There are a number of different 3D audio plug-ins available for game engines like Unity, but there have not been any solutions for mixing 3D audio for cinematic VR productions. Two Big Ears just released the 3DCeption Spatial Workstation, which is a digital audio workstation that integrates with professional audio production tools. The Spatial Workstation includes an Intelligent 360 Video Player to be able to preview audio edits in a VR environment, as well as the ability to mix ambisonics audio channels with additional audio inputs. Varun Nair is the VP of Products and co-founder of Two Big Ears, and I had a chance to talk about the 3DCeption Spatial Workstation, ambisonics, performance metrics, and how he sees virtual reality is facilitating the combination of audio post-production techniques from traditional game design and film production. You can register your interest in the Spatial Workstation here.
Listen
https://www.youtube.com/watch?v=i4ais9UcVvE
Become a Patron! Support The Voices of VR Podcast Patreon
Theme music: “Fatality” by Tigoolio
Subscribe to the Voices of VR podcast.
Rough Transcript
[00:00:05.412] Kent Bye: The Voices of VR Podcast.
[00:00:12.042] Varun Nair: I'm Varun, I am one of the co-founders at Two Big Ears, and I also lead product development across a whole range and suite of products. We're an audio company, we do everything from technology, that is low-level audio rendering technology for VR and AR, all the way up to workflow tools and complete pipelines to help people go from nothing to a fully working application. We are heavily involved in the interactive side of things like games and interactive VR experiences and have been for the past many years. And over the past two years or so we've also started looking at the whole cinematic virtual reality side of things and building tools and rendering technology to make a lot of those experiences come alive with better and immersive audio.
[00:01:00.147] Kent Bye: Great, so one of the announcements that you're just making here is that you're releasing what you're calling the CD-Ception Spatial Workstation, and maybe you could talk a little bit about what is included in this pipeline in terms of being able to create 3D audio with a cinematic VR setup.
[00:01:18.424] Varun Nair: I think it would be useful to also bring in a bit of context and why we decided to actually develop these tools in the first place. So much about virtual reality, as we know, is undefined, both from the technology and the creative side of things. Everyone in this space is starting to slowly understand what the medium is capable of and also what these new technologies are capable of. And one of the big issue for creatives and people putting together such experiences is having the right tools and a good workflow to go from an idea to a fully working product that sounds and looks great. On the more interactive side of things, we've got that covered thanks to game engines and all the audio middleware tools and engines out there, and many of the spatial audio plugins, including ours. But from the cinematic side, it's completely undefined. And there has been this huge gap of giving audio professionals the right tools to go and actually create great sounding 3D experiences that react to head tracking in virtual reality. So the tools we've developed allow sound designers to use existing audio workstations that they've probably spent years or a whole decade working with. So they can work with familiar tools and we augment those tools to allow them to mix directly to a VR device and design sounds directly to a 360 video and bring in linear post-production techniques and interactive game audio techniques sort of into the same place. And more importantly, go from there all the way down to rendering directly on the device and ensuring that the final application ends up recreating that audio mix that they've spent so much time creating and ensuring that it all gets recreated faithfully across the board.
[00:03:02.437] Kent Bye: And so, you know, I've done a number of interviews looking at audio within virtual reality and cinematic VR. Are you taking an approach which is kind of like an omni-directional binaural or are you using something like ambisonics in your approach?
[00:03:16.438] Varun Nair: So, ours is very much a post-production tool set, so you can, I guess the simplest way to describe it would be to think of it like how a game engine works, where you've got different kinds of input data, whether it could be 3D models, textures, and so on, and you've got the output, which is your game. Except in this case, your input can be all sorts of audio, whether it's mono, stereo, ambisonics audio, everything could come together in a single place, and then you can manipulate it, So you can use production sound, sound that is recorded on set using an ambisonics microphone or some of the more traditional recording techniques like lapel microphones, boom microphones, and work with all of the data, but also add extra data on top of it. So sound effects and music and additional dialogue. And you can then reposition all that audio. and get it to react to head tracking, get it to react to interactive elements within the cinematic experience. So it essentially is a middleware where all of these techniques and technologies come together and you've got this whole playground to take your project in whichever direction you'd like.
