WEBVTT 00:00:00.008 --> 00:00:01.071 Hey everybody. 00:00:05.067 --> 00:00:08.765 >> Have you ever noticed how hard it is, to locate a sound, 00:00:08.765 --> 00:00:11.138 in an, coming from an app or a game? 00:00:11.138 --> 00:00:13.157 It might even sound flat. 00:00:14.955 --> 00:00:16.610 My name's Steven Wilsons. 00:00:16.610 --> 00:00:18.070 >> I'm Adam Tuliper. 00:00:18.070 --> 00:00:18.870 >> And during this session, 00:00:18.870 --> 00:00:23.250 we'll show you how you can use Spatial Audio to accurately put 00:00:23.250 --> 00:00:26.540 a sound in a three dimensional space. 00:00:26.540 --> 00:00:30.340 And you'll be able to do that on HoloLens, Xbox 1, 00:00:30.340 --> 00:00:32.270 Windows Desktop and Windows phone. 00:00:34.360 --> 00:00:36.380 So what is Spatial Audio? 00:00:36.380 --> 00:00:38.120 Traditionally, audio has been flat. 00:00:38.120 --> 00:00:42.886 Whether there was headphones, 5.1 speaker configuration, 00:00:42.886 --> 00:00:45.870 7.1 speaker configurations. 00:00:45.870 --> 00:00:49.240 But now, Spatial Audio gives you the illusion that 00:00:49.240 --> 00:00:52.200 a sound is coming from a particular location 00:00:52.200 --> 00:00:53.320 in a three dimensional space. 00:00:54.750 --> 00:00:58.787 If you hear a sound coming from an application as if you would 00:00:58.787 --> 00:01:02.680 from real life, that really is Spatial Audio done well. 00:01:05.293 --> 00:01:08.800 At BUILD we're announcing Project Penrose and 00:01:08.800 --> 00:01:13.098 with Penrose we've exposed Microsoft's HoloLens audio 00:01:13.098 --> 00:01:16.950 technologies to your UWP apps and games. 00:01:16.950 --> 00:01:19.280 Again you can do this on HoloLens, desktop, phone, and 00:01:19.280 --> 00:01:19.880 Xbox One. 00:01:21.160 --> 00:01:23.262 This technology uses HRTFs and 00:01:23.262 --> 00:01:26.900 as such is optimized to work well with headphones. 00:01:28.270 --> 00:01:31.463 As a developer you'll see how you can now enable spatial audio 00:01:31.463 --> 00:01:33.440 through the AudioGraph API or Unity. 00:01:37.440 --> 00:01:38.449 As I mentioned before, 00:01:38.449 --> 00:01:42.340 to experience Spatial Audio with Penrose you'll need headphones. 00:01:42.340 --> 00:01:45.150 The speakers you have in this room will only give you a hint 00:01:45.150 --> 00:01:47.920 at what Spatial Audio can really do for you. 00:01:47.920 --> 00:01:49.450 You can try the HoloLens Academy. 00:01:49.450 --> 00:01:51.760 Many of you have tried before. 00:01:51.760 --> 00:01:54.810 You can download the code, we'll show you for the demos and 00:01:54.810 --> 00:01:57.600 you can also download the games afterwards. 00:01:57.600 --> 00:01:58.290 You can also come and 00:01:58.290 --> 00:02:02.850 find us to hear some of these demos in person. 00:02:04.370 --> 00:02:05.580 So what are HRTFs? 00:02:06.580 --> 00:02:09.955 HRTF stands for head-related transfer function and 00:02:09.955 --> 00:02:14.170 with our two ears we can hear sound in three dimensions. 00:02:14.170 --> 00:02:16.540 One is direction, so I can hear where a sound's coming from, 00:02:16.540 --> 00:02:17.540 the left or the right of me. 00:02:17.540 --> 00:02:21.050 I can hear whether the sound is coming from in front of me or 00:02:21.050 --> 00:02:22.010 back of me. 00:02:23.180 --> 00:02:25.560 I can distinguish elevation and 00:02:25.560 --> 00:02:28.120 I can also estimate the distance from where a sound is coming. 00:02:29.860 --> 00:02:33.849 Your head shape, your external ears, your internal ears and 00:02:33.849 --> 00:02:37.110 your brain will work together so 00:02:37.110 --> 00:02:38.880 that you can perceive location accurately. 00:02:40.342 --> 00:02:44.710 An HRTF will describe how a sound would reach your ears, 00:02:44.710 --> 00:02:45.810 your left and 00:02:45.810 --> 00:02:49.140 your right ears when it's coming from a location in space. 00:02:49.140 --> 00:02:51.810 So when you actually record a sound at a certain location, 00:02:52.970 --> 00:02:55.590 when you move away from that, an HRTF will 00:02:55.590 --> 00:02:58.920 describe the transfer that happens to the sound. 00:02:58.920 --> 00:03:01.333 And you will get the illusion that the sound was actually 00:03:01.333 --> 00:03:02.935 coming from where I just pointed at. 00:03:05.328 --> 00:03:08.029 Head Orientation Tracking is really important 00:03:08.029 --> 00:03:11.240 to improve accuracy of sound localization. 00:03:11.240 --> 00:03:14.670 And as you can move your head, you'll notice that it greatly 00:03:14.670 --> 00:03:18.655 reduces front-back confusion and improves elevation perception. 00:03:18.655 --> 00:03:22.274 We'll show you a demo in a couple of minutes of how we can 00:03:22.274 --> 00:03:26.053 use one of the headphones that are available in the market 00:03:26.053 --> 00:03:29.361 today to implement Head Orientation Tracking and 00:03:29.361 --> 00:03:32.376 just greatly improve that accuracy for you. 00:03:32.376 --> 00:03:35.390 In Windows 10 we introduced the AudioGraph API. 00:03:36.720 --> 00:03:41.600 So the AudioGraph API has a bunch of input nodes on one end. 00:03:42.870 --> 00:03:45.580 A device input node where you can connect a device 00:03:45.580 --> 00:03:48.810 like a microphone to a file input node. 00:03:48.810 --> 00:03:52.100 You can load a WAV file or an MP3 file with on an AudioFrame 00:03:52.100 --> 00:03:54.980 input node if you were already processing audio or 00:03:54.980 --> 00:03:57.430 video in the Windows stack. 00:03:57.430 --> 00:04:02.510 You can apply a fax to them or you can submix them. 00:04:02.510 --> 00:04:05.270 And once you're done submixing them or applying affects to them 00:04:05.270 --> 00:04:10.090 you can output them to a device, a speaker for example, a file. 00:04:10.090 --> 00:04:12.505 Or you can continue processing them in the Windows stack using 00:04:12.505 --> 00:04:13.674 the AudioFrame output node. 00:04:16.211 --> 00:04:19.671 Now in Windows 10 Redstone, we are adding Emitters and 00:04:19.671 --> 00:04:20.470 Listeners. 00:04:21.620 --> 00:04:22.550 What is an emitter? 00:04:22.550 --> 00:04:25.160 An emitter really is the source of Spatial Audio. 00:04:25.160 --> 00:04:26.550 It's where the sound is coming from. 00:04:27.640 --> 00:04:30.429 An emitter can either be an input source or a submix. 00:04:31.920 --> 00:04:34.480 The number of emitters that we support really depend on 00:04:34.480 --> 00:04:37.050 a number of factors in your application. 00:04:37.050 --> 00:04:39.640 One is, how much CPU do you wanna sacrifice? 00:04:39.640 --> 00:04:42.070 How much CPU does your device have or 00:04:42.070 --> 00:04:44.330 computing power does your device have? 00:04:44.330 --> 00:04:46.910 And does your device support hardware offload or not. 00:04:48.010 --> 00:04:49.690 An emitter can have an orientation. 00:04:50.690 --> 00:04:52.240 It can have a position in space and 00:04:52.240 --> 00:04:54.970 it can actually travel at a velocity. 00:04:54.970 --> 00:04:58.040 On the other end of sound there's the listener and 00:04:58.040 --> 00:05:00.830 that's where we perceive sound. 00:05:02.820 --> 00:05:05.810 We only support 1 listener at the current time per 00:05:05.810 --> 00:05:09.380 AudioGraph and a listener itself can have an orientation. 00:05:10.580 --> 00:05:12.290 It can have a position in space as well and 00:05:12.290 --> 00:05:13.670 it can also travel at a velocity. 00:05:15.100 --> 00:05:16.730 So let's have a look at that original graph. 00:05:17.880 --> 00:05:20.030 So now you'll see all of our input nodes can now 00:05:20.030 --> 00:05:23.190 have an emitter and a submix can have an emitter as well. 00:05:24.590 --> 00:05:27.223 And on the right hand side you'll see that our 00:05:27.223 --> 00:05:30.898 DeviceOutputNode now has a default AudioNodeListener object 00:05:30.898 --> 00:05:33.879 for which now you can use to configure where sound is 00:05:33.879 --> 00:05:34.653 perceived. 00:05:39.023 --> 00:05:43.752 All right, let's have a look at a first Hello World demo, 00:05:43.752 --> 00:05:46.561 we called it Hello Spatial Audio. 00:05:50.390 --> 00:05:51.780 Before we do that, sorry. 00:05:53.040 --> 00:05:54.946 Before we actually show you the demo itself, 00:05:54.946 --> 00:05:56.410 let me show you the graph we built. 00:05:58.670 --> 00:06:00.550 The graph is very simple here. 00:06:00.550 --> 00:06:05.398 We're loading one file input node with the file the siren 00:06:05.398 --> 00:06:07.130 of a car. 00:06:07.130 --> 00:06:10.330 We're positioning it in space by adding an emitter to it. 00:06:10.330 --> 00:06:13.000 And we're sending that spatialized sound that you'll 00:06:13.000 --> 00:06:16.180 get, we're sending that directly to our DeviceOutputNode. 