 So we have Mobile World Congress 2018 and who are you? Hi, I'm Tin Lin from Mediatek corresponding to Protamarketing. And you're launching the Helio P60. So this is OctaCore. Yes, and this is our most attractive SOC platform for mainstream and to high-end platform. It's the most powerful, right? So it's a Quad A73, Quad A53. Yes, we have 4-bit code, 4-small code, and for the 4-bit code it can help you to run some applications in higher performance. And it's also Mali G72, right? Yes. So very powerful GPU. Yes, Mali G72 can help people to enjoy some games without too much, with very good performance. And here you're showing, for example, that the computer vision engine right here is able to do very strong bokeh. Yes. You have a single camera? This is a dual camera. Because dual camera is a very hard feature for the mobile phone maker and they want to enable it. So we help them to pre-integrate the dual cam bokeh feature in our microphone. So they can enjoy this kind of technology without too much problem. And it does a very strong bokeh effect using the dual camera real-time computer vision? Yes, yes. We have two ISPs that can help our customers to do real-time bokeh and provide very good efficiency. You can see right now here. And what do you show over here? Okay. So here is another technology we provide to our customers. It's embedded in our P60. It's a hard way face detection. So you can see here, it can detect the face very quickly. And in different scenarios, it can identify the face very quickly. Compared to other solutions, you can see our face detection can be used for different scenarios. Alright. Cool. So that's important to focus faster. What is it for? Focus faster and find the face very fast. Alright. What are you showing over here? Next is another key feature of P60. This is our AI technology. So right now the P60 you run on these, what is this? This is our internal phone. We use it for our internal development. So it's a development phone for internally for media tech? Which... What kind of display do you have? This display is... It's not 120Hz, no? No. Can you support 120Hz? Can you support high? Yes, we can support 120Hz. So very fast UI, fast display? Yes. Alright. And this is about the AI? Yes. So what do you enhance with AI? Because AI can be adapted in different applications. So from media tech point of view, we would like to help all of the ecosystem player can integrate their AI application on top of our platform. So we support the standard enjoy neural network SDK and also help some key partner to integrate their AI application in P60. So you have a neural network on the P60? Yes. So how much performance do you have with the neural network? That depends on the application. But because the key point for the standard enjoy neural network SDK is a standard interface. So everybody can use this standard interface with porting to every platform without too much care. Is this called the mobile APU? Yes. We have in P60, we have mobile APU. This APU can help AI application to provide some hardware accessories function. So for example, face unlock, face recognition, for example, what else is possible? Okay, right now we have several different applications. This one is running on GPU and this is face recognition. So you can see the face recognition is very quickly. You can identify your face less than 0.1 second. Okay. So face unlock? Yes. And over here? Here we have 3D and emoji. Sorry. Yeah. Oh, sorry. So, an emoji is like an emoticon animated from your face, right? It's a new trend. Yes. Where people do all these advanced front camera bokeh, really? Okay. This is the single-camera bokeh feature developed by AXO. So, how does that work? You can see, it can identify you and blur the background. And this can do it in real time. So you can preview the front camera bokeh effect for this application. You can see the background is blur. Let me show you. Try on you? Yeah, you can see the blur effect is enabled. Okay. That's pretty cool. Yeah, yeah. Awesome. That's pretty cool. Because this is a single-camera bokeh solution. And it's also software solution. All of the OEMs they can adapt without power cost. Nice. So you don't need a dual-camera bokeh? Yeah, yeah. Thanks to the mobile APU, the neural network. Yes. This is running on our APU. So when you're running on our APU, the power consumption is about one-tenth of the... compared to the running on the GPU. Oh, one-tenth. One-tenth. One-tenth. One-tenth. Cool. And what are you showing over