NETINT Technologies about a Video Platform used by 97% of US Football Teams
Well, it's 8:00 AM, on Pacific Time anyway, so on the West Coast, I guess. Hey, welcome to this edition, another very, very exciting edition of Voices of Video. Casey, are you ready to have an awesome conversation? Absolutely. Let's do it. All right. Well, good. Well, let me introduce you. So as everyone knows today, I'm going to be talking with Casey Bateman, who is with Hudl, and you're going to hear all about Hudl. But we were just chatting, Casey and I, about how the company was founded. I'll
let him tell the full story, but they're actually one of the OGs in streaming video, and I have been fortunate enough to have been a part of, not Hudl, but some other companies in the same vintage. It's amazing how on one hand, 15 years ago is 15 years ago, and yet it seems like yesterday how far we've come in streaming video. Casey, welcome to Voices of Video. Thank you for joining me. This is going to be great. Yeah, thanks for having me. I'm excited to be here. Yeah. Cool, cool. Well,
let's start, obviously give a quick intro to yourself and tell the viewers what you do, and then talk to us about Hudl, what Hudl's building, and let's get into this. Yeah, absolutely. Yes, I'm Casey Bateman. I'm a principal engineer in the office of the CTO at Hudl. Specifically what I do is I look into our video technology stack. I kind of own that area, technically speaking. My job is to look at new technologies, assess the landscape, and to try and plot out a strategy for the company around our video vision and how we're to execute on that. It's a big job plotting the landscape and the vision of the future. But tell us about Hudl,
because you are a remarkable company, especially... Well, you're, I'm sure, a remarkable company, but in the area of streaming. I have to believe that probably not all of our viewers even know who Hudl is, and yet, I think everyone's going to be shocked when they hear the scale of operations that you run and what you do. So what does Hudl do? Hudl is... We are a technology company that focuses on empowering coaches, athletes, and fans' sports experience. We build a coaching tool platform that really helps coaches and
athletes work their game and get better at it. We have a platform then also for our fans who can't necessarily be at games, or are at games and want to enrich their experience at those games, or even just follow their favorite athletes and teams, just watch that content. We're like a full service sports video suite of products. Yeah. Yeah. You were saying OG Company. Yeah. Company started back in 2006. Got our first customer,
was actually in 2007- Incredible. ... and stuff, with the University of Nebraska. Wow. Wow. What's the origin story? Are the founders athletes that were trying to solve a problem, or technologists that were just looking for an application? Both? Neither? Yeah. Well, actually, yeah, our three founders is Dave Graff, John Wirtz, and Brian Kaiser. They were attending the Raikes School at the University of Nebraska, which is an entrepreneur computer science program, the stuff that the University of Nebraska has. Dave Graff, who is our CEO, was working with the football team at that time. Just to paint a picture of how it was back then for you, these teams, when they'd exchange video, they were doing VCRs. A lot of time they would have their video recordings,
they would cut everything and record to a VCR, and then meet other coaches, drive and meet them halfway. It was- And exchange tapes, right? Exactly. Literally. Exactly. This is what it was like, and it is hard to think, based on where we are now, there was a past in which that was the thing to do. Dave Graff had the idea at that point in time, and John Wirtz and Brian Kaiser joined in, and they developed a product they called Huddle. Now, it was spelled a little differently back then, like the actual, what you'd think of when you're talking a huddle, was the spelling. It was a smoke and mirrors product, that's what
we call it back then, that they took to Coach Bill Callahan at that time, for the University of Nebraska, and really sold them on it. So for that first year, they were using this new product, or Huddle at that point in time, to test it out with the school. And then in 2007, in June 2007, we officially signed the Huskers as our first client. It took about another year,
until about March of 2008, until we finally got our second client. If you know the football story behind Coach Bill Callahan, he wasn't with the Huskers very long, but then immediately after went to the Jets, and he brought Huddle with him. So our second ever client were the New York Jets. Professional team. Incredible. Right. Incredible. Right. So that was the idea, originally, is Huddle was going to be made to be targeting these college and pro teams and stuff. So for that first two years, we had a lot of interest from high school, small college teams, wanting to use the product. It was one of those things that, up until that point, we just turned them down. But in 2008,
decided we needed to shift, split the product into two sets, the whole Pro Suite, and then Hudl, and start tackling that high school market. And from there, Hudl just started exploding. Incredible. Yeah. By the end of 2008, we got an additional 12 high school clients. In 2009, we got up to 350 in total clients.