00:06:16.180 --> 00:06:18.900 The reason we're doing this is that it'll support hardware 00:06:18.900 --> 00:06:19.750 offload in the future. 00:06:20.780 --> 00:06:24.210 We're also adding a reverb path with reduced gain. 00:06:25.230 --> 00:06:27.790 And we're sending that through the SubmixNode where we actually 00:06:27.790 --> 00:06:29.590 are applying the reverb and 00:06:29.590 --> 00:06:31.340 as well sending that to the device output node. 00:06:33.350 --> 00:06:35.240 The code to do this is very simple. 00:06:36.960 --> 00:06:38.800 First off, we're creating an AudioGraph. 00:06:39.880 --> 00:06:43.212 We're setting the audio category which is GameEffects so 00:06:43.212 --> 00:06:46.650 that the proper docking policies are put in place. 00:06:46.650 --> 00:06:49.720 We're also making sure that we configure our output correctly, 00:06:49.720 --> 00:06:52.740 setting it to 48 kilohertz, setting it to stereo for 00:06:52.740 --> 00:06:56.250 our headphones and 16-bit flow. 00:06:56.250 --> 00:06:58.280 And then we're creating our graph. 00:06:58.280 --> 00:07:00.730 And if our graph creation is successful, we'll continue. 00:07:02.390 --> 00:07:06.230 Now our graph is really built from the output 00:07:06.230 --> 00:07:08.060 nodes to the input nodes. 00:07:08.060 --> 00:07:09.630 So the first thing we'll do is we'll create our 00:07:09.630 --> 00:07:10.344 device output node. 00:07:10.344 --> 00:07:16.194 We'll create our device output node saying on our AudioGraph 00:07:16.194 --> 00:07:19.466 CreateDeviceOutputNodeAsync. 00:07:19.466 --> 00:07:22.054 And once that's successful, our next step happens, 00:07:22.054 --> 00:07:24.300 and we keep building our graph. 00:07:24.300 --> 00:07:26.319 We're adding the reverb submix node. 00:07:27.690 --> 00:07:29.160 We're creating a submix node and 00:07:29.160 --> 00:07:30.750 we're adding it to our output node. 00:07:31.900 --> 00:07:33.690 We are also reducing the gain for that. 00:07:35.780 --> 00:07:39.070 And finally for our reverb, we're adding the reverb effect, 00:07:39.070 --> 00:07:41.760 we're creating, it and we're adding it to the submix. 00:07:43.430 --> 00:07:46.920 Now, our next step is to create that emitter. 00:07:46.920 --> 00:07:48.690 That's really where the sound will be coming from. 00:07:50.090 --> 00:07:53.720 We're creating an audio node emitter by giving it 00:07:53.720 --> 00:07:56.020 an omnidirectional emitter shape, 00:07:56.020 --> 00:07:58.350 which means that the sound will go in every direction. 00:07:58.350 --> 00:08:00.920 You can also create a cone, for example, where you will 00:08:00.920 --> 00:08:03.710 only hear the sound as long as you're in the projected cone. 00:08:05.710 --> 00:08:09.824 And we're positioning our emitter and we're using a right 00:08:09.824 --> 00:08:13.607 hand coordinate system for that so negative X is left, 00:08:13.607 --> 00:08:16.910 positive one is right, and then we have Y and Z. 00:08:16.910 --> 00:08:18.760 And Z is surprisingly for 00:08:18.760 --> 00:08:21.799 some people negative in the front and positive in the back. 00:08:23.970 --> 00:08:24.620 Our next step is, 00:08:24.620 --> 00:08:26.470 now that we have our emitter, we're loading our file. 00:08:27.770 --> 00:08:31.820 Currently for spatialization to happen, all of the sound you 00:08:31.820 --> 00:08:36.282 attach with an emitter needs to be mono and encoded in 48 kHz. 00:08:36.282 --> 00:08:42.790 So we're loading in mono 48 kHz car siren, and once 00:08:42.790 --> 00:08:46.770 we've done that, we're adding it, or combining, or sorry. 00:08:46.770 --> 00:08:50.950 We're creating the file input node and 00:08:50.950 --> 00:08:52.330 we're passing the emitter in. 00:08:52.330 --> 00:08:54.459 So at this point, we have our file input node, and 00:08:54.459 --> 00:08:56.750 we positioned it in space by adding our emitter to it. 00:08:58.839 --> 00:09:02.259 Once we've successfully created our emitter, we're adding it to 00:09:02.259 --> 00:09:05.920 our submix, and we're adding it to our output node. 00:09:05.920 --> 00:09:08.298 And at this point our graph's been completed and 00:09:08.298 --> 00:09:09.522 we can show you the demo. 00:09:18.368 --> 00:09:21.018 What's very neat, with this simple demo you'll be able to 00:09:21.018 --> 00:09:23.668 download the code afterwards, is that we have a car here and 00:09:23.668 --> 00:09:25.600 the car is positioned away from you. 00:09:25.600 --> 00:09:26.920 So the sound will be less loud. 00:09:28.100 --> 00:09:29.240 As I drag the car, 00:09:29.240 --> 00:09:32.370 we'll also update the emitter position, we'll also 00:09:32.370 --> 00:09:35.670 add velocity to the object which will create a Doppler effect. 00:09:36.780 --> 00:09:43.525 So let me give it a try. 00:09:43.525 --> 00:09:50.682 [SOUND] Very 00:09:50.682 --> 00:09:54.269 simple demo to add an emitter to your audiograph. 00:09:58.186 --> 00:09:59.464 >> I'm gonna wake up with that sound in my head. 00:09:59.464 --> 00:10:00.787 [LAUGH] >> Yeah. 00:10:00.787 --> 00:10:02.087 Well, we've gone through a couple of dry runs. 00:10:02.087 --> 00:10:04.508 [LAUGH] The next step is, 00:10:04.508 --> 00:10:10.310 how do you had head tracking to our AudioGraph? 00:10:10.310 --> 00:10:10.990 So what we did for 00:10:10.990 --> 00:10:15.170 that is, we're using the Jabra Intelligent Headset, 00:10:15.170 --> 00:10:20.175 which provides us with an accelerometer, e-Compass, 00:10:20.175 --> 00:10:23.810 [INAUDIBLE], GPS, etc. 00:10:23.810 --> 00:10:26.620 And we're using that one to update our listener 00:10:26.620 --> 00:10:28.070 orientation in the next demo. 00:10:29.880 --> 00:10:32.480 So from a graph perspective what we're just doing is we're, 00:10:32.480 --> 00:10:33.740 we will build the same graph, 00:10:34.860 --> 00:10:36.380 the UI will look a little bit different. 00:10:37.390 --> 00:10:40.069 But we'll start updating our audio node listener object. 00:10:44.120 --> 00:10:48.190 The first thing we're doing, is from this head set we'll read 00:10:48.190 --> 00:10:54.110 yaw, pitch, and roll as a vector three coordinate, 00:10:54.110 --> 00:10:55.460 x, y, z coordinates, again. 00:10:56.580 --> 00:10:59.386 Our next step will be to convert the angles we'll get, 00:10:59.386 --> 00:11:01.221 the url angels, to radial angles. 00:11:02.784 --> 00:11:04.290 Former is right there. 00:11:04.290 --> 00:11:05.740 It's just a simple math conversion. 00:11:07.410 --> 00:11:08.700 Then the next step is, 00:11:09.770 --> 00:11:13.832 to make sure that our orientation is set as part of a. 00:11:14.840 --> 00:11:18.280 Is often times used in the game development world 00:11:18.280 --> 00:11:19.540 to avoid gimble lock. 00:11:19.540 --> 00:11:23.470 It's a pretty well known problem there where to axis collide and 00:11:23.470 --> 00:11:26.160 you can't just move one coordinate anymore 00:11:26.160 --> 00:11:30.790 to make sure your object turns, all lright. 00:11:30.790 --> 00:11:33.400 Let's have a quick look at that demo. 00:11:33.400 --> 00:11:35.937 You'll see it's a spacial sphere demo and 00:11:35.937 --> 00:11:38.967 I also did the recorded session I'll go step by step 00:11:38.967 --> 00:11:42.081 through the code to actually make this demo happen. 00:11:53.082 --> 00:11:54.266 So what I'll do in. 00:11:54.266 --> 00:11:55.773 >> Think switch them out. 00:11:55.773 --> 00:11:56.298 >> Yeah, let me switch. 00:11:59.398 --> 00:12:02.298 In this presentation, we'll see sphere and 00:12:02.298 --> 00:12:05.426 if you let the demo run, I'll position a ball and 00:12:05.426 --> 00:12:08.420 a sun around you and you'll hear the sun move. 00:12:08.420 --> 00:12:10.220 It would work a lot better with headphones as I 00:12:10.220 --> 00:12:11.530 mentioned before. 00:12:11.530 --> 00:12:15.320 What will also happen is that as I move my head, 00:12:15.320 --> 00:12:17.820 you'll see my little bobblehead move along with me. 00:12:17.820 --> 00:12:19.094 And update the position of sound. 00:12:26.970 --> 00:12:29.720 >> Front position. 00:12:29.720 --> 00:12:30.890 Front right position. 00:12:32.390 --> 00:12:33.830 Right position. 00:12:33.830 --> 00:12:34.980 >> So as I'm moving my head now. 00:12:34.980 --> 00:12:36.460 >> Right position. 00:12:36.460 --> 00:12:37.920 >> For me the sound is moving with me. 00:12:40.000 --> 00:12:41.440 >> Right position. 