[00:04:21.738] Kent Bye: And so in the Unity interface, you're basically working with this kind of 3D space. And traditionally in film editing tools, it's pretty flat, pretty 2D. So with your workstation, how are you creating an interface for people to actually kind of create these 3D positional audio objects for a cinematic experience?
[00:04:43.872] Varun Nair: Sure. So what we provide in our tool set is a whole video playback mechanism that is closely linked to the audio workstations. So both systems share a lot of data between themselves. So a sound designer can play back a 360 video within the same system. and directly preview how the sounds are being positioned in space. So if you've got a helicopter flying over your head in the video, the sound designer can quickly position that sound right up there in their digital audio workstation and look around the video in real time and look at how it all reacts or can hook up an Oculus Rift and get all that head tracking working. seamlessly within the same environment. The key here is to not jump between different workstations and different setups, but to work in the same place and quickly iterate through ideas without having to get into more complicated workflows.
[00:05:41.142] Kent Bye: So I know a lot of audio post-production is used in Pro Tools. So is what you're creating something that is just a standalone application? Or is this something like a plug-in that you could continue to use, kind of like your favorite audio workstation?
[00:05:56.927] Varun Nair: Yep, so we provide a whole suite of plugins, a video playback mechanism, some video playback software, and a suite of additional standalone tools. All of these work together, but sort of center point is still your digital audio workstation, like Pro Tools. This is really key because software like Pro Tools are really, really good at audio editing and bringing post-production sort of soundtracks to life. Our tools augment that and sort of bring in that layer of interactivity and combine virtual reality and game audio workloads with the linear post-production workflow in that same environment, which sound designers are very familiar with.
[00:06:41.008] Kent Bye: And so yeah, maybe you could talk about some of the actual microphones that you foresee is going to be the leading solution. Because I know that I've talked to Jaunt. They're using more of an ambisonics approach. And then there's some people that were working on an omnidirectional binaural microphone. What do you see is going to be kind of like the leading approach for what actual hardware you would need to do kind of like a fully specialized audio for a cinematic VR experience? And then what the actual workflow would be for them to take that recording and then do some work in post-production in order to make it fully polished.
[00:07:19.004] Varun Nair: Yeah, so ambisonics at this point does seem to be the solution that solves a lot of problems. It's a technology that has been around for a really, really long time. So there's a lot of infrastructure as far as microphones and techniques to actually record audio. It's been around for decades. So it's a really convenient format. And it's also easier for camera companies or other microphone companies to build on the technology. Our pipeline workflow supports ambisonics right away. So if you've got an ambisonics, ambisonic recorder, you can just import it into Pro Tools or whichever DAW you're working with and use it right away. You can polish it and change some of the environment settings and so on so that it can fit better with the scene you're creating. I think a lot of the camera companies are going down the ambisonics route and it makes perfect sense because it doesn't restrict any of the post-production techniques. It opens up a lot more possibilities. And the good thing with ambisonics is that it can be translated to different playback mechanisms. So if you're using binaural audio, rendering it out into binaural audio, it could be used for that. Or if you're using some sort of a speaker system, it could be used for that too. So it's quite a convenient piece of technology. I think the bigger problem is also from the hardware side, where some of these microphones can be quite bulky. And when you're using a 360 camera rig, the biggest problem is finding the right spot to put that microphone. But we're working closely with many partners in this area, from hardware to software, and I think the next six to 12 months should be very interesting. We're gonna see lots of clever solutions that tackle this problem.
[00:09:04.097] Kent Bye: Yeah, so I guess, you know, what's sort of like the framework for, you know, the cost of this new workstation that you're putting out and some of the biggest features that you see?
[00:09:16.687] Varun Nair: The biggest features at this point are cutting down the number of barriers to go from just having your sounds all the way up to a fully polished application or experience. It's really quick. We sort of took a unique approach, at least as far as our history is concerned, a unique approach to releasing this product. It was ready early this year, but we just worked with a few select partners and we gave them special access to the tools and we've been working very closely with them. Many of these projects you'll hear about in the coming months. And so it has been built with a really strong feedback loop with some of the leading industry players in this space. So it's quite a refined workflow and more importantly we've worked on a lot of technology under the hood over the past many years that brings positional tracking and some of the newer VR technologies and moving up to where augmented reality is going and tool sets becoming a sort of nexus for a lot of these things to happen. As far as pricing, we've got various models in place, depending on the project and depending on the kind of solution they're going for. Some of the larger companies we're working with have got a lot of content or plan to have a lot of content over the next few years. So we tend to vary the pricing accordingly and tailor it to specific needs. Just like with games, we're also aware that there are a lot of independent filmmakers and game developers out there and it is always in our interest in supporting independent filmmakers, independent developers and smaller projects that don't really have budgets as well. So we think it's really important to invest in people who are experimenting and exploring the medium because that's the best way for all of us to actually learn. So we tend to tailor our pricing accordingly based on the needs of the project.
