of the streaming and download capabilities in their network. And it was surprising for the CTO to see that people want to view TV on a mobile phone. And this is just an example of what all the other operators around the world could be. This could be their situation in the future.

Now apart from the opportunities, Ericsson has also done its own studies on what services subscribers would like in the mobile media space. So we surveyed 40,000 people around the world in 11 countries actually, and this was done by an internal group called Consumer Lab, which just looks into user behavior and user statistics. And specifically for streaming and download of video content, what they found out was there's three main categories that people look for. It's the music category. the news and the sports categories.

So, in the end-to-end chain from creation of the content to the actual delivery to the handset, what are we doing in these different processes or different stages of the delivery of content? From the content creation side, as I said, we have mobility world, we have partners, we create packages of applications that we match towards operators.

So, if we approach, for example, a company called Amena in Spain, the way we approach them is we have a mobility world representative and he says, these are the things that we have, a thousand different things, pick and choose from them. Or we give them also information about how these services are, if they're successful in other markets, why they're successful, what sort of revenues, potential and business case.

What these applications have and how it would impact their network. So, from that side, we have mobility world as well as the tools that we sell with our solution to enable operators to create content on their own or resell that content or actually buy content from content providers.

On the other side, from the terminals, the activities that we mainly do is testing. IoT testing. We do IoT testing with our own handsets as well as all the other handsets from Nokia, from Motorola, from NEC and the rest of the terminal suppliers across the globe. So, we do support for that. Now, our main offering and our main, well, actually, our core competence in Ericsson is delivery, is integration and is the wireless network. Okay. And that's why we focus on that.

In terms of transport, we make sure that our platforms have the features and functionality that interwork with the core network and radio network, that the quality of service is maintained end-to-end between the handset, the radio base stations, the radio base stations connected to our switches, and then the switches connected to our streaming solutions. Okay. So, we not only address the features on the product itself. But we also.

Develop features on the radio and core side. Apart from that, we do integrate or we focus on business logic. The product that we sell focuses on the business logic that will allow operators to charge, to provision, and manage the different nodes that we're introducing. Because this is where we're coming from. This is where Ericsson's competence is. We've been building mobile network technology and integrating this. For more than 100 years. So, this is why we're focusing on this area.

Okay, so knowing our role and our function in the end-to-end chain, what's the challenges we see and what challenges do operators see? And I think my colleague from Buig will be discussing their views as well. But in our point of view, there's four main areas that are challenges in delivering mobile media content. First and foremost, it's having the right content and the right applications in operator networks. It's still not that easy to find that killer content. Then there's the handsets, the problem of making sure that every single content is delivered.

And then there's the operator networks. So we have to be able to integrate the operator networks into handsets, and those handsets can decode and play the different formats that are available in operator networks. But apart from that, we also look at end-to-end integration. All right, let's say I have the content out and my partners are set. I have the handsets ready.

Next thing that the operator asks for, or the next thing that is asked for, is how do I integrate the solution? How do I make sure that the billing interfaces are set, that the prepaid interfaces are set? How about my integration to customer care? How about my integration to network management? And finally, of course, there's how should I charge for this? And this is still an unknown area for a lot of operators.

There's so many ways you can charge for this. It could be subscription-based, event-based, it could be volume-based, it could be category-based, it could be one-time fee, or it could be combined with MMS, it could be combined with voice. There's so many things that could be possible on a way to charge for this. And that is a challenge because this will actually make or break the service.

All right, so what's our solution? What's Ericsson's solution for this? Well, basically, we have come up with a solution, an end-to-end solution that addresses or aims to address all the content, all the relevant formats, all the relevant protocols, connecting that content to the handsets or the devices that will receive that content. And what we've done in Ericsson, or what we focused on Ericsson, is the development of a network integration node. And what that does, of course, is do the integration aspects, basically develop features that bring about our core competencies.

And what it does is it allows us to have streaming services. Now, it could be a Sun streaming server or it could be a QuickTime streaming server. What we want to be, what we want the system to be is agnostic to its streaming server or streaming platform because we want to address the content space.