00:12:41.440 --> 00:12:42.480 >> I think that's a pretty cool demo- 00:12:42.480 --> 00:12:42.880 >> Right position. 00:12:42.880 --> 00:12:47.160 >> In how to show how you can use head orientation tracking to 00:12:47.160 --> 00:12:51.510 just make sure you have a more realistic experience of audio. 00:12:51.510 --> 00:12:53.785 And if you're just playing a game or something, just those 00:12:53.785 --> 00:12:56.115 small movements will help you locate the sound a lot easier. 00:12:59.098 --> 00:13:02.231 All right now that we've seen some of the simple introductions 00:13:02.231 --> 00:13:03.660 on how to use spatial audio. 00:13:07.308 --> 00:13:10.497 Let's talk about what are some real life applications you 00:13:10.497 --> 00:13:11.090 can have. 00:13:12.840 --> 00:13:14.730 First one is, augmented and virtual reality. 00:13:14.730 --> 00:13:15.810 It's very simple. 00:13:15.810 --> 00:13:16.780 You've all seen it. 00:13:16.780 --> 00:13:20.130 It's a very, it's a must have to create a realistic perception. 00:13:20.130 --> 00:13:23.330 If you're creating great audio, or if you don't create great 00:13:23.330 --> 00:13:25.800 audio in a virtual reality, it is easy for 00:13:25.800 --> 00:13:28.650 your users to really just snap out of that experience and 00:13:28.650 --> 00:13:29.650 it doesn't work well for them. 00:13:30.970 --> 00:13:32.624 Another experience you can create is watching movies. 00:13:32.624 --> 00:13:35.529 So as you have surround sound, what you can do with this 00:13:35.529 --> 00:13:39.077 technology is virtually position the speakers around your head. 00:13:39.077 --> 00:13:42.574 So you can take surround sound and actually listen 00:13:42.574 --> 00:13:46.080 to a surround sound experience with headphones. 00:13:47.150 --> 00:13:48.550 Another experience you can build, for 00:13:48.550 --> 00:13:51.710 example, is a music experience. 00:13:53.040 --> 00:13:55.470 If you use this technology you'll notice that the sound 00:13:55.470 --> 00:13:57.020 actually is more externalized and 00:13:57.020 --> 00:13:59.270 it just creates a more I think, 00:13:59.270 --> 00:14:02.620 it's just a better listening experience for the audio scene. 00:14:03.770 --> 00:14:08.290 What we've also experimented with was just positioning 00:14:08.290 --> 00:14:10.380 all the different instruments in a different location, and 00:14:10.380 --> 00:14:13.440 you get a very interesting perception of sound that way. 00:14:13.440 --> 00:14:15.770 And then, there's gaming. 00:14:15.770 --> 00:14:20.990 Now, how neat is it, if you play a 3D shooter and 00:14:20.990 --> 00:14:23.110 you hear somebody sneak up behind you. 00:14:23.110 --> 00:14:24.470 You can just turn around and shoot them. 00:14:24.470 --> 00:14:26.890 No depth of sound will be coming from behind you. 00:14:26.890 --> 00:14:30.550 That point, you just became a better gamer and 00:14:30.550 --> 00:14:32.490 we just raised the bar for gaming. 00:14:35.330 --> 00:14:38.460 One of the applications of Augmentor Reality does not have 00:14:38.460 --> 00:14:39.209 any visuals with it. 00:14:40.210 --> 00:14:43.110 It's the Guide Dogs project we've been working on 00:14:43.110 --> 00:14:46.700 and with the Guide Dogs project we're helping people 00:14:46.700 --> 00:14:49.850 that are visually impaired to navigate to the world more 00:14:49.850 --> 00:14:54.470 comfortably with the use of spatial audio. 00:14:54.470 --> 00:14:58.110 So as they're walking around we are tracking their location. 00:14:58.110 --> 00:15:00.350 And we're making sure that as our for 00:15:00.350 --> 00:15:02.350 example walking by a coffee shop, 00:15:02.350 --> 00:15:05.040 we give them an indication that there is a coffee shop. 00:15:05.040 --> 00:15:07.910 But we don't have to say things like there is a coffee 00:15:07.910 --> 00:15:12.078 shop 30 degrees from you at about 15 meters. 00:15:12.078 --> 00:15:15.920 Just by seeing there is a coffee shop and 00:15:15.920 --> 00:15:19.380 positioning the sound then we'll be able to look at where that 00:15:19.380 --> 00:15:23.430 shop is and we just don't interfere as much with their 00:15:23.430 --> 00:15:28.390 daily experience by having all that extra thrown at them. 00:15:29.560 --> 00:15:31.950 Some of the key technologies that are being used here 00:15:31.950 --> 00:15:33.540 are spatial audio. 00:15:33.540 --> 00:15:36.340 But also background audio which has become a lot simpler 00:15:36.340 --> 00:15:39.540 in the new version of Windows 10 and we'll talk about that. 00:15:39.540 --> 00:15:41.967 You've seen the headset that we've been using through 00:15:41.967 --> 00:15:44.559 BlueTooth, and it also uses Microsoft mapping services. 00:15:49.758 --> 00:15:52.700 I mentioned virtual speakers before. 00:15:52.700 --> 00:15:55.440 And we'll build a demo here using Big Buck Bunny, 00:15:55.440 --> 00:15:57.830 which is my favorite video of all time. 00:15:59.000 --> 00:16:01.610 This one's a little bit more complicated, this graph. 00:16:01.610 --> 00:16:04.730 So what we're doing is we're taking a video file 00:16:04.730 --> 00:16:06.910 with 5.1 speaker audio. 00:16:06.910 --> 00:16:08.550 We're passing that to a media clip, 00:16:09.860 --> 00:16:13.580 which is a media composition class. 00:16:13.580 --> 00:16:17.080 And to that clip we're adding the spacial audio effect. 00:16:17.080 --> 00:16:20.240 So what we'll be doing here is that we'll get an audio frame 00:16:20.240 --> 00:16:23.360 into spacial audio effect and we'll take the six channels, 00:16:23.360 --> 00:16:27.730 we'll split them up and we'll position them in different audio 00:16:27.730 --> 00:16:29.860 frame input nodes and we'll put them around my head. 00:16:32.190 --> 00:16:34.640 That's pretty much what the graph will look like. 00:16:34.640 --> 00:16:38.010 It's a graph here, you're fairly familiar with at this point. 00:16:38.010 --> 00:16:40.840 So as I mentioned you have the audio frame input node, you add 00:16:40.840 --> 00:16:43.930 an emitter to it and you send it to an audio frame output node 00:16:43.930 --> 00:16:48.280 instead of a device this time so that you can loop the result, 00:16:48.280 --> 00:16:53.300 the spatialized stereo version of the sound back to the effect, 00:16:53.300 --> 00:16:56.440 replace the original audio, the 5.1 audio, 00:16:56.440 --> 00:16:59.230 send it back to the media clip to the media element. 00:16:59.230 --> 00:17:00.660 And here, to your speakers. 00:17:01.800 --> 00:17:03.490 So how cool is that? 00:17:03.490 --> 00:17:04.580 All the sudden, 00:17:04.580 --> 00:17:07.490 I got a 501 surround system just using my headphones. 00:17:09.400 --> 00:17:12.040 So the code here is fairly simple. 00:17:12.040 --> 00:17:14.870 But and it's creating the media clip first. 00:17:14.870 --> 00:17:15.880 I'm loading my file. 00:17:17.080 --> 00:17:19.710 And then, through the audio effect definitions, 00:17:19.710 --> 00:17:21.219 I am adding my spatial audio effect. 00:17:22.300 --> 00:17:24.330 And the meat of the code is obviously in the space, 00:17:24.330 --> 00:17:25.170 how do you effect? 00:17:25.170 --> 00:17:28.320 I'll share the code out, after this presentation. 00:17:29.950 --> 00:17:32.880 Our next step is to create a media composition object, and 00:17:32.880 --> 00:17:34.400 add a clip to it, and 00:17:34.400 --> 00:17:36.160 then just set it as a source for the player. 00:17:38.270 --> 00:17:39.087 So, have a look at that demo. 00:17:53.970 --> 00:17:57.242 So, as this video plays, what you'll notice is that the sound 00:17:57.242 --> 00:18:00.457 as sounds playing throughout the video, different speakers 00:18:00.457 --> 00:18:03.393 here in this will light up as they are receiving volume. 00:18:03.393 --> 00:18:13.393 [MUSIC] 00:18:23.034 --> 00:18:26.624 So you've now created a much richer listening experience for 00:18:26.624 --> 00:18:28.570 people with headphones. 00:18:28.570 --> 00:18:29.880 Very simple. 00:18:29.880 --> 00:18:33.555 One of the things you notice is that as soon as I minimized my 00:18:33.555 --> 00:18:37.080 application, it stopped playing which is not great for 00:18:37.080 --> 00:18:41.660 video but it definitely is not great for music scenarios. 00:18:41.660 --> 00:18:44.190 So what we're announcing today as well is a change 00:18:44.190 --> 00:18:46.450 in how background audio works and 00:18:46.450 --> 00:18:49.830 background audio works great with AudioGraph as well. 00:18:49.830 --> 00:18:53.640 So what you do now is you add the background media playback 00:18:53.640 --> 00:18:54.360 capability. 