here? Okay. Dual-front camera, no? Yeah. You can... This is our reference solution for the 3D Stratolite. You can... Use the 3D Stratolite to do some face recognition. Ah, so this is face unlock using AIR? Yes. Infrared. Infrared, yes. And because it's... So the infrared beamer and infrared camera? Yes. Right? Yeah. Infrared camera. Yeah. One IR camera, one RGB, and one laser dart projector, and one flood illuminator. Nice. How do you cast this hardware to do this type of face unlock? Oh, this module. Maybe the factory price is $5, $10 more? The phone maybe? Yeah. We're working with another vendor. As for the price, we don't know yet, but it's a little bit expensive, but the accuracy rate is quite high. Yeah. All right. And this stuff can also be at the back of the phone, right? It doesn't have to be only at the front. Yeah. It can be at the back. Yeah. But the application you usually use in front, in the front? Yeah, because you want to do the face unlock. Yeah. And also maybe you will use it for the payment, so front camera will be much more reasonable. But what I think maybe is more interesting is to do a 3D scanning of the space. You want to... Like Project Tango, right? Yeah, yeah. And Project Tango is in front? Yeah. Apple is at the back? Yeah, yeah, yeah. Right? But that's a different user scenario. Yeah, different user scenario for the rear camera. The use case is just like Project Tango or maybe AR solution. They maybe came for some maybe application. Yeah. Yeah. Yeah, AR related application. Can we walk more and look around? Yeah. So you have some more demos over here. So you can do high quality voice over LTE. Yeah. Vi LTE is video over LTE, I guess? Yes. Wait a moment, I need to find someone. Switching 2.0. What is that about? Yes. Say other phones have two antennas. So if you block one of them, you lose some signal. So the transmission is lower. But we have three different antennas. So if you cover one of them and you block it, it switches automatically so that you don't miss the signal. And you don't lose any of the speed? Yeah, you don't. Full speed? Yeah. So like for example the iPhone at one point, you could not use it with the left hand. Yeah, so that's... That was what you fixed. Yeah, we fixed it. And provide much more practical solution to the phone maker. So that's part of P60? Yeah. And you have a power efficient modem? Yeah. Hi. Hi. What we're doing here is a real time simulation where we're using one of the MediaTek Helio features of connectivity, which is the Wi-Fi hotspot. And what we have here is an LTE signal generator which connects with these two boxes. In these two boxes we have one phone with the chip of our competitor and then another phone in the other box with the chip of... The P60. Yeah, with P60. And what we do here is that... Oh, sorry. So it's at different bit rates? Yeah, at different bit rates of downloading we see that our technology, what does... It consumes less power than our competitor, which means saving up to 40% of the power. And this is a real time demonstration. 50% power saving at 25 megabits. Yeah. For Wi-Fi. For Wi-Fi hotspot. But we went a little bit further than that and we also tried all these features. LTE standby. I thought hotspot. Voice LTE, voice INLT data link. And in all of them, our chips were saving power in compared of our competitor chips. Is this something you can do with AI? No, it's not AI. That's something else. It's relevant only for more than... Is it more than technology? Yeah, more than technology. All right. It's on 12 nanometers, right? Yeah. No, no, no. Yeah, 12 nanometers. So it's a new design with TSMC. Very power efficient. Yes, yes. And good price. Yes. High performance, good price, right? Yes, yes. All right. Do you have more demos? Yes, we have a... Here. In our PCC, we also provide some... You can enable the Palm Express while it's charging. While it's charging, yeah. Yes. So, does it charge fast? Yeah, it charges fast. Yeah. How fast is it while it's charging? Is it the fastest while it's charging the world? Yeah, it's fast while it's charging. In the receiver side, it can charge it with 10 watt. 10 watt? Yeah. 10 watt like... around two times to the conventional... Two times? Yeah, two times. Yeah, it's fast. So, right here, we're looking at... and this is the transmitter power. It's a 13 watt right now. But it's 1.38 amps? You don't need to care about it. Consider about it? Yeah, you just need to check the... So, if it's a 4,000 mAh battery, how long it will take to