So if an operator says, "I want this type of codec and this other streaming server provides it," then we put that in. And I think QuickTime is really helping us there in terms of the new protocols that they're supporting and 264 is coming up. So it could happen that operators will start asking for that, and we will deliver it with that.

We do also provide content upload for the content providers, content storage rating for that content, and a subscriber database to hold the information about which services that subscribers want. Apart from that, I'd like to explain to you how our solution, which is ECDS, fits into the radio network and the core network. The way it works today is we have support for both circuit switch and packet switch, which is quite a key strength because a lot of the existing solutions today only address the packet switch space. So we have a lot of different solutions.

We have a lot of different solutions. This is more wireless technology infrastructure. So you have your radio base stations connected to controllers, controllers connected to gateways, and then finally it's on the Internet or within the operator's walled garden. And then we have integration to content creation, content management, HLR, and charging systems. Okay? So that's the first part.

So, I told you that one of the challenges is integration and that's why we focus on that. We make sure that we have integration points to the billing gateways, we have integration points to network management systems. So, when we talk about network management systems, it could be HP OpenView, it could be Compact TAMIP, anything with SNMP support.

We have integration points towards a AAA server where basically we get the IP to MSISDN mapping. Now, this is important because on the wireless network, you need to find out what MSISDN is so that you can charge for it. So, it's like if you look at it from the web perspective, it's your key to access streaming services.

Okay, apart from that, we're looking into integration with or creating functions for MMS. Okay, integration to that and VIG is a video interactive gateway which allows us to have circuit switch services. Okay, on the encoder side, we do sell products based on QuickTime and we do promote QuickTime based encoders like the Final Cut Pro and the QuickTime Pro, is it? Yeah, QuickTime Pro. But we do have also other partners that we have. we package along with our solution.

Okay. Now, let's say we address the integration part. How did we address the application part? How do we make sure that there's enough applications out to the handsets? Well, we have a Mobility World program where we focus on making sure that developers have the ability to create. Then, after the creation process, we take those applications, verify it in our test labs, make sure that it's tested properly, it integrates with operator networks, and then we go to market with that.

Okay. From the developer's site, we have www.ericsson.com slash Mobility World. Okay. And there you'll have all the SDKs, not just for streaming and download services, but you have APIs and SDKs for location-based information, for parlay services, which is voice-related, and there's also SDKs on how you can push SMS. So, basically, anything with regard to mobile network technologies and providing applications on that, you can find the development tools there. We have support services, forums, and we, of course, have success stories of search. We have certain application providers who have made it big with us. Okay.

On the verified side, we try to as much, since we package Applications and services towards operators, we need to make sure that it's tested very well because basically we do first-line support for those applications. And the first thing the operator says is, Erickson, why doesn't your application work? So we need to do very stringent tests on these applications, and we do that through a process, and we have a lot of things that we look into in the testing process.

After the testing and verification, then we go to the market. And there's roughly 30 centers worldwide where we have mobility world representatives. And these mobility world representatives, the only thing they do is sell applications. They go to each operator and ask them, "What do you want? Where are your services headed? How can we help you?" And that allows you, as an application developer, to push your content without actually having the manpower to fly across all the globe. So we do the front end for you.

On the terminal side, we're heavily relying on the standard, as is Apple. And we've been drivers for the packet switch streaming standard for a while now. During the times when it was three or four years ago when I was trying to sell streaming solutions, but a lot of the streaming solutions were proprietary. There was a proprietary codec from guys from Seattle, proprietary codec from this guy, and there's like four solutions, all proprietary.

And we couldn't live with that. As Ericsson, being a driver for standards and being a driver for GSM, we needed to make sure that there is a standard way to transfer these files and encode these files. And we did that through the packet switch streaming standard. So what came out of that is UA prof information, which basically allows you to find out the terminal capabilities of the phones that you send.