00:18:55.420 --> 00:18:58.590 And with that new capability you can now write all of your chord 00:18:58.590 --> 00:19:01.200 in one process versus the two process model you had to do 00:19:01.200 --> 00:19:04.720 before where In the one process you had your why and 00:19:04.720 --> 00:19:07.890 in the second one you controlled all of your playback which 00:19:07.890 --> 00:19:08.930 was very confusing. 00:19:11.390 --> 00:19:14.290 Now with the one process model when your application is put to 00:19:14.290 --> 00:19:17.310 the background the XAML textures are unloaded and 00:19:17.310 --> 00:19:19.920 the memory footprint of your app comes down. 00:19:19.920 --> 00:19:21.440 The next thing that happens is that you get 00:19:21.440 --> 00:19:23.880 a notifications saying you're minimized. 00:19:23.880 --> 00:19:26.620 And you can clean up more resources if needed. 00:19:26.620 --> 00:19:29.390 And if needed means if you're over the memory limit. 00:19:29.390 --> 00:19:31.810 If you are actually over the memory limit you'll get another 00:19:31.810 --> 00:19:34.840 event and you will need to clean up your resources or 00:19:34.840 --> 00:19:37.830 you will be potentially closed as an application. 00:19:38.850 --> 00:19:42.680 When you're foregrounded again and your visibility has changed. 00:19:42.680 --> 00:19:44.920 You get a notification for that as well. 00:19:44.920 --> 00:19:46.508 And XAML's rehydrated. 00:19:46.508 --> 00:19:50.215 So it's a very simple with a couple notifications how you now 00:19:50.215 --> 00:19:53.718 implement a background audio in Windows 10 Redstone. 00:19:57.955 --> 00:20:00.462 >> All right, that was pretty awesome. 00:20:00.462 --> 00:20:03.215 You can take C# and work with the AudioGraph API. 00:20:03.215 --> 00:20:06.179 And the first time I heard that it was neat because you're 00:20:06.179 --> 00:20:09.331 to really stress that point you're taking all this multiple 00:20:09.331 --> 00:20:12.420 channels and making my regular headphones here just like it 00:20:12.420 --> 00:20:13.860 was almost on fire but one. 00:20:13.860 --> 00:20:17.478 All right, so that's one option that Steven showed to be able to 00:20:17.478 --> 00:20:19.220 work with the AudioGraph API. 00:20:19.220 --> 00:20:22.620 One of my favorite programs of all time outside of 00:20:22.620 --> 00:20:23.710 Visual Studio is Unity, 00:20:23.710 --> 00:20:26.020 a very popular gaming engine middleware. 00:20:26.020 --> 00:20:28.710 And to be able to use our spatial audio API behind 00:20:28.710 --> 00:20:31.150 the scenes with Unity is actually pretty easy. 00:20:31.150 --> 00:20:33.586 So we worked to create a couple neat demos to show you that is 00:20:33.586 --> 00:20:36.121 gonna showcase I think some of the audio features and how to do 00:20:36.121 --> 00:20:38.676 it inside of Unity and some of the cool things that you can do. 00:20:44.005 --> 00:20:47.760 Now audio in Unity has a couple different facets to it. 00:20:47.760 --> 00:20:49.670 We have 2D audio and traditionally think of when 00:20:49.670 --> 00:20:51.630 you're playing a game, a 2D game. 00:20:51.630 --> 00:20:53.550 That's the easiest use case to look at. 00:20:53.550 --> 00:20:56.290 The background music to your game is a pretty consistent 00:20:56.290 --> 00:20:58.410 volume while you're playing your game. 00:20:58.410 --> 00:20:59.880 Doesn't matter where you go in your game, 00:20:59.880 --> 00:21:01.040 it plays pretty consistently. 00:21:01.040 --> 00:21:02.250 It might fade out, might fade in, but 00:21:02.250 --> 00:21:04.130 it's not positionally sensitive. 00:21:04.130 --> 00:21:05.910 The same thing can happen in a 3D game as well. 00:21:05.910 --> 00:21:07.310 You can have 2D audio and a 3D game. 00:21:07.310 --> 00:21:09.530 You have a first person shooter. 00:21:09.530 --> 00:21:11.560 When you're running around shooting zombies you've got 00:21:11.560 --> 00:21:13.600 a heavy metal truck thing in the background. 00:21:13.600 --> 00:21:15.580 Also 2D audio in a 3D game. 00:21:17.610 --> 00:21:19.236 Now next you have 3D audio. 00:21:19.236 --> 00:21:22.670 3D audio as it stands out of the box is really, it's a mix 00:21:22.670 --> 00:21:26.420 between volume control and panning from left to right. 00:21:26.420 --> 00:21:29.320 Based on the position of the listener to the various 00:21:29.320 --> 00:21:30.730 audio sources so 00:21:30.730 --> 00:21:34.270 the closer you get to a sound the louder it's gonna be. 00:21:34.270 --> 00:21:35.810 If you hear a gunshot off to the left hand side, 00:21:35.810 --> 00:21:37.650 you're gonna hear it out of the left hand speaker. 00:21:37.650 --> 00:21:39.930 If you rotate essentially the audio listener you can now 00:21:39.930 --> 00:21:42.110 hear it out of both speakers pretty evenly. 00:21:42.110 --> 00:21:43.750 And you get closer, it's gonna get louder and 00:21:43.750 --> 00:21:44.875 louder and louder. 00:21:44.875 --> 00:21:46.490 Until you reach a distance where it won't get any louder. 00:21:48.260 --> 00:21:50.822 Then we have spatialization. 00:21:50.822 --> 00:21:52.447 You might think that 3D audio spatialization, 00:21:52.447 --> 00:21:53.485 there is a difference though. 00:21:53.485 --> 00:21:57.200 Again, 3D audio is still pretty much flat and plainer. 00:21:57.200 --> 00:21:59.730 Spatialization as Steven showed in some of the demos, 00:21:59.730 --> 00:22:00.940 you can hear things behind you, 00:22:00.940 --> 00:22:02.780 the police car when it goes behind you. 00:22:02.780 --> 00:22:05.650 So it's not just panning or volume control. 00:22:05.650 --> 00:22:08.420 Out of the box, Unity doesn't have spatialization with it. 00:22:08.420 --> 00:22:09.090 But however, 00:22:09.090 --> 00:22:12.350 it does have the ability to have a spatialization plugin. 00:22:14.530 --> 00:22:18.350 So we have an MS HRTF spatialization plugin for Unity. 00:22:18.350 --> 00:22:20.540 You can also add it to the project, or 00:22:20.540 --> 00:22:23.190 as was just announced I believe yesterday with Unity 5.4 00:22:23.190 --> 00:22:27.800 that has Hollow Lens in preview that has it out of the box. 00:22:27.800 --> 00:22:29.330 And future versions of unity will that have 00:22:29.330 --> 00:22:30.010 plug in there out of the box. 00:22:30.010 --> 00:22:31.840 You just need to make sure you go in and 00:22:31.840 --> 00:22:34.190 enable spatialization in your project. 00:22:34.190 --> 00:22:36.720 And you can do that in project settings audio. 00:22:36.720 --> 00:22:39.370 Select a particular spatializer that you want to use. 00:22:39.370 --> 00:22:41.910 And then you're all set to go from the plugin stand point. 00:22:41.910 --> 00:22:44.790 Now you still need to configure a couple of other things. 00:22:44.790 --> 00:22:47.840 But now what are we doing with this spatialization? 00:22:47.840 --> 00:22:51.940 Inside Microsoft, we've actually tested over 350 people. 00:22:51.940 --> 00:22:55.530 We measured the size of their ears, size of their head and 00:22:55.530 --> 00:22:58.170 measured actually 96 different biometric 00:22:58.170 --> 00:23:01.180 pieces of data about each one of those individuals. 00:23:01.180 --> 00:23:04.040 And we've come up with a pretty standardized head model 00:23:04.040 --> 00:23:07.080 that fits a good portion of the population out there. 00:23:07.080 --> 00:23:09.430 We can essentially predict on top of that, 00:23:09.430 --> 00:23:12.320 how you're going to hear audio by the way you look. 00:23:12.320 --> 00:23:15.060 So personalization of these HRTF, head related transfer 00:23:15.060 --> 00:23:17.760 functions, is actually being used in HoloLens currently. 00:23:17.760 --> 00:23:20.640 And we're gonna be bringing this to future versions of Windows 00:23:20.640 --> 00:23:23.220 as well, to be able to use this personalization aspect of HRTF 00:23:23.220 --> 00:23:25.089 in your other applications. 00:23:27.820 --> 00:23:29.840 Now spatialization options, this is important. 00:23:29.840 --> 00:23:30.960 We've already configured a plug-in, 00:23:30.960 --> 00:23:32.460 we have a next step that we need to do. 00:23:33.595 --> 00:23:37.740 We can do it manually, and doing it manually you can spatialize, 00:23:37.740 --> 00:23:40.160 we've got it going to an audio source. 00:23:40.160 --> 00:23:42.590 In a Unity project we typically drop audio sources all over. 00:23:42.590 --> 00:23:44.506 If there's a waterfall, there's an audio source over there. 00:23:44.506 --> 00:23:45.356 There's a zombie, 00:23:45.356 --> 00:23:47.610 has an audio source over there as we'll see shortly. 