Wow. Then by 2010, we were up to 1300 high school. Wow. Wow. That's hockey stick adoption. Was it just purely viral? I guess- Yes. ... it almost
makes sense. You talk about network effects. If two coaches were driving to meet, all it takes is one of those coaches to adopt it and you automatically get the other one, right? Because- Exactly. ... they'd be like, "Wait a second, you don't have to go drive..." What a pain. That could burn a Saturday afternoon, or whatever.
Absolutely. It burns your time. A Monday night, or whatever. Yeah, wow. Especially these high school coaches. They want as much time as they can doing what they love, and driving back and forth is probably not what they love. Yeah, incredible. One of the statistics that I think we put in our poster art for this particular interview, was that you have 97% of US, and I guess we should say American football, right? Right. We are going to get in to talk about the technology. So anybody who's listening and saying, "Wait a second, are you going to talk
about what they built?" Yes, we are. But I think sometimes, as technologists and engineers, just my observation is, we're sometimes almost too quick to dive in to talk about the nuts and the bolts and the protocols and the standards, and without the context of the application, how a product's being used, why something is built a certain way, sometimes it can either lead you to not realize really the full breadth of the significance of why something was built the way it is, or you may be asking, "Well, why are they using that standard, that technology? That doesn't make sense." Well, when you understand the devices, for example, then you're like, "Oh, okay, I get it." So that's just my caveat to say, bear with us here if you're looking to get quickly into the nuts and the bolts, we're going to do that. But you have 97% of American football teams, so that is across high school and colleges, the high school, college pro? What exactly is your addressable market there that you look at? For the American football market, yeah, it is 97% of the high school football teams in the US are all using the Hudl platforms. Amazing. Really, the ones that aren't, are generally when you start getting into the six man football teams- I see.
... like smaller schools. Small schools, yeah. Yeah. Wow, that's incredible. And then, you have schools and customers outside the US? Or are you- Yes. Regionally, where are you? We are actually global. We have a big, big presence in the UK area. We have a big presence over in Sydney, Australia. I hear they also have football over there, but the ball looks a little different. They do. They do. It's a little different,
and they're not carrying it as much. That's right, that's right. Yeah. Hands not so much. Hands off, right? Yeah, that's right. That's right. I guess they also use their heads,
although I think I'd rather use my head to hit their ball than American football. Oh, yeah. No kidding. Using my head to hit another large person. Yeah. Yeah. Okay. We've definitely expanded. We service over 40 different sports right now when we make up Hudl,
and there's over 230,000 different organizations- Incredible. ... and teams. Yeah, we are global- Incredible. ... for sure. Yeah. Yeah. Absolutely amazing. Well, we have talked through the years, and so I've known about Hudl and worked with Hudl for a little while, but until that point, I too was just absolutely shocked at, how is it possible? Being around the industry a little while you can sometimes feel like, oh, I know everybody who's built the significant platforms, and then Hudl pops up and you go, "Wait a second, how have you guys been hiding out?" Yeah, congratulations. Now, you were not a part of the original founding team, right? You joined after- Right. Yep. ... the first platform was built? Okay.