And we also have a standard for the data that you send content to or is retrieving content from you. There is a specific standard on DRM, but it's very limited. It took them a day and a half to hack it, and now it's not that difficult to hack that. And then for the codecs, we have AMR, narrowband, wideband, AAC, which is the next generation MP3, H263 and MPEG-4. H264 is coming in R6. We have scalable polyphony, media.

And as you can see, all the things that Apple has been working with, RTSP, RTPSDP, these are things that is in the standard as well. Apart from that, we have for scene descriptions, SMILE 2.0, and for text and images, we have the following. You can get this off of the 3GPP site, www.3gpp.org. You can get the latest information of what's happening in the standards. Okay, so that's addressing the terminal space.

I think now I'll focus mainly on what we've done with other customers, or some of our customers. I'll focus on three because the rest, we actually have roughly 15 customers, but I can't really disclose them to you. But I have three samples that I will be discussing. All right, so we have Mobitel in Slovenia. They were our first customer, actually, and they started launching their services in 2003. And in terms of services, they listened to us and put news, put sports, and put music content there.

Apart from that, they had their own thinking of putting a lot of adult content, which is really taking up there. Don't know why. Okay. But yeah. They also have live TV. And in their launch, during less than two weeks since the introduction, they had 100,000 clicks to various erotic content.

Well, we'll see maybe that's the way forward for this, but there was a lot of interest in special weather programs where people were stripping, and there was a lot of services around that. Now, basically what happened is after they launched, after a year of launch, they were able to penetrate 40% of their subscribers. So that's quite a quick penetration rate.

I mean, that's one of the best they have, actually, in their span of launching services. And right now they have roughly 445,000 users. They are growing at 8,000 per month. So big. So they could reach 100 pretty soon. In terms of content providers, we have given them some content providers, and we've given them support in providing them with the content that they're looking for. Okay, so we have discussions now ongoing with Sony Music and MTV.

Another operator that we work with is Amena Spain. They have been very good supporters of Ericsson. And they launched services such as TV programs from Canal+. And they have also music, sports, cinema, and fashion content. In terms of content providers, there is Grupo Prista, which is a media company in Spain. And the way they charge for this is basically... Ah, one thing that I missed on the other slide was charging. All right.

The way they charge for these services is they charge on the time the person is browsing content. So it's time-based. For Amena Spain, they charge on time-based charging as well. And there's no specific charge. But they're now looking into prepaid support. And once that happens, I think then they'll go into the broad base. Because if you look at Europe, it's mainly prepaid subscribers there.

Okay, then there's Telefonica. They launched in December 2003. They launched almost the same time as Amana because they're competing with one another. And they have the same set of services. They have Spanish TV channels. And in terms of charging, they do have prepaid support today, and we did that for them.

Apart from the success that they have with the service, what I haven't really said here is the integration that we've done. Integration into fault management systems, integration into their content management systems. And we've been very successful in quickly deploying that. Actually, we had one customer where they wanted to launch within a month and we were able to deliver.

Actually, it took us one week to install, two weeks to integrate. That was one of our best times in integrating a platform for one of our operators. And the price levels for Telefonica is one euro per video. Another important customer that we have is Sunrise. And the difference between Sunrise and Telefonica is that we have a lot of different customers.

And the difference with this one is we provide them with a hosting service. So they don't have anything. We provide everything from the platform, the streaming servers, and the applications, and we aggregate that content for them. They only pay us for usage, basically. And this has been very successful.

They launched in July 2003. We've used a lot of the Mobility World Partners, EMW. That's Ericsson Mobility World Partners. They price per clip. That's the difference with the rest of my previous examples. So they price per clip. And they also have subscriptions for groups of content. So you pay a subscription fee, let's say HBO content, and then premium prices for each of the clips. We recently launched with them Emuse, which is a music service. And this music service, allows subscribers to look into, let's say, Britney Spears' website, and find out the latest information about her, her ringtones, her music, and also the music.

Ask for MMS content that they can show their friends. And we had in Asia last year a slogan, which is make money by making people famous. Because how operators will make money is if they can make the subscribers famous with their friends. Oh, look, I have the latest football goal in Euro 2004 by, let's say, Portugal. Or maybe here, if it's the San Francisco Giants.