00:23:47.610 --> 00:23:48.730 We have a pretty cool zombie demo. 00:23:48.730 --> 00:23:52.070 And you need to make sure that you check off spatialized. 00:23:52.070 --> 00:23:54.270 The important thing to note is that if the plugin is not 00:23:54.270 --> 00:23:55.940 enabled, spatialized won't show up. 00:23:57.640 --> 00:23:58.730 Then you choose a spatial blend. 00:24:00.280 --> 00:24:01.880 It's a little slider to determine whether or 00:24:01.880 --> 00:24:02.920 not you want 2D or 3D. 00:24:02.920 --> 00:24:04.560 And there's different use cases for using that. 00:24:04.560 --> 00:24:06.420 Sometimes if you get close to your sound, 00:24:06.420 --> 00:24:08.280 some folks will make it over to 3D side. 00:24:08.280 --> 00:24:11.330 But for spatialization, you want that all the way over to 3D. 00:24:11.330 --> 00:24:13.920 You want a 3D sound spatialized. 00:24:13.920 --> 00:24:16.940 Not to manually configure it and you can do this on any of 00:24:16.940 --> 00:24:19.330 the audio sources in your seam that you want to be spatialized. 00:24:20.480 --> 00:24:22.350 There's another method that you can use to auto configure 00:24:22.350 --> 00:24:23.650 all of this. 00:24:23.650 --> 00:24:25.830 It requires a little bit of code to enumerate your game objects 00:24:25.830 --> 00:24:27.859 to do that and we'll look at that in just a couple slides. 00:24:29.190 --> 00:24:32.590 But first, let's talk about the plugin parameters and 00:24:32.590 --> 00:24:34.500 how we can talk to the spatialization plugin. 00:24:36.210 --> 00:24:38.080 Any spatialization plugin can define their own 00:24:38.080 --> 00:24:38.870 custom parameters. 00:24:38.870 --> 00:24:41.690 They can communicate with from inside of Unity. 00:24:41.690 --> 00:24:43.470 The Microsoft HRTF plugin. 00:24:43.470 --> 00:24:46.220 We have two main parameters that you communicate with. 00:24:46.220 --> 00:24:49.140 The first is a room size, or room model. 00:24:49.140 --> 00:24:52.020 If you're in an office, there's a little bit 00:24:52.020 --> 00:24:54.100 of reverb that's added to the sound in there. 00:24:54.100 --> 00:24:55.060 And your ear senses that. 00:24:55.060 --> 00:24:57.210 Now this is not reverb added for effect. 00:24:57.210 --> 00:24:58.030 If I'm playing guitar, 00:24:58.030 --> 00:25:00.710 I might add a reverb effect on there, that is an effect. 00:25:00.710 --> 00:25:03.220 Sound that you hear in the real world, has a little bit of 00:25:03.220 --> 00:25:06.550 reverb, depending on the room, added to how you hear audio. 00:25:06.550 --> 00:25:11.380 And so our HRTF functions try to model that to make it so, 00:25:11.380 --> 00:25:13.800 you can hear the sound in a realistic way. 00:25:13.800 --> 00:25:15.390 So our HRTF code will say okay, 00:25:15.390 --> 00:25:17.700 you've specified a small room model. 00:25:17.700 --> 00:25:18.870 Then we're gonna add and 00:25:18.870 --> 00:25:22.310 integrate, a little bit of reverb onto the HRTF this way, 00:25:22.310 --> 00:25:23.620 so it sounds realistic. 00:25:23.620 --> 00:25:26.300 You have the ability to specify a medium room. 00:25:26.300 --> 00:25:28.050 In here, I would specify a large room. 00:25:28.050 --> 00:25:30.380 Or number four, as you're basically going outdoors, 00:25:30.380 --> 00:25:31.550 we're gonna have an extremely minimal, 00:25:31.550 --> 00:25:34.050 or nill amount of reverb added to that. 00:25:35.750 --> 00:25:38.270 Next, we have a roll off which allows us just some fine tuning 00:25:38.270 --> 00:25:39.425 of volume curves. 00:25:39.425 --> 00:25:40.530 It's a algorithmic curve and 00:25:40.530 --> 00:25:42.710 we can essentially say here's your minimum gain and 00:25:42.710 --> 00:25:46.120 maximum gain, how we can effect the decibel depending on 00:25:46.120 --> 00:25:48.800 how close or how far away we get from the audio source. 00:25:48.800 --> 00:25:50.520 Just some extra abilities to tweak those values. 00:25:51.858 --> 00:25:53.700 Now, using Unity it's as simple as this, 00:25:53.700 --> 00:25:56.950 on a particular gain object you add a little bit of code. 00:25:56.950 --> 00:25:58.970 First line is common practice across Unity, 00:25:58.970 --> 00:26:02.540 that's how you get a component, an audio source. 00:26:02.540 --> 00:26:05.700 We get a reference to our audio source, now we simply call, 00:26:05.700 --> 00:26:07.280 set spatial Azure float. 00:26:07.280 --> 00:26:09.700 That communication the plugin, the tele-plugin in 00:26:09.700 --> 00:26:13.310 this particular case, value one is room size, and 00:26:13.310 --> 00:26:15.310 then we're going to pass in room size small. 00:26:15.310 --> 00:26:18.540 That's it, that's how you configure an audio source for 00:26:18.540 --> 00:26:20.680 spacialization, enabling a plug-in, 00:26:20.680 --> 00:26:23.760 and communicating with a plug-in to say for 00:26:23.760 --> 00:26:25.740 this audio source we're gonna use a small room size. 00:26:26.940 --> 00:26:28.930 Now I mentioned we can also do this automatically. 00:26:28.930 --> 00:26:30.650 I showed the manual configuration. 00:26:31.790 --> 00:26:33.210 If we want to auto configure it 00:26:33.210 --> 00:26:35.260 we can use just a little bit of code. 00:26:35.260 --> 00:26:38.780 Let's assume we have a use case that, in a small room 00:26:38.780 --> 00:26:41.480 we have 50 different objects we need to spatialize in there. 00:26:41.480 --> 00:26:43.700 What we can do is we can add a tag. 00:26:43.700 --> 00:26:46.550 A tag is just some text you assign to an object in Unity. 00:26:46.550 --> 00:26:49.610 And we can tag all of our particular objects in that small 00:26:49.610 --> 00:26:52.190 room with a tag "SpatialEmittersSmall". 00:26:52.190 --> 00:26:53.210 It can be any text you want, 00:26:53.210 --> 00:26:55.170 I just chose "SpatialEmittersSmall". 00:26:55.170 --> 00:26:56.080 And so that line of code 00:26:56.080 --> 00:26:57.590 enumerates all of those game objects. 00:26:57.590 --> 00:26:58.240 Gives them back to you. 00:26:59.490 --> 00:27:02.560 Next, you can just write a little bit of code to configure 00:27:02.560 --> 00:27:04.970 the audio sources for each of those game objects. 00:27:06.120 --> 00:27:06.844 And then next, 00:27:06.844 --> 00:27:09.685 you just communicate with those particular plug-ins, for 00:27:09.685 --> 00:27:11.260 the spacialization plug-ins. 00:27:11.260 --> 00:27:13.109 You say for each of those objects, go ahead and 00:27:13.109 --> 00:27:14.209 set the room size to small. 00:27:14.209 --> 00:27:16.954 If you had a medium room size you could have spatiaizer 00:27:16.954 --> 00:27:19.480 emitter medium, spatializer emitter large. 00:27:19.480 --> 00:27:21.364 And so it gives you just one bit of code that auto configures 00:27:21.364 --> 00:27:23.363 everything for you so you don't have to worry if you missed any 00:27:23.363 --> 00:27:24.566 options or didn't send correctly. 00:27:28.409 --> 00:27:32.225 For this next demo we worked with a partner, Dream Tapp, 00:27:32.225 --> 00:27:34.900 to implement a Zombie super rocket. 00:27:34.900 --> 00:27:35.620 I love zombies. 00:27:35.620 --> 00:27:37.110 I love zombie shows. 00:27:37.110 --> 00:27:39.890 So there's a couple considerations when you're 00:27:39.890 --> 00:27:44.160 making a game or an experience with spatialized sound 00:27:44.160 --> 00:27:46.650 you can't just drop any audio anywhere. 00:27:46.650 --> 00:27:49.580 Because there's things that folks that have been doing game 00:27:49.580 --> 00:27:50.310 and app development for 00:27:50.310 --> 00:27:52.210 awhile they might have some expectations. 00:27:52.210 --> 00:27:53.710 Like let's just throw sound in there. 00:27:53.710 --> 00:27:56.470 Position becomes extremely important as we'll see. 00:27:56.470 --> 00:27:57.921 So, let's check out this demo. 00:28:09.820 --> 00:28:12.595 All right, first let's go ahead and play it. 00:28:16.915 --> 00:28:17.762 You'll notice right away. 00:28:21.756 --> 00:28:22.859 There's some zombie sounds. 00:28:26.260 --> 00:28:28.153 Let's actually go ahead and maximize this. 00:28:33.076 --> 00:28:34.130 Okay now listen to the audio. 00:28:35.800 --> 00:28:37.650 We are currently spalialized as you can see. 00:28:37.650 --> 00:28:39.350 Kind on the left hand side of the screen. 00:28:39.350 --> 00:28:41.060 Spatial audio is on. 00:28:41.060 --> 00:28:42.380 We are in a small room size. 00:28:42.380 --> 00:28:47.885 Let's change the room size to medium and large. 00:28:47.885 --> 00:28:51.472 [SOUND] >> [INAUDIBLE] 00:28:51.472 --> 00:28:54.282 >> None. 00:28:54.282 --> 00:28:54.960 Small. 