charge? In general, it depends on the phone. But in general, it's around two and a half hour to three hours. With the whole charge, yeah? Yeah, yeah, yeah. It charges to be similar to wire, almost similar. Similar to USB Type-C? Yeah, yeah, USB Type-C, yeah. Oh, okay. Cool. Because in wire charging, wire charging also has a fast charging and a normal charging. So, the wireless charging speed is quite similar to the wire. And so, in the phone, there's a big coil? Or where's the coil to do the wireless? In the center. Center? Yeah, center. What's the cost to add wireless charger in the phone? You mean the cost? The minimum factory price for wireless charging. Oh, it's... Is it five dollars or ten dollars? No, no, just less than two or three dollars. Two or three dollars? Yeah. So, every phone should have it, right? Yeah. Why do some phones not have it? I don't know. Because they choose to have a better... Yeah, yeah. I think every phone should have the wireless charger. So, this is the fastest chi? Yeah, fastest charger. And so, to do this, you have a receiver and transmitter? Yeah, this is a receiver. In the phone? In the phone. And this is a transmitter. It was at least 3181A transmitter. And how much is hardware cost to make? Less than five. Five. So, everybody should use this, right? Yeah, I think so. Okay. No need the cable anymore? Yeah, no need. Can you also do data transfer or you don't do that? No, we don't do that. Just power. Just power. Okay. Cool. Thank you. Here are another key feature about our P60. So, here are our another key feature. In our P60, we have very good thermal control. So, maybe you can introduce the thermal control. So, is it running right now in here? Yes. And you can see that the competitor A is completely burning up. I mean, it's very hot. This one is also quite hot. But there you find, you're blue. Yeah. Right? Yeah. They are the same, right? Yeah, almost the same. And the key factor of this thermal control is, you can see the thermal in this competitor phone is just focusing on this area. But in our phone, you can see the thermal is being spread in average area. So, does that have to do with how you manage the CPU? Yeah. The core pilot is your multi-core system, right? Yeah, yeah. We will have some technology to detect how the application will need the computing power. So, when we know how they want to consume the power, we will provide just in time the power consumption today and the performance. Is this part of the AI or no? No. Right now, it's no switcher. Yeah, yeah. The multi-core switcher technology. Yeah. And the A73 core is a very optimal high-performance core. It's better than the A72. In the X30, it was the A72 dual core. Now, quad core A73, very optimized, right? Yes. We should say the normator is A72 or A73. This kind of technology can apply in different arm architecture. So, we developed this kind of technology and helped the CPU to run in and provide a power consumption performance to the application just in time. Do you need something with the sensor? Yes, yes. We can show you another one. It's a little bit busy right now. Maybe we can go another. So, this is the core pilot. Core pilot, yes. So, what is it showing right here? This is the current. Yeah. P60. And the current for nonmeter. And this is the power gain compared to P60. So, you're comparing 10 nanometer chipset to 12 nanometer to 40 nanometer? Yes. And there's a 70% reduction compared to 40 nanometer solution. Yes. So, we're better than... But a 10 nanometer is even... Can it be less? Yes. Yeah. The... That's big, yeah. It's 14 to 12, yeah. All right. So, this is to explain where the 12 nanometer is placed in terms of power consumption. Yes. To run these kind of games. Yes. So, of course, 10 nanometers could be better, but is this one higher? Yes, even higher. Why? Because the GPU is better than yours? There are several, several conditions, right? Okay. I can maybe explain. Because you can see this 14 nanometer is 12. The nanometer is 10 nanometer. So, from the theoretical, the power consumption, the 12 nanometers should be in the middle. Yeah. But you can see the difference between this application. We have better... We have more... You will... The power consumption is almost approaching the 10 nanometer. Okay. This is the application. So, our compiler technology can help some application running on this chipset with very good power performance. But sometimes some application may not gain so much gain, but you can see, at least, it's almost similar, but