Oh, here's the latest tackle. All right. Summary. There is a lot of opportunities, and it's happening now. It didn't happen as much last year, but this year we see a lot of positive results, both with our operators and with operators that have launched streaming and download services. Ericsson wants to be your business partner, and we want to take you to our operators. We'd like you to join us in Ericsson's Mobility World program. The link is there. If you need to get contacts, just please approach me and be part of the mobile and media convergence. Thank you.

Thank you, Eugene, for an excellent presentation. So coming up next, we've got Dr. David Guyard, who's going to now talk about a similar topic, but from the perspective of an operator. So with that, I'll hand over the clicker. My name is there. It doesn't move. It doesn't change. I work in the Department of New Technologies at Bouygues Telecom. It's a French operator. I will make a brief presentation in a moment. The topic of this presentation is video services of a mobile network.

And here are the different aspects I will deal with. First, I will speak about my company, because we are a small operator. And it's interesting to know why we are making video right now. I will talk a little bit about iMODE. But my colleague in next time slot will have a deeper presentation of it, I think. So I will be rapid.

I will also talk about the network constraints and the type of networks we are operating, and also the handsets, because it is one of the most limiting devices to deploy video. And then I will go deeper into the subject, that is I will present two simple services that are launched, that is video mail and video downloads. And as a conclusion, I will make some Talk about future trends, both in the content aspects, how to encode the content, and the transport problems, how will the transport evolve in the future.

And that will be all. So Buick Telecom. In the previous slide, there was a way to pronounce it. It's very difficult to pronounce it for American. So the network started in 1996. We have 6 million subscribers. It's approximately 18% market share in France. And we are operating GSM/GPRS network. We have a nationwide coverage for both technologies.

And since we are only French operator, we have a lot of roaming agreement around the world that makes us possible to have access to iMODE services while we are being here in the States. So it's quite important for us. We intend to evolve our network. In 2005, we are making some sort of test in France with Edge technology, which is a very important evolution for us because we will, from one day to the other, gain a nationwide coverage as a position to 3G networks that most of the time operate on spots.

And with Edge, we will gain a four-fold throughput. So very interesting to have bigger files to download, for instance. In 2002, we launched iMODE. Internet for the mobile phones, but in a simpler way than, for instance, WAP, because it reuses protocols existing for the web. So mainly it's HTTP based, and it uses email instead of MMS.

One particularity of it is that the operators specify completely the services and also the phone functionalities. That is, with other operators, we exactly define the layout of the phones, the functionality that's inside, the keyboard, everything, so that the users exactly know from an ergonomic point of view how to operate this phone, which is not always the case with very sophisticated phones.

The idea behind that is to make it very simple for everyone to use the services. So far we have mostly 10% of our clients that have shifted to iMODE since we launched it. And finally, well that is the topic of the presentation, we launched with this one, we launched the video services end of May of this year.

So back on iMOD in a few words. It is a technology that is licensed from NTT.com. We deployed it in 2000 in Japan. It is licensed to, I think, approximately 10 operators worldwide, mostly in Europe and one in Australia. It has 42 million users, so it's quite important. Well, maybe 40 million in Japan, but it's increasing overseas.

I think there is a problem with the slides, something that comes again. Concerning the network, For packet data, we use GPRS. That is, we have throughputs in the range of 40 kilobits per second in the downlink and 20 kilobits per second in the uplink. It is the best effort network.

That is, well, if you don't have any radio resource, then you have to wait a little bit. And most of that, you also have the fact that voice may be primitive over data. That is, that if a lot of people want to place phone calls, then you have to wait for your content.

Handsets is a big limitation today. Most of the JPRS handsets are 3+1, that is they can support 30 kilobits in the downlink and 10 in the uplink. It's a little less than the theoretical throughput available. For the IMOD onsets, we specified that they should support 3GP file format. That is for the video, they use H.263 on MPEG-4 video codecs in resolution SQCIF, that is 176 by 144 pixels.