[00:11:20.013] Kent Bye: Great, I know that back in April of this year, Anastasia Davana wrote a whole just epic write-up comparing all the different audio options. And it looked like actually that the plugin from 3Dception, your company Two Big Ears, kind of came up on top in terms of all of the different features that were supported, especially when it comes to, one of the things that she did there was actually implement this app, an Android app called Skeletons versus 3D audio where you could actually listen to all the different solutions. And so maybe you could talk a little bit about like what are the things that you're doing in your plugin, both for the Unity plugin that you don't see other companies doing quite yet.
[00:12:06.513] Varun Nair: One of the reasons why we started 2BigEars was trying to find a better way to render 3D audio across devices all the way down from low-powered devices to high-powered gaming computers. This forced us to really look at things at an atomic level and rework a lot of the binaural techniques from scratch. I'm really glad we did this. We spent a lot of time a few years ago and since then building on our rendering algorithms and it was time worth invested because we've been able to build a very scalable and high-quality solution that works from low-powered Android devices all the way up to the more powerful gaming computers. That is to do with our underlying binaural rendering technology, but we've been working on a lot of cool features that add on to it and help create that sense of immersion, ranging from room modeling algorithms to real-time game engine, geometry-based audio reflections and occlusion of sound. All of these are components that just build on the binaural rendering technology and make it sound better. So the cinematic tool set is built on the same rendering technology so it makes it really scalable and one of the key things with this whole cinematic workstation is that we're probably one of the few or maybe even the only company at this point that's got native browser support so you can actually get an app working in the browser right away. It's still early days, quite experimental because of the state of the browsers but This underlying technology has helped us quickly scale and support a lot of platforms easily while still using the same workflow and we've got quite a unique team from different backgrounds so it actually helps hugely when it comes to creating better workflows and interfaces to design better audio.
[00:13:59.532] Kent Bye: What is the sound format that you're actually using that's being put into this 360 degree video? Because how are you able to kind of like track the head movements? And then what is it that is able to kind of translate that sort of HRTF into kind of like what audio is actually coming through?
[00:14:19.288] Varun Nair: We've had to develop our own audio format and codec to actually make all of this happen. So it borrows a few ideas from the games industry where audio is usually packaged into what we call a sound bank. So it's essentially a single file that's got a whole bunch of audio and metadata. We borrowed a few ideas from there and we worked on very high quality lossless algorithms so we can crunch all of that audio and metadata down into a really small file. Some of the early tests we had done with the system, one of the key requirements a lot of the studios had was to have really small asset sizes because they're either streaming it over the net or downloading all this content to an Android phone that usually doesn't have too much space. So we've worked quite hard on crunching it all down into a single format that is easily translatable across all these systems. But that said, it is an interesting time in this whole space. There are lots of 3D audio standards that are being worked on and there's a lot of competition between them as well. We've taken the approach of just waiting and watching to see how these standards build over time. And with VR being so new, I think we still don't really know what those standards should be like. So we are very much focused on contributing to an open environment and we should be able to share a lot more about what's happening under the hood over the coming months.
[00:15:42.814] Kent Bye: There seems to be specific issues when it comes to VR and latency and I know that visually there's like the motion of photons that has to be around 20 milliseconds. What is the kind of threshold for latency when it comes to audio and the specific challenges that are involved with being able to dynamically change the sound based upon how people are moving their head?