00:28:57.210 --> 00:28:59.685 Medium and large. 00:28:59.685 --> 00:29:01.705 And if we turn it off you can hear the instant effect. 00:29:01.705 --> 00:29:05.662 [SOUND] Right, everything is kind of 00:29:05.662 --> 00:29:09.092 a consistent volume almost like you have headphones on but 00:29:09.092 --> 00:29:10.983 it's all panning the same way. 00:29:10.983 --> 00:29:14.722 Now let's turn around here. 00:29:16.589 --> 00:29:18.912 And we have the scariest zombie of all. 00:29:18.912 --> 00:29:21.760 [SOUND] >> We have a zombie that looks 00:29:21.760 --> 00:29:22.429 frightening like me. 00:29:22.429 --> 00:29:26.663 [SOUND] >> [LAUGH] Let's off that 00:29:26.663 --> 00:29:27.748 zombie. 00:29:27.748 --> 00:29:33.587 >> [LAUGH] >> Now listen in here. 00:29:33.587 --> 00:29:34.342 [MUSIC] 00:29:34.342 --> 00:29:36.084 We have a flame overhead. 00:29:36.084 --> 00:29:46.022 [MUSIC] 00:29:46.022 --> 00:29:49.353 We have some zapping wires, and lastly, 00:29:49.353 --> 00:29:51.930 we've some speakers in there. 00:29:51.930 --> 00:29:54.264 [MUSIC] 00:29:54.264 --> 00:29:57.042 As I turn away. 00:29:57.042 --> 00:29:58.285 We can hear the position. 00:29:58.285 --> 00:29:59.502 The sound is up there. 00:29:59.502 --> 00:30:00.326 It's coming from up there. 00:30:00.326 --> 00:30:01.046 Spatialize. 00:30:01.046 --> 00:30:05.167 [MUSIC] 00:30:05.167 --> 00:30:07.150 So, with these sounds, there's a couple different things. 00:30:07.150 --> 00:30:08.650 We heard foot steps as we were walking along. 00:30:08.650 --> 00:30:09.652 So, let's look at my character, here. 00:30:16.919 --> 00:30:19.923 If I scroll back, let's talk about two things, first, 00:30:19.923 --> 00:30:21.780 that we use to enable this. 00:30:21.780 --> 00:30:25.450 Edit, Project Settings, Audio, and 00:30:25.450 --> 00:30:27.400 here we have MSHRT a spatializer. 00:30:27.400 --> 00:30:28.320 So we've added the plug in. 00:30:29.790 --> 00:30:34.120 Now anywhere, when we have an audio source, 00:30:34.120 --> 00:30:35.770 we've configured our audio sources 00:30:37.750 --> 00:30:42.010 to have our spatialized settings and those were set to 3D. 00:30:42.010 --> 00:30:43.359 So our sources are all set up for us. 00:30:44.880 --> 00:30:46.185 But why do we have so 00:30:46.185 --> 00:30:50.598 many different options here on our first-person controller? 00:30:50.598 --> 00:30:54.331 There's one, two, three, four, five, six, seven, 00:30:54.331 --> 00:30:58.640 eight different audio sources just on my main character. 00:30:58.640 --> 00:31:00.310 Why so many sounds for just one character? 00:31:00.310 --> 00:31:02.280 Well, as we started this, and 00:31:02.280 --> 00:31:03.580 we had something like the footsteps. 00:31:03.580 --> 00:31:07.310 Footsteps, typically an audio source, is about here on a user. 00:31:07.310 --> 00:31:08.959 When you're playing footsteps from your 00:31:08.959 --> 00:31:11.800 mid-section in a spatialized world, it sounds weird. 00:31:11.800 --> 00:31:13.380 When you play in a non-spatialized game, 00:31:13.380 --> 00:31:14.670 it's just kind of what your 00:31:14.670 --> 00:31:16.780 brain is used to in playing a non-spatialized game. 00:31:17.990 --> 00:31:20.005 So we took those sounds and we put them on the floor and 00:31:20.005 --> 00:31:21.740 they suddenly sounded like they were coming from the floor, 00:31:21.740 --> 00:31:22.310 as you'd expect. 00:31:23.370 --> 00:31:25.460 Now overhead we have one each for rain, so 00:31:25.460 --> 00:31:27.140 we have a left channel and a right channel for rain. 00:31:28.190 --> 00:31:31.380 Now our spatialized audio sounds should be mono, so we have 00:31:31.380 --> 00:31:34.000 slightly different sounds to make sure they're not phasing 00:31:34.000 --> 00:31:35.480 for left and right and they're playing overhead. 00:31:35.480 --> 00:31:38.100 You ever see the old cartoons where the storm cloud follows 00:31:38.100 --> 00:31:39.360 you around wherever you go? 00:31:39.360 --> 00:31:40.760 That's exactly what happens here. 00:31:42.010 --> 00:31:44.840 And the reason that we do that is rain is always overhead. 00:31:44.840 --> 00:31:46.030 It's positionally placed and 00:31:46.030 --> 00:31:47.890 it follows us around wherever we go. 00:31:47.890 --> 00:31:49.940 We don't need multiple ones set around our scene. 00:31:49.940 --> 00:31:52.710 If we walk notice our audio goes with us. 00:31:54.360 --> 00:31:56.930 We did the same left and right for an airplane. 00:31:56.930 --> 00:31:58.970 But there's a special case that comes with something 00:31:58.970 --> 00:31:59.589 like an airplane. 00:32:02.150 --> 00:32:05.100 When somebody records the sound of an airplane what do they do? 00:32:05.100 --> 00:32:07.193 Do they hang on the edge of an airplane with a microphone? 00:32:07.193 --> 00:32:08.017 They don't! 00:32:08.017 --> 00:32:10.166 [LAUGH] They're on the ground they're listening to that 00:32:10.166 --> 00:32:10.669 airplane and 00:32:10.669 --> 00:32:13.680 as such because the speed of the airplane you get Doppler shift. 00:32:13.680 --> 00:32:17.360 So, a sound out of the box that you record of an airplane 00:32:17.360 --> 00:32:19.800 has Doppler shift associated with it. 00:32:19.800 --> 00:32:21.370 So that's one of the settings you need to make sure that 00:32:21.370 --> 00:32:21.890 you turn off. 00:32:21.890 --> 00:32:24.910 We don't want additional Doppler sound included on 00:32:24.910 --> 00:32:26.450 audio that already has a Doppler built in. 00:32:26.450 --> 00:32:28.740 That was one of the challenges of getting an airplane sound. 00:32:29.800 --> 00:32:33.300 That also follows us overhead because the outside or 00:32:33.300 --> 00:32:34.710 any building you're in is gonna basically 00:32:34.710 --> 00:32:36.759 act like a big acoustic box or something like an airplane. 00:32:38.590 --> 00:32:40.635 Zombies same thing as the player. 00:32:40.635 --> 00:32:42.220 Zombie's groaning and 00:32:42.220 --> 00:32:45.670 moaning comes from the zombie's mouth which would be right here. 00:32:45.670 --> 00:32:46.220 And zombie 00:32:46.220 --> 00:32:47.970 walking sounds would have to come from the zombie's feet. 00:32:47.970 --> 00:32:50.989 It also makes sense but something to be aware of when 00:32:50.989 --> 00:32:54.089 you're designing these particular experiences. 00:32:54.089 --> 00:32:54.751 >> All right. 00:32:54.751 --> 00:32:55.795 Thanks Adam. 00:32:55.795 --> 00:32:57.105 Actually, can I show one more thing? 00:32:57.105 --> 00:32:57.607 >> Yeah, go ahead. 00:32:57.607 --> 00:33:00.001 >> I know you're gonna talk about profiling and 00:33:00.001 --> 00:33:02.836 performance but this is a pretty good graph that shows 00:33:02.836 --> 00:33:05.990 you immediately what happens with performance and audio. 00:33:05.990 --> 00:33:08.260 Let's go ahead and get my mouse up here. 00:33:11.500 --> 00:33:13.170 Scroll down to audio. 00:33:13.170 --> 00:33:16.920 This graph right here, this is my total audio CPU percentage. 00:33:18.600 --> 00:33:20.242 So let's go ahead and unpause this. 00:33:23.291 --> 00:33:28.460 Just watch the graph and let me shut-off spatialized audio. 00:33:28.460 --> 00:33:30.490 Turn it on, turn it off. 00:33:32.510 --> 00:33:33.770 There's a hit for spatialized audio. 00:33:33.770 --> 00:33:35.392 In this case it's pretty minimal but 00:33:35.392 --> 00:33:38.060 it is one thing to be aware of and using the profiler built-in 00:33:38.060 --> 00:33:40.677 unit you can actually see the audio CPU percentage that your 00:33:40.677 --> 00:33:42.322 experience is currently taking up. 00:33:44.390 --> 00:33:47.895 >> So let's talk about some tips while Adam 00:33:47.895 --> 00:33:51.410 is trying to get ready for his next demo. 00:33:51.410 --> 00:33:55.930 One of the things that is important for our spatialization 00:33:55.930 --> 00:34:00.428 technology is to make sure that all spatialized audio is mono. 00:34:00.428 --> 00:34:04.010 We also wanna make sure you target 48khz audio because 00:34:04.010 --> 00:34:07.500 in the pipelines, it will actually be converted to 48khz. 00:34:07.500 --> 00:34:10.705 AudioGraph itself will actually throw an error at this point if 00:34:10.705 --> 00:34:12.220 you do not use 48khz audio. 00:34:14.090 --> 00:34:17.070 What Adam and I learned is that when you get graphs from 00:34:17.070 --> 00:34:21.380 designers, they're not always at scale, and real live scale. 00:34:21.380 --> 00:34:22.390 >> Usually not at scale. 