Joseph Kucif. And for the audio, we use two codecs, one dedicated to voice, which is IMR that comes from the GSM world, and the other one which is AAC. We have a very simple data protection for Wordlock. It's very simple but still very efficient because we specify the phone in a manner that the content can't go outside the phone. So even that very weak protection is enough to convince the content provider to be confident in the technology. And from the recording point of view, we have found we're equipped with camera so that we can send a video as an attachment to email.

So if we look at-- there are two constraints when you design video clips for mobile phones. One constraint is the time that people have to wait before they can see the clip. That is, if I go onto a website, I click on the URL for the clip. If I wait more than 30 seconds, then I don't wait more than 30 seconds. I quit. We made some tests two years ago, and that was a conclusion for 90% of the users.

When we combine this constraint with the fact that we have a very limited throughput on the network, we end up with a constraint on the file size that is approximately 100 kilobytes. And if you look at the table, then you can see that the download duration time on Edge is approximately between 20, 30 seconds. And the upload duration is more important. It's 80 seconds. One thing interesting to see is that the phones today are not operated by a multitask operating system. So when I download a clip from the website, well, I wait, but I can't do anything else.

If I sent an email with a video attachment, it's the same problem. So from a user perspective, it's very problematic, and we have to find some solution. And one solution is you increase the bandwidth, so you go to Edge, and then you have shorter upload/download time. And that is the way we're going to.

We also have a problem with video quality because if we have very small files, if we want to have a duration acceptable that is in the range of 15, 20 seconds to give the users a piece of information, not only a still image, then we can't have video bitrate very high. We are in the range between 32 to 64 kilobits per second. I will show you later on what is the type of quality that we can have on very hard content that is on football games, for instance.

If we look at video mail services, it's a service very intuitive because it's like an email. You just shoot video with your phone, and then you open the email. Email client on the phone, you put the video as an attachment and you sent it. Very clever, very simple. Or if you have the authorization to do so, if the video that you downloaded from the web is not a lot, then you can forward it to someone else.

I said iMod is reusing some technologies existing on the web, so what is good with that is that you can address any email on the web, so almost everybody. And also you can address some MMS recipient if you have their phone number. There is in fact a getaway into the network that makes some translation between the email format and the MMS.

On the recipient side, well, you could have different case. If it is inside an iMod network, then it's very easy. There is no problem. If we talk to an OMS client, that's what I talked about. And if you talk to someone on a PC on a fixed network, then you can look at the video using, for instance, QuickTime, which is very easy to get.

Video download is the other service. So it's HTTP download. It has a little overhead during the transportation, so we had to think about that when we look at how to charge the content. And it's very important because the users pays for transport. Here I put different captures in order to show you the steps.

The idea is when I go onto the iMODE portal, there is-- maybe I'll show you in a moment-- there is a touch on the keypad. And the keypad, if you press it, you just go onto the website. And this is the portal you see on the left. And then it's very similar to web navigation. So I'm going to show you a video clip. If we can switch to the camera.

We can put the light a little down. I will play it again before making some comments. So as I mentioned, there is a fixed-- You can see the length of the video here. Well, it's a little better on the screen of the phone, because there is degradation in the process.

But there you can see the size. Well, you have duration of about 12 seconds. It's a video clip encoded at 64 kilobits per second. And there is no audio on it. So that gives you an idea of the kind of quality you can have. That type of content is very hard to encode because, well, there is the grass with a lot of tiny details. There is a lot of moving objects that are the players and also some small objects like the ball.

So you end up in having that type of quality. Could we switch back to the other, please? So if we look at how all that is charged, we have on iMOD something very simple. Well, at least in the beginning it was very simple, but it's very hard to live with it while it evolves. It is simple in the fact that we charge one cent by kilobyte. So if you just make email, that's right, because an email is something in the kilobyte range, two or three kilobytes.