[00:16:06.303] Varun Nair: Latency is a complicated area, especially on Android. Even for audio, Android is a bit infamous for its audio latency. But that said, companies like Samsung have managed to implement their own solutions. So, the Gear VR actually has low latency. And some of the more recent Android devices have a low latency path as well. But that's more at a hardware and audio driver level. When it comes to HRTF and binaural processing, if you follow the more textbook implementations or the most standard implementations of binaural audio, you find that latency can be quite high depending on the algorithm you implement. This was also one of the key factors we took into consideration when designing our algorithms a few years ago. I was trying to achieve as little latency as possible. At this point, we had a miniscule amount of latency that is almost negligible to the system. When it comes to end users actually detecting that latency, I think the good thing with audio is that in some cases, a larger amount of latency is a bit forgivable, but really depends on the use case. If it is something that is interactive or reactive, the latency needs to be quite low. Sometimes 20 to 30 milliseconds sometimes can work okay. But if it is a more passive experience, you can get away with larger latency. Some of the Android devices out there tend to have latencies of about 80 to 120 milliseconds, which is quite a lot. So it really varies and depends on the experience as well.
[00:17:42.253] Kent Bye: When you're creating an experience like this, how do you go about actually measuring and troubleshooting or debugging the latency?
[00:17:51.695] Varun Nair: That's really difficult, especially on mobile devices. So latency can be added in, can happen in multiple stages. So there's the algorithmic latency, which is within the binaural rendering solution. If that solution does add any latency, that can be quite easy to measure. You pass audio into the system, record that audio back out. and you sort of know how much that signal is delayed by. When it comes to latency at the hardware level, that can be quite difficult to measure. The Android audio team gave a talk about, I think, two years ago at Google I-O, where they came up with a clever hardware solution to measure audio latency, where they used the LED on the Android phone to blink at the same time the audio event had to be played back. They measured the difference between both of those streams to actually figure out what the latency of the system is. There's a lot of documentation online and lots of techniques to actually go about measuring it.
[00:18:54.700] Kent Bye: Maybe you could just describe what are early reflections?
[00:18:58.003] Varun Nair: Sure. So in real life, if you heard me speaking in a room, what you hear is a mix of my voice just by itself, and the sound of my voice reflecting around objects in the room. If you are in a larger space like say a church and if I spoke you would hear my voice and you would hear a lot more of these reflections happening and because you're in a larger space some of those reflections take time to happen or might bounce off multiple surfaces before you actually end up hearing it. It's those reflections that actually provide a lot of crucial information when it comes to spatialization. It helps us understand where a sound is in space. It helps us understand how far that sound is from us. And it also helps resolve a lot of spatial cues that our brain then uses to position that sound. So if you are in a church, you hear all those reflections. And if it's a pretty big church, you might hear those reflections running for as long as maybe three or four seconds, because it's in a really large space, the reflections tend to bounce off a lot of surfaces and then reach you. So it could take a few seconds for that to happen. So if you look at all those reflections happening right from the origin of the sound all the way down to when those reflections are inaudible, we call that whole chunk reverberation. And reverberation can be broken down into multiple components where one of which is early reflections. Early reflections are the first few reflections that reach your ear. So usually, depending on the size of the space, this could be anywhere between the origin of the sound to about 80 milliseconds, but that number is variable. So these are the first few reflections that bounce off objects in the room. And these are the reflections that are crucial to spatialization. They help us understand the position of that object. They help hugely with what we call externalization. That is, it is the factor that makes you believe that the sound is out there in the environment and not just playing back inside your head.
[00:21:11.086] Kent Bye: So with the 3D Ception Spatial Workstation, I know in the Unity plugin you have a number of presets of different reflection presets. Are those same reflection presets going to be available within the 3D Ception Spatial Workstation? Or is it going to be something that is going to be derived from the live recording of the Ambisonics?
[00:21:36.196] Varun Nair: The Spatial Workstation does include the Early Reflection Modeling System with similar presets. And we'll shortly be releasing a huge update on that whole front, both on the game engine side and the Spatial Workstation as well. with a lot more features and resolution across the board. So that is available in the Spatial Workstation as well. So the sound designer can use existing recordings from the location. They could use the early reflection system we have. They could also combine it with any other reverberation solutions they use in post-production. So it's quite a modular system that can fit into multiple components.
[00:22:19.365] Kent Bye: And maybe you could talk a bit about, like, what is the inputs that you're taking for ambisonics? Like, what are the different channels? And then how is that used to kind of create a sound field?