Yep. I joined in 2013, right at the beginning of the year there. Okay. All right. But you're telling me that you were scaling. 2010, you had, I think, 1300 schools. So clearly, 2013, well, let's face that's 10 years ago now. So you've been there through some growth. Why don't you tell us what the platform looked like then? Just
architecturally, how did you build it? Were you building on the cloud? Did you own your data centers? Talk to us architecturally, and then maybe walk us through what it looks like today from an architectural perspective. Yeah, absolutely. Thinking back to that 2010 time range, Hudl, we mainly targeted stop-and-go sports, or clip based sports is what we've referred to them as, or that's kind of like the American football market. That was our bread and butter. Still is our bread and butter, even though we've got so many other products. Back in 2010 through 2012, Hudl was clip based, and I say clip based because most of those sports, it was, like I said, stop-and-go. When you're looking at a football game, it's anywhere from a five-second play to maybe a 40-second play for one of those really long time periods. That's a really good point. I remember you made that comment to me at one time when we were
talking, and I went, "Yeah, that's right." American football, in particular, I guess you care if the running back is repositioning themselves and whatever, but let's face it, what I really care is once the play starts. There's a significant amount of time where the play hasn't started, players on the field and whatever. So that's what you mean, just to be clear, about- Correct. ... the clip based? Yeah.
Yeah. There's a lot of dead time that coaches really don't want to see. They want to see the action. Yeah, that's right. Yeah. Again, they're busy, so they don't need to sit through all of that. Right. So at that point in time, our video came in individual clips. We used WMVs at that point in time. We used WMV because it played
really nice with Microsoft Silverlight, if you remember the Silverlight Technologies? Yeah. Yeah. We used Silverlight back then because it allowed them to build a pretty rich experience where we could actually draw directly on the player canvas. We could do more intelligent clipping and stitching of video files together, which is an integral part to the coaching workflow. They need, at certain point in times, much like if you're watching an ESPN broadcast and they're breaking down the film, they need to telestrate, essentially, on the video play, and draw out where motion should be, where routes need to be. We used Microsoft Silverlight at that point in time with WMV files. We started
shifting into MP4s quickly after that, as MP4 became the streaming standard, and was definitely the streaming standard at that point in time. Yeah, so most of our platform at that time was all online. Actually in 2009, we signed with AWS. We were very early adopters of Amazon Web- 2009. Wow. Wow. You're an OG on AWS as well. That's right. That's right.
Yeah. All of our video files were then hosted. We host our video files on AWS and stream them down through the Hudl.com website. Yeah, at that point in time, Hudl definitely looked very different. It was that bright orange color that you would see almost all over the place and was really designed to be optimized for a football coach's experience. Spreadsheets are almost like a spreadsheet or table looking
area where they could enter in play-by-play data, as what happened, what were results, and then you'd have your video player Chrome. The original experience was all in the browser, and I also was a part of building some demos with Silverlight back in the day and showing some very cool interactive features. It's interesting you bring that up. I can remember, I was a part of a consortium called DECE. We were doing the digital locker in the original, original days of transactional VOD. If you purchased a file on one particular platform, or if you rented it, not only was it only viewable on that platform, so like Voodoo, you couldn't watch it on, well, back then, let's say Apple, but you couldn't watch it on some other competing platform. But
you couldn't even watch it on a different device on Voodoo. So if you originally transacted it, purchased or rented, whatever that transaction was, on your connected television, which way back in the day there were like three, but you couldn't then watch it on your laptop. I can remember some very, very cool demos that Microsoft did with various partners showing just some really cool... basically how you could take the interactive DVD or Blu-ray experience and translate that to streaming, which again, I think a lot of us forget. Even though now, it's a little bit archaic and resolutions are very, very low, but there were some pretty cool Blu-ray and DVD experiences, interactive menus and all kinds of really cool stuff. That's an aside though. You were streaming into browsers. Now, today,
what is the dominant playback mechanism? You must surely support and have very strong support for mobile devices, I'm guessing? Actually, playback is still split between the different devices almost perfectly in half. I just looked up the numbers here- Oh, interesting. ... to get some updates. I would say half is still web traffic, and we have various protocols we stream over, MP4, HLS, for varying platforms. And then, yeah, we do also have our Android and iOS applications where a
lot of our athletes utilize those to watch their video and do some of their review. We're getting into the fan streaming space more and more, so we have a whole host of smart TV apps as well. If someone uses one of your apps, do they get some additional functionality? Or is it just convenience, not needing to use a... and of course some devices, maybe a browser's not the best playback mechanism. I'm thinking connected television, as an example, but-