And when you start to download files that are in the range of 100 kilobytes, then it makes you something like a euro or a dollar, which is approximately the same for a video, and starts to be really expensive. And on top of that, the economic model that the content provider want to use is the same as the Ringtone economic model that is they want to charge for a video. So if you subscribe for a month, and during that month you could get a number of videos, and at the moment it's in the range of two to five videos.

So I will talk a little bit about content creation. Some consequences of the file size that we have. Since it is very limited, then we need to work on the video. And while it's not our job, but we have to help the content providers to work on that.

The story about this is, in iMOD, we have content providers coming from a reasons that are very different from video. We have people making newspapers, we have people selling ringtones, and all of a sudden they want to put some video on their websites. And we have to tell them, well, if you want to do so, you have to use such tools, you have to follow such rules, and, well, we have to define those rules.

So video sequence must be very simple, that is, not too many sinkers. There might be a contradiction with the kind of trailers that we can see where everything is going very fast and in the end you don't see anything if you encode it right away. We have to also optimize the sound. The sound is something very -- to which the user is very sensitive, in fact.

If you put two videos with soundtracks encoded at different qualities, then although they are the same, the users will prefer the one with the best soundtrack. So we have to live with that. And there are some contents. For instance, if you have a news, then someone is speaking, you can use without any problem a voice codec.

But if you have someone singing, if I start to sing right now, if I want to encode that, then if I want to encode that, if I use an iMarketDec, it's the end of the world because there will be some glitches in the soundtrack and it's unacceptable. So we also have to define how to encode that. And to do so, we have the AAC codec, which also requires a higher bit wedge. So there is a compromise between the amount of data that you put into the audio and the amount of data you have in the video.

For the digital right management, there are several ways to do it. At the moment, the only tool that we have to create content and to have the protection is QuickTime Pro that completely supports the file format defined by Entity.como. And while it can be seen as a limitation for certain content providers. But it enables the protection. If we have a look at different parameters we can use, here there is the different profiles optimized for duration or video quality. And you can see the video quality and audio.

I'm looking for the-- OK, duration is there. So if you want a high quality video that was the parameter set for the football sequence, then you encode at 64 kilobits per second with 10 to 15 frames per second. That number of frames per second is largely sufficient due to the definition of the image. And you can use an IMR codec or you can not use an ECODEC. if you just want to emphasize the action.

And then if you don't mind that much about the video or if you have something very static like head and shoulder sequence, you can encode the video at 32 kilobits per second and you end up with a duration that is twice the duration of the high quality video.

If you look how it works in QuickTime, well, when you export a file, you have a source file, a clean source file in QuickTime. When you export it to 3GPP file format, you have different tabs in which you can select the format of the file. So there it's the mobile MPEG-4.

It's the format defined by entity. It's exactly the same thing as 3GP, except that there is the DRM on top of it. You can define the definition, also if there is a nodule tracker or not. And then, You specify all your encoded audio on the video, and it's very intuitive to do it. So it's a really nice tool.

Well, so all that is what can be done today. The service has just been launched, so we don't have any figures to provide to say, well, it works very well, or the users are really satisfied with the quality. Still, we are working to define the service we will have in the coming years, taking into account the evolution of the network. And that's why I'm talking about the future trends.

Concerning the contents, one big need is the improved encoding tools. Someone yesterday mentioned the fact that MPEG-2 is now all standard, that improves every day in coding efficiency. MPEG-4 has hit the market very recently, so we evaluated a few tools and we have similar results, but we think it could be really improved by the use of two-pass encoding, for instance, better analysis of the sequences. And also we think that the encoder, the core of the encoder will also evolve.

We hope so. We hope there will be a lot of progress done. Then we can increase the file size in order to have longer clips or to increase the quality. So here there is a trade-off, what the users in the end want. And it's not a very easy question because depending on how we charge the content, what if I charge one euro for ten seconds of video and if I charge two euros for ten seconds of video, but with a file that is twice the size, then maybe the users, although the video quality is better, they won't understand. So we really have a problem of how to charge the content.