[00:22:30.935] Varun Nair: Ambisonics is based on a concept of spherical decomposition. It's quite a math-heavy topic. But when you use an ambisonics microphone, you end up recording a multi-channel audio file. The majority of the ambisonics microphones out there tend to record first order ambisonics. So it's first order ambisonics all the way up to higher order ambisonics which is usually third order. So with the first order ambisonic recording you end up with four channels of audio. And these four channels can then be imported into your audio workstation and run through our plugins, which can then help you reorient that sound field, change various parameters in that sound field, such as how diffused that sound field is, how much of that reverb you actually hear in that space and so on. And a lot of information is then extracted from that Ambisonics recording and then passed down to our system. which is then combined with all the other audio within your workstation that is then spatialized into a full binaural sound field.
[00:23:37.834] Kent Bye: What is the difference between first, second, and third order ambisonics then?
[00:23:42.941] Varun Nair: It's usually down to spatial resolution. I think the general consensus is that higher order ambisonics can sound better, especially as far as spatial resolution is concerned. So by spatial resolution, I mean the accuracy and positioning of a sound in space. But higher order ambisonics also result in more channels of audio and can also result in more complicated microphone setups, which is why a lot of the commercial ambisonic microphones are first order, because they are simpler and cheaper. But you do get more complicated microphone setups that do higher order as well.
[00:24:21.182] Kent Bye: So is that higher order, like let's say you're shooting a scene with like two or three actors, would you individually mic each of the actors, measure the distance of how far they are from the camera, and then, you know, what kind of inputs would you be needing in order to kind of add more audio input sources?
[00:24:38.910] Varun Nair: Sure. You can think of an ambisonic microphone and just like a 360 video camera. All it does is it just captures audio around what ideally should be a single point in space. That's all it does. It just captures all the audio around it and describes it in a format that can then be used to reconstruct that sphere of audio around the microphone. We usually recommend sound designers and production teams to use an ambisonic microphone, but to also use lapel microphones and individual actors and boom microphones if they can place a boom microphone in that environment, but to record as much additional audio as possible. Because in a lot of cases, ambisonic recordings can work well by themselves, but would need reinforcement. And in some cases the ambisonic recording would be alright but it might not sound good enough for that project because maybe the project requires a hyper-realistic soundscape and all that ambisonic microphone gives you is just quite a realistic representation of the location. And there are other complications to it. A lot of times, production companies tend to choose locations for a shoot based on how it looks and not how it sounds. So you might have the most beautiful location, but you'd hear the traffic off in the distance, or you might hear a plane flying over your head, which might not be needed for the soundtrack. So in a lot of cases, the ambisonic mic recording can be used as reinforcement or even as a reference, which a sound designer can then use to reconstruct soundtrack from scratch. And that is already a common practice in cinema. Almost everything you hear in a movie has been reconstructed.
[00:26:21.014] Kent Bye: And so how do you kind of position those additional lapel mics in that 3D space then? You know, do you have to actually know how far away each of the actors were from the ambisonics microphone?
[00:26:34.057] Varun Nair: You could take measurements, but in a lot of cases it can be done by ear. A really experienced sound designer would be able to just take that sound and position it by ear and based on what they see visually. So that is already what they do in cinema when it comes to 5.1 or Dolby Atmos systems or anything like it. And in a lot of cases, a realistic measurement might not be what you need. A lot about audio is about trying to design it in a way so that the audience feels something. So you might want to get one of the dialogues to be louder simply because you want to increase tension in the scene. or give the listener or the user the impression that the actor is getting a bit closer to them. Well, visually, they might not. So there are lots of special tricks that can be done under the hood. And while a measurement might be useful, it can very easily be done by sight and hearing.
[00:27:30.433] Kent Bye: When it comes to the additional CPU load that is required to be able to render out the spatialized audio, I recognize that when you're using something like a game engine, there's a lot of that CPU power that'd be driven towards actually generating the visuals, but yet, in a cinematic VR experience, it seems like that might be a little bit less intensive on the CPU and GPU in order to actually drive the experience. So does that mean that you're able to do more sophisticated audio processing for a cinematic VR than you would say something that's a video game?