Yeah, using our apps, and it depends on which app they would use, but yeah, they would get some additional features and availability. For example, our iOS app, we had the idea that anybody has this phone with them at any given time can capture different moments in sports. That's a great capture mechanism for getting video into Hudl. So we do have capture in our iOS and Android applications as well. The only sports that I played were... I was in track. I was pretty competitive in high school and ran a little bit in college, so I can't exactly relate. I never played football, never played basketball, but I can imagine a coach standing on the sidelines now, everybody has an iPhone with them or an Android device. You could whip that out in theory, I'm just assuming this
is the function, and capture 5, 10, 15 seconds of something, and then you upload it, right? Yeah. And then it's available. Yeah. That's interesting. Wow. Okay. Well, a lot of streaming conversations are in the distribution context, and that is the one to many, or the one to a very large number of manys, meaning, certainly for live events, like a Super Bowl, if we're going to stay within the sports conversation. Every year it's like, "Well, how much larger was this year's audience than last year's for streaming?" But it seems to me, based on what you're describing, your use case for the platform, from a platform perspective, is different. Do you run events
where you might be streaming to millions of people a single event? I'm assuming you don't. Yeah, not really. We don't really have those big events. Most of the time we're streaming just directly to teams' feeds. To their fans, to their athletes and coaches. But on a Friday night, for example, especially if we're talking high school sports, there must be a peak time where all of a sudden your platform is pretty crowded, right? Because that's when- Absolutely. ... a lot of schools are playing. There's even a time of the day that you can look at from East Coast to West Coast,
if we're just going to talk about the US. That's a different type of challenge. And believe me, I'm not suggesting one is easier or necessarily harder than the other, because it is a challenge when you have a stream and it's a single stream or a small number, but it's going out to a million people, okay? There's scale and there's challenge in that. But it's different when you might have 10,000 simultaneous streams that individually might just be going to a very small number, but yet your platform perspective, you have to power them all equally, and you have to provide a great service. Can you give us some insights into how your platform is built, and maybe even some surprises that someone might say, "Well, why did you make that choice?" And then you're like, "Oh, well, because I don't have to deliver to a million people, but I do have to be able to run on a Friday night," whatever, "X number of hundreds or thousands of streams reliably without question." So yeah, give us some insights of... Yeah. Absolutely. Yeah. You hit the nail on the
head there earlier when you were talking about we are built definitely different. Most of the time our content is coming from teams and coaches and is targeting their athletes, their fans. So we will have a recording, these two hour long events, they may have a couple hundred views on them. So we're not talking big numbers, which definitely, when you're talking about the distribution side, the number one thing you might immediately think is, yeah, our caching strategy is very different. Where a lot of these sites, they may hit for really high cache rates, I'm very happy with a 40 to 50% cache hit rate. That's actually fantastic performance, in that situation.
But Disney streaming would be like, "Oh no, we have a serious problem here." Exactly. Exactly. So the things that we have to optimize for in that situation are going to be regional optimizations on trying to pull content down to the users. We look at things like pre-caching content, bringing content, either when we're talking our HLS streams, bringing fragment by fragment content down to the users, just as they're watching the stream, we know what the next chunk of video they're likely to watch is going to be. We try and get that pulled into these PoPs in
the regions where they are watching these videos. So pre-filling of the cache is a big thing for us. Also, of a diversity. I know this is important across multiple organizations, whether or not they're getting millions of views on a single stream or not, but also the diversity of PoPs available in diversity of CDNs. We run a multi CDN environment here, and that's because we don't usually have isolated chunks watching the same stream across multiple areas. You might have isolated chunks of users watching our content, but then across the state, someone's not watching that content. We need to- Yeah, understand.
Yeah. We have to target, essentially, our users for what's going to be the best performing CDN potentially for their region, for their exact ISP and where they're accessing from. So how we target our users, what CDN we provide to them, is slightly different than what I would say you might see out of a YouTube or a Disney, or something like that. So what I hear you saying, and I don't want to put words in your mouth, so correct me, but almost always when you're talking about these really large scale events, where it's streaming a single stream to millions, it's about cost reduction. So very often the CDN decisioning gets down to... and it is splitting fractions of pennies, because it's meaningful when you're delivering to millions.