00:34:22.390 --> 00:34:25.570 >> Or usually yes. 00:34:25.570 --> 00:34:28.620 And the thing that is really important with spatializing 00:34:28.620 --> 00:34:32.150 sound is understanding that the spatialization technology we use 00:34:32.150 --> 00:34:34.330 assumes the units to be in meters and 00:34:34.330 --> 00:34:36.590 so one unit equals one meter. 00:34:36.590 --> 00:34:40.040 So when that world scale is off, spatialization sound and 00:34:40.040 --> 00:34:43.050 the distance curves and such will also be off a little bit. 00:34:45.180 --> 00:34:47.130 Adam also touched on HRTF performance. 00:34:47.130 --> 00:34:50.360 So, it is important to note that since we're Windows 00:34:50.360 --> 00:34:54.360 we run on many, many different devices, and different devices 00:34:54.360 --> 00:34:58.154 will support different amounts of supported voices. 00:34:59.300 --> 00:35:03.691 Now, It's also important to realize that you will budget for 00:35:03.691 --> 00:35:04.888 spatial audio. 00:35:04.888 --> 00:35:09.042 If it's 5% you'll have less spatial voices than you have 00:35:09.042 --> 00:35:09.740 10%. 00:35:09.740 --> 00:35:13.757 But again the 10% on a less powerful device might result in 00:35:13.757 --> 00:35:15.268 less voices. 00:35:15.268 --> 00:35:19.100 So, it is important for you to keep that in mind when you're 00:35:19.100 --> 00:35:20.249 designing your application. 00:35:21.720 --> 00:35:25.510 Below I have listed some common techniques on how you can 00:35:25.510 --> 00:35:29.410 deal with devices that are less powerful. 00:35:29.410 --> 00:35:32.220 So if you have a graph, the first graph on the left that has 00:35:32.220 --> 00:35:36.350 two emitters, two spatial objects on it. 00:35:36.350 --> 00:35:39.080 And you need less, or you have less spatial audio 00:35:39.080 --> 00:35:42.540 capability what you could do is take the emitters off those two 00:35:42.540 --> 00:35:45.027 individual nodes and sub-mix the sound and 00:35:45.027 --> 00:35:48.520 basically co-locating them by adding another emitter to it. 00:35:48.520 --> 00:35:51.630 So you would do that when sounds are really close in distance and 00:35:51.630 --> 00:35:53.140 traveling at a similar velocity. 00:35:54.690 --> 00:35:57.880 One of the other things you can do is put a priority 00:35:57.880 --> 00:35:59.850 on the spatial object you have. 00:35:59.850 --> 00:36:03.065 And if your spatial object count is not high enough or 00:36:03.065 --> 00:36:06.644 when you can render you might drop one of the emitters on one 00:36:06.644 --> 00:36:07.901 of the input nodes. 00:36:09.590 --> 00:36:12.010 What's important to notice is that spatial audio is 00:36:12.010 --> 00:36:12.960 a simulation. 00:36:12.960 --> 00:36:17.630 So as Adam mentioned before, you expect and you need sounds that 00:36:17.630 --> 00:36:22.145 come from the right location as you're creating your sound. 00:36:22.145 --> 00:36:24.995 That means footsteps, you wanna put an emitter that 00:36:24.995 --> 00:36:28.390 actually targets the footsteps at the bottom. 00:36:28.390 --> 00:36:29.650 Because that's really what you're used to and 00:36:29.650 --> 00:36:31.490 that's where the sound is coming from. 00:36:31.490 --> 00:36:33.220 Same goes for when somebody is speaking. 00:36:33.220 --> 00:36:35.636 You wanna actually put a speaker or 00:36:35.636 --> 00:36:41.554 an emitter where their head is Another 00:36:41.554 --> 00:36:45.790 important thing to note is that vertical location matters. 00:36:45.790 --> 00:36:47.270 So what I mean by that is that 00:36:48.460 --> 00:36:53.570 as you now have a high channel, it's important to understand how 00:36:53.570 --> 00:36:55.060 well that works at certain degrees. 00:36:55.060 --> 00:36:58.680 So anything above 30 degrees, 30 degrees works well. 00:36:58.680 --> 00:36:59.820 So if I'm looking at Adam and 00:36:59.820 --> 00:37:03.800 Adam is speaking at me that almost works well for me. 00:37:03.800 --> 00:37:07.060 But an airplane flying over or speakers or 00:37:07.060 --> 00:37:10.930 such are really good to locate because of that fact. 00:37:12.140 --> 00:37:14.710 The next thing Adam mentioned was 00:37:14.710 --> 00:37:17.549 the importance of room models and the good use of room models. 00:37:19.020 --> 00:37:22.290 When you are spatializing sound you need to use room models to 00:37:22.290 --> 00:37:25.080 really represent the room you're in because that will give you 00:37:25.080 --> 00:37:26.750 the correct distance cues and such. 00:37:28.040 --> 00:37:30.310 And so don't use them as an effect, but 00:37:30.310 --> 00:37:31.800 use them to really model the room. 00:37:33.840 --> 00:37:37.440 Something you might have experienced here is that 00:37:37.440 --> 00:37:40.000 expectation of wave simulation. 00:37:40.000 --> 00:37:43.038 So a rain always comes from above so immediately your brain 00:37:43.038 --> 00:37:46.230 thinks that the rain really is coming from above. 00:37:46.230 --> 00:37:49.388 And the next demo, if you hear of an eagle fly over, 00:37:49.388 --> 00:37:52.841 you will naturally be able to localize that sound better 00:37:52.841 --> 00:37:56.160 because eagles oftentimes fly over your head. 00:37:56.160 --> 00:37:56.800 Unless you're in trouble. 00:37:58.230 --> 00:38:01.920 And real world sounds are more effective for the same reason. 00:38:01.920 --> 00:38:04.420 Sounds that you know and 00:38:04.420 --> 00:38:06.190 sounds that are always in a similar location, 00:38:06.190 --> 00:38:08.540 you'll be able to place them better in those locations. 00:38:10.500 --> 00:38:11.860 When you're normalizing a sound for 00:38:11.860 --> 00:38:16.160 spatial audio, it's important to think about normalizing 00:38:16.160 --> 00:38:20.430 the sound as if it was recorded 1m away from you, and as such, 00:38:20.430 --> 00:38:22.760 the spatialization effect will work really great. 00:38:24.320 --> 00:38:29.120 You do not need to add any additional distance attenuation, 00:38:29.120 --> 00:38:31.520 so the spatialization technology 00:38:31.520 --> 00:38:33.170 panels will take care of that for you. 00:38:34.220 --> 00:38:36.002 If you have a sound in your UI, for 00:38:36.002 --> 00:38:38.330 example, is really important. 00:38:38.330 --> 00:38:42.331 You can add dynamic compression to make sure that that sound 00:38:42.331 --> 00:38:45.382 sounds louder when it's really important. 00:38:45.382 --> 00:38:49.631 All right, with that being said, let's move on to our next demo. 00:38:49.631 --> 00:38:51.548 >> My favorite demo. 00:38:51.548 --> 00:38:52.730 Well, the other one's good, too. 00:38:52.730 --> 00:38:53.580 They're different. 00:38:53.580 --> 00:38:55.440 Zombies and dungeons. 00:38:55.440 --> 00:38:56.430 You can kind of see a theme here. 00:38:57.560 --> 00:38:59.570 This next demo that we worked with another partner, 00:38:59.570 --> 00:39:04.340 Silicon Storm, and this going to be using the HTC Vive, one of 00:39:04.340 --> 00:39:07.770 the various capable VR headsets that are on the market, that 00:39:07.770 --> 00:39:12.230 actually can integrate quite easily with our audio runtime. 00:39:12.230 --> 00:39:16.030 And so on this one, this is a headset. 00:39:16.030 --> 00:39:18.020 I have stage tracking that's gonna be set up. 00:39:18.020 --> 00:39:19.626 But there's one thing I wanna show you inside of 00:39:19.626 --> 00:39:20.469 the experience first and 00:39:20.469 --> 00:39:22.251 then we'll actually do a quick run through on here. 00:39:33.535 --> 00:39:35.485 All right, let me set these on the floor just to make sure 00:39:35.485 --> 00:39:42.875 these are ready And I'm back. 00:39:42.875 --> 00:39:45.680 [LAUGH]. 00:39:45.680 --> 00:39:48.100 So we can see this level that we've got laid out here. 00:39:48.100 --> 00:39:50.957 And we're gonna start out on this far edge of the cliff 00:39:50.957 --> 00:39:52.745 here and there's gonna be wind. 00:39:52.745 --> 00:39:54.173 So take a moment to really appreciate and 00:39:54.173 --> 00:39:54.887 listen to sounds and 00:39:54.887 --> 00:39:56.576 try to understand where they're coming from. 00:39:56.576 --> 00:39:58.215 Imagine yourself on the edge there. 00:39:58.215 --> 00:40:01.195 I'm gonna be immersed so I'm gonna be in a slightly different 00:40:01.195 --> 00:40:03.599 place than you guys but when we go up the side here, 00:40:03.599 --> 00:40:04.822 we're gonna hear wind. 00:40:04.822 --> 00:40:07.446 Think, if you're going up the side of a cliff, and 00:40:07.446 --> 00:40:09.530 the sound of wind is gonna be different. 00:40:09.530 --> 00:40:12.650 So we've taken time to add different things to 00:40:12.650 --> 00:40:13.620 these locations. 