better than the average. Nice. So, new benchmarks for SoC power consumption. Yeah. CPI architecture of the VDTEC. Oh, this one here is the one. Energy away scheduler plus. EAS plus. Yes. Thermal management UX military. All right. And there was a sensor right over here. What is this about? Yeah. Let's introduce another solution. This is our biosensing solution. Nice. So, what kind of sensor do you have here? Does it take a block pressure military? Yeah. We have six bioscience like here. We can output heart rate, blood pressure, oxygen level in your blood, and the fatigue also present in this. Also, the ECG in the picture from one minute. How do you do the blood pressure? Oh, okay. We test a big block around the arm. No, no. We test the blood flowing speed from heart to the finger. One finger. So, do you use something like this? Yeah, you can take it this way. How does it go to the finger? I don't know. You need something else? No, we figure out information from the thumb. This way. You need another sensor or this is enough to do blood pressure? Blood pressure, yeah. This does blood pressure? Yeah, you can test it by yourself. Do you want to check? All right. Oh, yeah. Okay, you try? Yes. Let me connect to the Brutus first. Okay. So, is this going to be available? Yeah. Is this going to be 56381? It's already available. For mass production. So, you can see ECG, right? And the ECG. The first coming is heart rate. 71. Then we're coming out with the oxygen level. 98, 99. And then we're coming out, how do you say the blood pressure? We measure the pulse with transition time from heart to the fingertips. So, give you the blood pressure. And then finally we're coming out with the exhaustion level and the pressure level. Like breathing? No, not breathing. It means you are tired or nervous. Something, emotional index. So, like this. So, my blood pressure is normal. And we upload your data to the cloud. It might give you an instant report here. So, my heart rate is in the normal region. The same for the SPO2 in the normal region. What's the SPO2? The oxygen level in your blood. And here. In your blood. Normally it's above 95. If you below 90 or feel dizzy. Or when you are climbing into the high mountain. Oxygen level. Oxygen. Oxygen. Oxygen. Oxygen. Oxygen. Oxygen. Oxygen. And then there's fatigue. Fatigue, yeah. Exhaustion level. If you are exhausted, the index will go high. And the pressure level. If you feel high pressure, the index will go high. This is a high risk. If you have irregular heartbeat. You have to try it. I will try it. This is great. Interesting. This already on the market? Yeah. How much is the price for something like this? The reference price is 6 dollars. How much? 6 US dollars. 6. 6 US dollars. Just 6. 6 US dollars. That's 60 or 60. 6 US dollars. 6 for this kind of sensor. What is inside here? So, there's an IR. Inval raid. And the photo diode. Photo diode? 3 sensors inside. What's over here? It's just the metal. Connection to the chip. Only this one is doing. And then the MT6381. Is that an arm cortex? No no no, just at the front end. Just a front sensor. How long? How do you charge this? The tragedy is that it's just a coin battery, kind of the last, it's over one year. So it's on the market, many people use this? There are two kinds of, one is some customer they put in to embed the phone, so you see like the major 90s way. Nice, you haven't embedded it in the fingerprint reader? Yeah, it's just a fingerprint reader, you see, the major process. So some customers they put this on the phone, some use the accessory. Because people touch their phones all the time, that's great. So can I see this one again? Where does it go? It works like this way. Major 90s way. So when you touch there? Fingertips, like this way. You put the other one there? Yeah, because we have two close-ups. 60 seconds? Yeah, give you 6 miles. Is this FDA approved? It depends on the customers. Is it accurate? It's very good. How does it compare with taking what pressure on the arm? Let's see. Our heart rate is CO2 and ACG can pass the FDA. Our pressure is close to the FDA. Close? Close to the FDA. So it's not very far from the accuracy of the FDA? Very close, very close. Because the arm is annoying? It's annoying and it's a bigger volume. Bigger volume? This one is just one like this. Yeah, you should try it. So is it already mass production? Yeah, mass production. So the first cell phone was coming out on April and it's coming out on May this year. Cool, that's awesome.