But I didn't hear you mention... obviously you care about cost. I'm not suggesting you don't, but is it true that your decision is not strictly necessarily based on cost, and it's performance, as you stated? Or it's even just purely about coverage? Like if one CDN doesn't have coverage in a market, well, you need someone who does, though? Absolutely. Yeah. I would say, yeah, cost is always going to be important. You're right. But- Of course, yeah. I'm not implying that it's not, but- No. Absolutely. But yeah, mostly it's just the availability of the network and the performance. Yeah. Interesting. It is the more important thing. If a coach can't access their content because of a
couple regional PoPs, or one CDN are out, then you can be assured that our support team is going to have their phones ringing. Going to get a trouble ticket filed. What is the role of quality with this content? I'm asking the question because really, I don't know. Is this one of those shocking things where coaches care about quality even more than... or is there a different bar of acceptable quality, therefore you can encode at maybe lower bit rates? Are there some sports that are more perceptive to quality than others? How do you approach that? Quality is really important in the sense that there are a few things that we have to get absolutely right in encoding our videos. That's going to be the field details and jersey details. One of the things that you don't really- Oh, good point. ... think when a coach is breaking down content, you'd like to think that... A coach knows the names of their players. Obviously they know the names of their players, but when you're watching a video and you may be getting a video shot from a press box on a football field, everyone's wearing a helmet, all you're seeing is jersey numbers. You need to be able to quickly identify who did
what and where. Clarity on numbers, and one of the things we see is with a lot of those lower bit rates with some encoders out there on the market, some of those finer details, the jersey lines and stuff like that. Also with hair, you get people with longer hair, it blocks jersey numbers, and the encoder creates these artifacts or unnecessarily blends together jersey numbers, and features just together, and it makes it more difficult for coaches to see. So, yeah, we do generally have to provide a slightly higher bit rate or use an encoder that is going to be optimized for providing those clear lines, and also bolder colors. One of the things
that's really easy when you're talking about an outdoor game, is getting that color blending. Some encoders work really well at keeping your colors bold, not desaturating content too much. Other encoders may not do as great a job. To take this back, quality plays just a really big part of the sports video experience, because when we're watching content, we use the identifying characteristics, like jersey number, field lines, to help tell the story of what's happening on the field. Interesting. Right. So those- Yeah. Makes sense.
... things have to be clear and easy to read. Yeah. Yeah. Yeah, that's interesting. That completely makes sense. So what is the architecture? What I hear you saying is, is that a large percentage, if not nearly all of the source material, you're actually not generating. Are you putting cameras out in the schools, or is it on the school and the coach and the players to- We started off, it was on the schools and the players. We have, over the past several years, been developing an in-house camera. I've actually got-
Makes sense. ... my development rig of it right here. It's this- Hey, cool. Cool. Let's see that- You can't really see... There you go. There we go. There we go. Okay. Yeah, there we go. Yeah. That's our- Cool. Cool. ... Hudl Focus camera. Basically, it's a stationary camera that gets set up in gyms, gets set up on football fields in different areas that will record. Uses a collection of sensors,
maps them all together in a 3D environment, does player tracking and other AI to help track the ball. And we create smart tactical feeds of video that is uploaded directly to Hudl that way. Interesting. Okay. But overall, video can come from a coach from their iPhone, from an expensive camera. A lot of these Texas schools, they have nice expensive rigs and stuff they love using- They love their high school football in Texas, don't they? Yes, they do. By the way, I grew up in Colorado, and I grew up when Nebraska and CU were the absolute powerhouse teams, so yeah. Yeah. So I know a little bit, even though I didn't play the sport, I know the culture. I can relate to it. Yeah, it actually is really amazing. If someone's not been
exposed to it, just go to a high school football game in Texas, or back in the day in Colorado on a Friday night, and it's quite an experience. It is. They take it serious. It's a way of life for them. Yeah. Yeah, totally. Okay. So you have content coming in from a lot of different sources, many of which you don't control, meaning that it could be very high quality, very well shot, great lenses, great cameras, all that stuff, and others, it's a phone and a shaky one at best.