[00:28:07.052] Varun Nair: I wish that were true because I think with very simple 360 or cinematic VR experiences, you might not need a lot of CPU, but with some of the more high quality ones, there's a lot that a lot of the visual effects companies and some of the production companies out there are doing under the hood to just get the video to render better. So a lot of CPU resources can be spent on video decoding if they're not using any sort of hardware acceleration. And there's a lot of discussions happening right now about the H.264 video codec not being good enough for 360 video because depending on compression you might actually end up seeing some of those stitch lines come together. So a fair amount of CPU resources is used for video decoding and adding a layer of interactivity to that video. And if you are rendering it out to a VR device, there is additional processing that needs to happen for head tracking and all of the other features in there. But compared to games, I'd say there is maybe a little more CPU budget for audio. But what we are also dealing with is a range of mobile devices and a range of desktop devices where some of these devices might be powerful enough or some might not. So it's still really important to have an audio engine that's very efficient and scalable across all these systems.
[00:29:31.649] Kent Bye: What have been some of your own internal benchmarks in terms of the load that you have using spatialized audio? If you were to do just a normal 2D audio solution versus using your approach of doing spatialized audio, what are you seeing in terms of both mobile and desktop PCs?
[00:29:51.118] Varun Nair: Let's tackle game engines and real-time game audio first. So generally just using 2D non-HRT if not binaural audio is quite cheap as far as CPU resources are concerned because computationally there's nothing much happening under the hood when you need to position sounds in normal stereo or 2D. Binaural processing is known to be an intensive process and generally a lot of the spatial audio solutions tend to compete on computational resources and quality. So we've generally seen that when it comes to binaural processing, comparing it to stereo, you might need about, I could give you a quick example actually. So if you're running a game on a decent spec Windows machine, maybe normal stereo panning might take up about let's say 0.05% of CPU. If you did use one of the 3Dception products, it would be about 0.07, 0.1% depending again on the CPU. So that's the kind of difference we're seeing. And the interesting thing is that a lot of these mobile devices are getting faster and almost as good as some of the mid-level desktop devices out there. The only major issue with mobile devices is power consumption, and that's not just to do with audio, that's to do with every single component. With the cinematic side of things, we've worked on a system that, as I mentioned previously, is quite scalable. Internally, we've got some clever algorithms that limit the amount of CPUs you use. As far as the application developer's concerned, they just need to know that about 8-10% of the CPU is just allocated for the whole audio engine, so this is not just spatial audio rendering, but streaming the audio files, playing back those audio files, a lot of the event-based mechanisms. So on the Galaxy S6, for example, all you need is about 5% for audio or 6% for audio and that's enough. On a Nexus 5 you might need about 8-10% and we've tried to limit it to about 10% of CPU on a single core so that there's still enough resources available for video, user interface, head tracking, rendering, or any of the other components that need to go in place.
[00:32:10.763] Kent Bye: Nice. And so, yeah, just to kind of wrap things up here, I'm curious about what you see as kind of like the ultimate potential of virtual reality and what it might be able to enable and how you see 3D deception kind of in that picture.
[00:32:27.836] Varun Nair: With everything we're doing, we're just looking at everything we're doing in this space is really starting steps to the bigger picture. We see the bigger picture being the convergence of mediums, which you could think of as, say, games and cinema converging into some sort of a single medium where you really no longer know whether it's a game you are in or a movie you're in. We are starting to see great examples of that already. but also convergence as far as technology is concerned. So moving from virtual reality to mixed reality and bringing in augmented experiences sort of into the same platform and audiovisual experience. And we see RL as moving forwards and developing scalable and high quality audio solutions and not just rendering technology, but also the tools to cater to this convergence and it's been really interesting to see this already happening over the past year where we've had people from strong post-production backgrounds starting to embrace some of the game audio technologies and workflows and people from game audio workflows starting to embrace a lot of the techniques and technologies from post-production. We're starting to see that happen at a really small level and our expectation is that it's just going to blow up into the metaverse.
[00:33:51.805] Kent Bye: Is there anything else that's left unsaid that you'd like to say if people want more information and they want to get their hands on this 3Dception Spatial Workstation? What do you suggest they do?
[00:34:02.669] Varun Nair: They can just head out to our website. We are very, very well known for a good support team. So just head out to our website. You can get access to it. If you've got any questions, just shoot us an email and we'd be more than happy to respond and help you make better and more awesome projects.
[00:34:19.159] Kent Bye: OK, great. Well, thank you so much for joining me today.
[00:34:22.682] Varun Nair: More than happy to. Thank you.
[00:34:24.724] Kent Bye: And thank you for listening. If you'd like to support the Voices of VR podcast, then please consider becoming a patron at patreon.com slash Voices of VR.