Then what you saw on the phone is a phone with a QVGR resolution. So you put a QC content on QVGR. You can zoom it. You can make it full screen. But then you have some artifacts due to the extrapolation. One interesting thing would be to increase the image definition in the 3GP file.

If you do that, then if you go from QC to QVisio, then the surface of the image is doubled, mostly doubled. It means that if you want to keep the same quality, you have to double the file size. So it's a never-ending problem. Whatever you change, you have to improve something.

So we can look at the encoding tools. We can also look at the codecs. There is the H.264 MPEG-4 AVC that is being introduced. There has been some press release and so on. We should gain from that 50%. For the same quality, we should gain 50%. But we are waiting for something working for mobile phones, because we have the encoder problem and also the availability of colleagues for the phones. And then another improvement could be the use of AAC+, where we could achieve higher audio quality for half the bandwidth. So also an improvement we're waiting.

For the content, there are two evolutions. One is progressive download, because as mentioned earlier, If my file size is too long, and if I have to wait until the end of the file size is downloaded, then maybe that waiting time is too long for me and I quit. So the idea with progressive download is to start downloading the file, and as soon as there is enough data into the buffer, start playing it.

So this improves the user experience with a click and play. I click, it plays. And it allows encoding of some contents with higher bitrate than the available throughput on the network. So that's very interesting if we want to increase the quality of the image. Then the other way is to look at streaming for video on demand or for live channels.

It's possible today, but one limitation is that if we have one user streaming at 40 kilobits per second in a cell, then it might be the only GPRS user in the cell. So it's a strong And it will be possible more easily when Edge is deployed and, of course, UMTS.

A problem with that is we need to introduce new elements in the network. We need to introduce streaming server. We need to define an architecture that is That is good, both for us and for the content providers, because there are big content providers like TV channels. They can host streaming at home. But there are small content providers. We doubt that they will be able to put the money and the knowledge to make the investment of the money and the knowledge. So we have a big problem.

So that is the where. The Omni is also in the range of this. It's also the same kind of problem. A big argument that is often given, the argument is, well, if you put some streaming, if you have live streaming, it is for specific events like a football match.

That is, when there is a football match, you will send SMS to the users and you say, well, there has been a goal. And what you expect is that everybody is going to go to the streaming server and make some requests. And if the dimensioning of the network is not good, then you have a big problem. It doesn't work.

So, how many is something very, very difficult to resolve. One last point is, one of the last point is the interoperability problem, as was mentioned in the previous presentation. You can have, you can find a lot of brands of streaming server that may not be interoperable between them. And we have to be sure that the day we launch the service, we don't have any problem with that because the user won't understand.

And last evolution about the content is how we charge it. Well, I talked a little about that. So if we increase the file size, then automatically the price tend to increase with the business model we have today. Another problem might come from the standard bodies, because if you read into the 3GPP standards, then the MPEG-4 profile that is used is limited in bitrate. And theoretically, we could not go beyond that bitrate today with what we have in the specification. So we have a problem.

It might not be a problem because with the phone I show you, you could put video with higher bitrate. So we just could say, well, we don't mind. Except if one day we have a manufacturer that says that, well, it was not in the specification and I can't support 128 kilobit per second.

For content providers, we have a problem of compatibility. That is, if I launch a terminal today that supports a certain amount of codecs, And one year after I start to ship H.264 terminals, then a content provider, what it will do is it will say, well, I have one encoding chain. I will shift my encoding chain to H.264. And for the guys who bought a phone one year ago, then it doesn't work anymore. And we have to deal with that. And every time we think about the specification of future phones, we keep that in mind.

So also the content creation tools are important. I talked about increasing the performance of them. There is also the fact how to manage different profiles for different phones. The last point we foresee as a big problem is the DRM aspects. There are already a lot of phones on the market shipping with memory cards.

And on those phones, DRM is not always handled properly. And we have to be very careful about that because without any protection solution, there is no content provider that follows. The future trends and difficulties that we will have to face as operators and also as providers of solutions for encoding and terminal solutions, for instance. Okay. Thank you.