Yeah. So then, are you doing some processing in the cloud? Do you transcode that? What do you do with that video once it's on the platform, and do you have a decision tree? Are you using some sort of, even, I've got to get it in there, AI, to maybe do some sharpening or some processing or something, to improve quality? For most video, the way that we've operated for so long is we've taken the single encoding ladder approach. We didn't have a smarter way to provide more bits to a certain stream. It's just historically how we've done it. And some of that is around cost management. When you've got all these streams, we've got 100 petabytes of video out on S3 in our video library right now, it's a lot of video. It's a lot. And in order to keep our costs down, but also provide the best possible experience, we generally just test to find what is going to be the optimal bit rate in which 95% of our users have top quality for the highest renditions that we have. And that's historically how we have done
things. Now, we are looking at exploring a little deeper, maybe providing different experiences and stuff. We're exploring, obviously, some of these new codecs that are out here, like HEVC, AV1, do a great job at providing that better quality video off the bat at those lower bit rates that can- Yeah, lower bit rates. Sure.
Moving forward, I think we're looking to optimize, possibly on a per sport level, which- Interesting. ... we talked a little bit earlier. Yeah, there are some different requirements for different sports that might just have a worse experience. And thinking, water polo is a big one. Like most people in the video world, you've got the moving water. There's so much motion. Yeah, water is super challenging to encode. All those pixels are in motion, continuously. There is no step, by very definition, with water.
Right. I know one of the things that we've had to do in special situations like that, is we have to slow down the encoder significantly. In those situations, you have to pay for a little additional compute time on that, and then, yeah, you have to up the bit rate on those streams to provide a much better experience. Are you standardizing on 1080p for what the coaches and the players are looking at? Is there a need to deliver 4K? Yep. It depends on the sport and the level there. The NFL, they still run on 720p. Isn't that amazing? It is amazing. However,
our coaches, most of them, they want 1080p. Our football, we are delivering 1080p. Yeah. I have to ask a question here, and you can answer it as sufficiently general as possible to not give away any secrets or inside... but it still is shocking to me how broadcasts, and you mentioned the NFL, and not singling out the NFL because it's true, major leagues go around that are running on... sometimes there's even MPEG-2 happening. Now, I'm talking more in the broadcast distribution chain, but my point is, is where there's literally two, three, sometimes four generation old technology that they're still using today. 720p or interlaced. It is shocking. People just don't know how much interlaced is still happening even in live TV production. Now, interlaced, that's largely, I think, gone,
but three, four, even five years ago, there was still a tremendous amount of interlaced content that had to be then deinterlaced. Meanwhile, everything's being streamed at 720p, or 1080p, or 4K, all progressive, and yet the production chains are so old. Why do you think that is? Well, yeah, it's funny you mention that. I'm asking just a general question. Why is that? And 720p at the NFL. Why is it that a high school team in Texas can have better quality than the professional league? Yeah. Real quick. It's funny you do mention
that. Over 50% of our video we get in, we do have to deinterlace as well. So yeah, it's still out there for sure. Wow, amazing. I can tell you that generally, when you're talking about these bigger programs, I'll speak specifically to the NFL, there's just the concern around security. Security, it's [inaudible 00:38:37]- Oh, interesting. ... just more worried about- So their view is the lower the resolution, the less useful the video is? Is that- Well, not necessarily. See a lot of them, they actually still record to SD cards and they have runners that will run video content from the sidelines up to the press box and stuff. Some of them are under the belief that going to 1080p,