00:40:13.620 --> 00:40:17.390 Now, the specialization settings, we can set manually. 00:40:17.390 --> 00:40:19.820 We showed little bits of code that we can use to essentially 00:40:19.820 --> 00:40:23.100 talk with, the specializer plug in. 00:40:23.100 --> 00:40:25.550 And we can simply do basic things, 00:40:25.550 --> 00:40:27.976 the set specializer float, the demo's that we looked at before. 00:40:27.976 --> 00:40:32.160 Set specializer float, in this particular case, 00:40:32.160 --> 00:40:37.380 we're just setting this to be a room of none, in other words, 00:40:37.380 --> 00:40:38.560 we are outdoors. 00:40:38.560 --> 00:40:41.040 We're not in a small room, we're not in a conference shaped room, 00:40:41.040 --> 00:40:42.600 we're not in a supermarket. 00:40:42.600 --> 00:40:43.650 We're outside, so 00:40:43.650 --> 00:40:45.560 it needs to change how it's gonna do the processing there. 00:40:47.260 --> 00:40:48.542 All right, let's go ahead, and run this demo, 00:40:48.542 --> 00:40:49.144 let me maximize this. 00:40:58.040 --> 00:40:59.060 Find my right controller here. 00:41:04.402 --> 00:41:05.563 Hopefully, I don't fall off the stage. 00:41:05.563 --> 00:41:06.398 [SOUND] 00:41:15.433 --> 00:41:18.580 And as it figures, my headset's not tracking, there we go. 00:41:18.580 --> 00:41:23.409 [SOUND] Gonna restart this real quick, 00:41:23.409 --> 00:41:24.970 for some reason, my headset's not tracking here. 00:41:24.970 --> 00:41:27.420 Let me just move back, so I can see me. 00:41:27.420 --> 00:41:29.543 The way that this headset actually works is we have, 00:41:29.543 --> 00:41:31.231 sensors that are set up on the stage here, 00:41:31.231 --> 00:41:33.740 and so this gives me kind of a realistic tracking throughout my 00:41:33.740 --> 00:41:34.386 environment. 00:41:36.488 --> 00:41:37.963 [SOUND] 00:41:50.277 --> 00:41:53.335 And, of course it's not tracking me in my environment. 00:41:53.335 --> 00:41:55.660 [LAUGH] There we go, all right, now we're in. 00:41:58.393 --> 00:41:59.482 Now take a moment, 00:41:59.482 --> 00:42:02.341 we have eagles that are flying up overhead there, 00:42:02.341 --> 00:42:05.960 we have the sounds around us, we have flags that are flapping. 00:42:07.720 --> 00:42:10.710 You can hear the winds, we kinda got that edge over there. 00:42:12.520 --> 00:42:15.063 I'm gonna violate a little principle of VR, and 00:42:15.063 --> 00:42:18.164 fast essentially move myself through this world using this 00:42:18.164 --> 00:42:19.542 little way point system. 00:42:19.542 --> 00:42:21.319 Now automatically notice, or Steven, 00:42:21.319 --> 00:42:22.631 are you over there, Steven? 00:42:22.631 --> 00:42:23.489 >> [LAUGH] Yes. 00:42:23.489 --> 00:42:26.671 >> We can hear the edge of this wind zone here. 00:42:29.226 --> 00:42:31.026 We're at the very edge of it, let's go into it 00:42:31.026 --> 00:42:41.026 [SOUND] 00:42:43.866 --> 00:42:45.667 All right, so hopefully you felt 00:42:45.667 --> 00:42:48.893 a little bit of that chill that was kind of on the edge there. 00:42:48.893 --> 00:42:51.352 Now we're coming up to our dungeon here, 00:42:51.352 --> 00:42:54.802 Steve you want to bring me up to that spot there, next one. 00:42:54.802 --> 00:42:58.372 All right, 00:42:58.372 --> 00:43:03.370 let's go into. 00:43:03.370 --> 00:43:09.790 We go into our area, our flags shut off outside. 00:43:09.790 --> 00:43:11.110 Our eagles shut off outside, 00:43:11.110 --> 00:43:13.110 we need to make sure we don't hear those sounds any more. 00:43:13.110 --> 00:43:15.070 As soon, as we've entered into this region we've shut 00:43:15.070 --> 00:43:16.020 the sounds off. 00:43:16.020 --> 00:43:17.529 Now we have torches burning on the side. 00:43:20.830 --> 00:43:21.840 Let's move to this area. 00:43:23.930 --> 00:43:26.247 Again, we can literally hear the fire burning on the sides. 00:43:33.388 --> 00:43:35.730 And in the dungeon, we have a couple of different traps 00:43:38.484 --> 00:43:40.381 There's a slight creaking going on of the, 00:43:40.381 --> 00:43:42.943 the lumber that's kind of holding entire place up here, or 00:43:42.943 --> 00:43:44.240 helping to hold the place up. 00:43:46.070 --> 00:43:48.161 And, I can see there's a trapper right there, up on top, so 00:43:48.161 --> 00:43:49.606 let's check out what that guy's about. 00:43:54.772 --> 00:43:55.712 >> Up. 00:43:55.712 --> 00:43:59.683 [SOUND] >> Stop tracking. 00:43:59.683 --> 00:44:04.792 >> Okay, okay. 00:44:04.792 --> 00:44:07.752 >> Right when the trap drops it stops tracking. 00:44:07.752 --> 00:44:09.065 Did you see it or no? 00:44:09.065 --> 00:44:09.990 >> Yeah, I did. >> You saw it? 00:44:09.990 --> 00:44:13.170 >> Yeah. >> Because the other thing that 00:44:13.170 --> 00:44:14.270 we can do >> You said, 00:44:15.840 --> 00:44:17.400 you were able to see the trap drop on there? 00:44:17.400 --> 00:44:19.501 All right, there's on other thing, that I'll show you, and 00:44:19.501 --> 00:44:21.210 we don't actually have to have the headset on. 00:44:21.210 --> 00:44:25.994 The cool thing about technology nowadays 00:44:25.994 --> 00:44:30.640 is we can pretty easily switch back, and 00:44:30.640 --> 00:44:36.516 forth between, what we're doing with a headset, 00:44:36.516 --> 00:44:41.588 and what we're gonna do without a headset. 00:44:46.047 --> 00:44:49.567 >> So over this trap, on the right hand side here. 00:44:49.567 --> 00:44:51.561 There's a lava pits that we passed by. 00:44:51.561 --> 00:44:54.055 [SOUND] One more trap. 00:44:54.055 --> 00:45:02.413 [SOUND] So 00:45:02.413 --> 00:45:04.671 arrows that have audio sources on them, so 00:45:04.671 --> 00:45:06.310 we can hear them as they fly by. 00:45:08.080 --> 00:45:11.070 And then as we, if you notice, they hit the back wall, 00:45:11.070 --> 00:45:13.365 hopefully you heard it somewhat, behind you. 00:45:13.365 --> 00:45:15.530 Cuz, those hit the back wall, 00:45:15.530 --> 00:45:17.548 and they made a sound at that location when they hit. 00:45:17.548 --> 00:45:20.580 So the specialization aspect, I hope you can hear it brings to 00:45:20.580 --> 00:45:23.550 live, now with headphones on, it's kind for the side speakers 00:45:23.550 --> 00:45:25.940 you've got kind of an idea where those sounds are coming from. 00:45:27.100 --> 00:45:31.252 >> Great, thanks Dan for the demo, 00:45:31.252 --> 00:45:36.230 so Let's switch to, the deck again. 00:45:36.230 --> 00:45:39.270 All right, so that's the end of our presentation. 00:45:39.270 --> 00:45:42.030 I hope you've seen, couple great demo's that demonstrate, 00:45:42.030 --> 00:45:44.450 how you can add spatial audio to games, but 00:45:44.450 --> 00:45:48.220 also how you can add spatial audio, to your applications, 00:45:48.220 --> 00:45:50.270 in udev lp using the audiograph API. 00:45:53.000 --> 00:45:54.300 We can't wait to really see, 00:45:54.300 --> 00:45:56.790 what kind of spatial audio experiences you create, and 00:45:56.790 --> 00:46:00.270 you can reach out both to me, and Adam afterwards to, 00:46:01.350 --> 00:46:04.820 you know, we'll help you out with building those experiences. 00:46:04.820 --> 00:46:08.170 You can also watch the separate recorded presentation, 00:46:08.170 --> 00:46:11.070 on the spatial sphere on channel nine. 00:46:11.070 --> 00:46:14.820 And we will post code samples for the code samples, for 00:46:14.820 --> 00:46:16.960 a lot of the demos you've seen here today on github. 00:46:17.970 --> 00:46:20.830 And, we'll also post the games for 00:46:20.830 --> 00:46:24.390 which we can't always, share the full resources. 00:46:24.390 --> 00:46:26.280 But we'll put those online as well. 00:46:27.700 --> 00:46:30.750 We'll post the link to both of them on our channel 00:46:30.750 --> 00:46:31.469 nine video page. 00:46:32.600 --> 00:46:35.620 All right, well thank you for your time today. 00:46:35.620 --> 00:46:36.850 I hope you had a great time, and 00:46:36.850 --> 00:46:40.270 I hope you're ready to build some spatial audio experiences. 00:46:40.270 --> 00:46:41.120 >> Thank you all very much, and for 00:46:41.120 --> 00:46:43.050 the folks in the room, we're gonna stay a little bit after. 00:46:43.050 --> 00:46:44.670 So we'll be here to answer questions, and if you wanna 00:46:44.670 --> 00:46:46.190 check out the headset check anything with that. 00:46:46.190 --> 00:46:47.100 Come on up, and talk to us too, 00:46:47.100 --> 00:46:49.070 thank you all very much, enjoy the rest of build. 00:46:49.070 --> 00:46:50.980 >> Thanks a lot.