they would fill up those significantly. A lot of these hardware devices that they use to record to SD cards only support 720 feeds. Crazy. Yeah. They're- So it's like literal hardware limitation that could be as simple as, I have a file here, everything will play it back except for this one device, and therefore I have to reduce my experience because of this one single device in this somewhat complicated distribution chain? I have to say I'm not surprised by this, because again, getting back to the interlacing issue and other... in my past dealings with broadcast, whenever I've asked... like around 4K, that's where I really became exposed. "Why
is it that 4K is so hard to adopt?" And everybody says, "Well, it's the cost." It actually isn't the distrib... Yes, there's a distribution cost associated, but it's the cost of supporting throughout the entire pipeline. Correct. And it can be as simple as my graphics... box. I'm blanking out. They have a name for it. But that it doesn't support 4K, and there we have to go through all kinds of gyration. We have to basically convert it from 4K to 1080p, and then we upscale and it's like, "Well, now we lost it, so we just run 1080p all the way through." And you're like, "You mean just because of
the graphics overlay device, that hardware, the consumer's missing out on this beautiful experience?" Like, "Yeah, that's it." Wow. Yeah. And that was the other reason, is cost. Cost is the big thing there, especially jumping to 4K. When you're talking 4K, a lot of these boxes that these coaches have in their office that they're watching on, they're not going to support the latest HEVC. They're going to be 1080p televisions and-
Yeah, it's H.264, and yeah, it's an old TV. Yeah, interesting. So you mentioned AV1 and HEVC, and it's certainly of interest, and we've had conversations about those codecs. Platforms have different strategies for adopting, shall we say, next-gen codecs. One strategy can be, "We're just always going to use the latest and greatest." And then there are the more conservative platforms, are like, "Yeah, we're not necessarily going to be the first to adopt." Where are you in this
continuum? And then tell us why. Because, again, I think the why is important. It's one thing to say, "Well, we're pretty conservative," but then, is that because people are on devices that just simply don't support AV1, as an example? Yeah, but tell us where you are. Yeah, absolutely. Yep, you're right. The simple answer is yes, we're going to be a little more on the conservative side of this. Why that generally is, is when you look at our user base, our power users are going to be coaches, and you have a big mixture of the types of coaches and the technologies that they're comfortable with. When you look at our user base, where they're
running on, whether or not it's personal devices or even school devices, keep in mind we're looking at high schools and stuff, and a lot of public school systems aren't going to have access to the latest and greatest technologies out there. Yeah. Yeah. Yeah. Yeah, that's true. Generally we got to look at what devices are our users accessing our content, for what platform they're accessing as well. I think when you start talking more of our fan content, we get to branch out beyond the high school environment and start looking at just device share around the world. Yeah. When you start, once again, getting into our core platform, yeah, we're dealing mostly with school computers and stuff like that. Yeah. Yeah. Interesting. Yeah. Generally we are, like I said,
more conservative on that. We are getting into the HEVC world, given the Chrome announcement here last year about supporting HEVC playback- That was good for you, I'm sure. That's really big for us. Since the bulk of our users, about 98% of our web traffic is going to be either on Chrome or on Safari, both of which will support HEVC if the hardware capabilities are there. That's right. That's right. Currently, we're sitting at about... I double checked it again last week,
we're just about 71 to 72%, right in there, of our users have that hardware capability- Amazing. ... to decode HEVC. Amazing. Wow. Wow. How do you know that? Yeah. What we are doing, we ran a couple tests. So really fun thing you can do for anybody wanting to [inaudible 00:44:39]. There are a couple HTML5 video APIs you can say for, can play video codecs. I have to look exactly what it is,
but essentially, by specifying your codec tag, you can determine if the system can properly build the decoding chain, connect the pins, can decode it and stuff. So that is based on whether or not Chrome or Safari tells us that they absolutely do have support, or is most likely supporting this stuff. So definitely support- Got it. Got it. ... gives us approximately about 65%, and then the additional, there's about 10 to 15% there on top of that, that are in the some support. We assume that about half of those actually can support.
Actually can, yeah. Yeah. You mentioned that you have a custom app across certain platforms. Do you put a software decoder in there, or would you contemplate that? Or are you 100% dependent on just the hardware support on the device? Right now, it's going to be hardware, entirely dependent on that. And some of these newer codecs, we've looked at potentially even LCEVC- Like AV1, for example. ... or AV1. Yeah. Oh, LCEVC. Yeah. Right. And putting those in there. And just for our users and the time they're spending on our app, we don't see as good of a battery life. That's going to be your big thing that you see when you're putting in a software decoder on these newer products, is battery life really takes a hit and stuff. It hasn't been worth it to us, especially when you're weighing at the development
cost to put something like that in, with the, all right, if the majority of our users actually need this for playback, does it even make sense for us to encode a quality of that to them? So generally, when we are looking at new qualities, new codecs to use, it's dependence on hardware support, or a codec that is not going to sacrifice battery life significantly in mobile devices, are what we're going to have to look at. I know some of the... like dav1d for AV1, is really, really power performant, which is to say that it actually is amazing what it can do and how little energy consumption it needs. At the same time, there's always that trade-off, right? You still have to push the user even to upgrade your apps so you can build all this, and then if an insufficient number of users actually upgrade their app on the device, then they don't even have access. It's a lot to push upgrades even. Yeah. I would say even then though, we do have really
good adoption rate of our new app versions when they come out. We're generally looking at about- So, people are... Yeah. 15 day to a month to the majority of our users getting updated apps. And the majority is what? 90% of the base? Correct. Yeah. Yeah. Yeah. So, [inaudible 00:48:12]- Which is good. That's really good. AV1 has been something we've actively been looking at since Apple announced that there was AV1 support with the HLS spec. Yeah. We didn't talk about DRM. Is DRM required
for any of your content or on the platform? Not currently, no. It's something that we haven't heard. We're waiting to hear back from our coaches on whether or not that is actually required. I would say it's not likely going to be something that we will ever need. Yeah. Yeah. Yeah. Understand. What's coming for Hudl in terms of the platform? Are there any neat things that you're focused on, or even you're just dreaming about building that you can talk about? Yeah. Absolutely. Again, two big things that I'm
really excited about that we're working on. The first, as just coming hot off our conversation here with HEVC, we are looking to upgrade our platform to provide HEVC for our coaches. Amazing. We're really excited that we're going to be able to deliver a visually lossless quality to them, which is really big for us. The other one, I would say, really excited about this technology, is what we call our on-demand rendering system. This is something that we are currently fully utilizing for all of our secondary qualities in the US. So all US-based accessed video is currently using this in production, but it is
the ability to encode the secondary renditions of our ABR ladder, so our lower qualities, on the fly. So this is something where we're able to cut down our storage footprint in half, which, when you're talking about the sheer size and the amount of content we have- Yeah. That's very meaningful. Yeah. When you think about it, in a lot of situations I would say that the math wouldn't work out and you'd be spending significantly more. Our access patterns are that this coaching video is generally going to be accessed over the first 14 days of its life. Then it will just sit there, and the chances of it being accessed again are very small. So if you can deliver those- What is your SLA? How long do you have to
keep that video on the system? Yeah. We used to say it was indefinite. It's going to be based on the coach's subscription, their package- I see. I see. ... how many hours of video. We got to keep that video around as long as the coach keeps it in their account.
I see. I see. Okay. So is this, I buy a bundle of hours, let's say, and then I'm about to run out, and then I as a coach have to decide, I really don't need those, and I could, whatever, choose for deletion or something, or just delete them, I guess- Right. ... on video? Yeah. Got it. But yeah, it's an interesting problem, and it's interesting that you're pointing out that it's now becoming possible to trade compute cycles, or I'm sorry, trade storage for compute cycles, meaning that it's actually more cost-effective to transcode on the fly and not store those additional renditions, than it is to just have them prerecorded and stored. I've heard pretty consistently, and of course, obviously we provide methods to do that very, very cost effectively. So for a lot of our customers, they've done that math and it makes sense in almost every context to do it that way. So yeah. Yeah, storage cost, even though storage cost is being driven... it's more benefited from Moore's law. Storage is still cheap, right?
Yeah. Storage is cheap until you're talking hundreds and hundreds and hundreds of petabytes. And when you get up to a certain scale, there was- It's no longer cheap. Yeah. Yeah. Yeah. It becomes meaningful. Yeah. Casey, I really want to thank you for the conversation. It was wonderful. One of the things that we work really hard to do on Voices of Video is to have meaningful conversations just like this one. You shared a lot of great insights
and can't wait to have you back. Maybe after you get HEVC deployed a little more fully, we can have you come back and talk about lessons learned and give us an update. Yeah, and thanks for having me. It was a great conversation.
2023-11-20 10:01
