The VideoVerse

TVV Ep 11 - Without challenges, we don't have anything to do.

January 27, 2023 Visionular Season 1 Episode 11
The VideoVerse
TVV Ep 11 - Without challenges, we don't have anything to do.
Show Notes Transcript Chapter Markers

In this episode of the Video Verse, CP (CTO of Mola) joins us to discuss how technology is impacting user engagement on platforms. From understanding metaverse and merging lean back and lean forward experiences together; to leveraging content for increased engagement - CP has a lot of wisdom to share with us! Tune in now for an insightful conversation about tech, users and everything in between.

Watch the full video version.

Learn more about Visionular and get more information on AV1.

[Announcer] Welcome to "The Video Verse."

Nathan: Hey everyone, and welcome to "The Video Verse." I am joined today by my esteemed colleague, Zoe Liu. Sorry, mispronouncing your name already. And David Lee, and we're joined here with CP. I'm really excited to have you here on the show, CP. Can you just quickly introduce yourself, tell us who you are, and what you are currently doing at Mola?

CP: Yeah, sure, Nathan. It's actually a pleasure to be here on the program. So as mentioned, my name is CP. I'm actually the CTO of Mola, one of the home brew OTT platform provider coming up from domestic Jakarta, Indonesia. Over the past year, we have actually successfully ventured into the rest of the Asia, like Singapore and Malaysia. At the same time, we are establishing our presence in UK, as well as Italy.

Nathan: Wow. So you guys are all spreading out through Europe as well then?

CP: Yes.

Nathan: Awesome. That's really cool. So before Mola, you came from HBO, and I want to talk about that, but before that, let's go back even further, and can you tell us how you got started in the world of video codec and encoding? Kind of what was your first encounter with codec? How did you get into it?

[00:00:33  CP's experience with codec]

CP: Oh, that's a long, long time ago. So my first, I was really, really dropped in, actually they pushed me into the swimming pool. And I was actually asked to spearhead for South Asia, the so-called professional post-production world. Because Apple at the time was introducing Final Cut Pro into the market.

Yes, and we do not have any expertise on the ground. I actually came from an engineering background, no video. No post-production. So that's why I really was pushed into the swimming pool and forced to swim.

Nathan: Yeah, no kidding.

CP: Without any lifeboat. So actually that is the time where I started looking at just video, per se. And codec actually comes up as a form, supplementary from the angle because as an engineer, you don't just do the application and click here, click there. You really wants to know what's happening. So I spent time to really understand the different posts codec there is, and what makes the world works from production to distribution as well.

Nathan: Okay. So what codecs would that have been in those early days for Final Cut? Was it ProRes at the time or was that before ProRes?

CP: Yes, so we actually started, we actually started with ProRes as the base for the post at this site.

Of course, yeah. So we have the ProRes HD, we have the ProRes HQ, and then later on, we have the XQ and so forth.

Nathan: Of course. So you got thrown in from the production side, but then distribution was a natural progression from there, I guess? If you edit your video, now I wanna distribute it?

CP: Correct. So from the production angle, we then have two kind of distribution. We have actually what we call original distribution where it actually allows the video to be sent in a high quality to our partners for further editing. And then we have the really streaming, not really streaming, in the distribution media mechanism, the impact force, impact tools that is meant for really simple lightweight delivery.

Nathan: Okay. And so this is all, I don't wanna say uncompressed 'cause it's all been compressed from ProRes, but uncompressed compared to like, H.264 delivery over the web, is that right?

CP: Definitely, yes.

Nathan: Okay. Wow, that's really interesting. And it's another world of codec and encoding is all of that production side of things, which is a world I live in. I edit in Final Cut, and I use ProRes all the time, but then I have to deliver and have to choose my delivery format. Talking about moving from HBO, your experience there, over to Mola, was there a big difference technology-wise? Was a different market, or was the market different, I should say? Now we're talking about the delivery of it from HBO versus a Southeast Asia focused audience. Was there a big difference between the two that you had to address?

CP: I think in a certain way there is, but in general, the directions, or rather, the final destination is almost the same. In HBO coming from the cable world, one of my key tasks was actually to look at the digital distribution as well. And that includes HBO Go streaming platform for Asia. So from that angle, the focus into the OTT world has becomes predominantly strong as focus for the organization. In Mola, of course, starting from a digital native world, from the OTT platform initiative, right now, we are also looking moving back into the linear cable world. So to me, it's a very interesting mix because it is like I'm driving on the left hand side or driving on the right hand side and trying to balance both, right? I came from the left hand drive environment, moving towards the right hand drive. Right now I'm in the right hand drive environment trying to rebuild the left hand drive for this organization.

Nathan: Wow, so-

Zoe: It's a very interesting analogy about your experience.

Nathan: That's a great analogy, yeah. So you started out with those production formats, you moved towards the distribution for OTT, and now they're asking you for the cable, everything's going digital as well, is that right? Is that what's driving you?

CP: Yes, yes. Correct, I think the key difference will be, in this case, for the new world, even going back into the linear space, it is really digital driven as compared to before, but it's really a little bit more RF and satellite.

Nathan:  Okay. And if that's all digital, what codecs are they using?

CP: Oh, actually, in terms of the code, it is still something that is, I will say that there are many ways to do it. For distribution, in terms of streaming, we usually can be also using things like . That's one format that we use is also Zig C that allow us to actually deliver our final production as in final stream to affiliates. And there's also things LTTP, yeah. That's actually developed by some other organizations. That gives us efficiency as well. I think for this angle, it is still evolving because predominantly in the past for this huge base distribution, it has always been using the satellite for its actually multi-class mechanism. But as research IP, there is this, I would say, bottleneck in the IP world when needs to comes to a wide distribution point that the satellite can achieve much more efficiently. And, but I think the technology is picking up with more vendors coming into the space, learning the efficiency of how the satellite are using and then moving that into the IP world while maintaining the, I would say the SLA as well as a response cycle for the distribution itself.

Nathan:  Very interesting. So, at Mola then, I know that you actually did a lot of engineering work yourself on the platform. You're CTO, but you've been hands-on with the technology, is that right?

[00:07:49 Various tech stack and build modular]

CP: It depends how you define hands-on. If you're saying that I go in the coding, nah, I can't do that anymore. I can code to save my skin if I need to. Unfortunately, I don't need to save it, so. But what I do is that my position actually as CTO in Mola is really to look at the various tech stack. And for me, tech stack, it is not a singular big monolithic design. It has to be modular, it has to be flexible. So from a technology standpoint, even back from the HBO days when we started with HBO Go, we are always looking at specific component and how to maximize that particular module, for example. At same time, integrated with the REST modules for the whole tech stack tool. I think that is always the piece that is the most challenging, but is the most satisfying when you get to the end of the road. I will say that almost everywhere is challenging, right? Because if you look from a platform perspective, right from subscription, right from video delivery, even things like the CMS of the control plane, the player, the VRM, and so forth, everyone of them has its own set of challenge. And the available is so huge because there are so many ways to do that one single part. And it also impacts on the upstream as well as the downstream of things. So the challenge of a technology team is also always looking at all these pieces and ensuring that the handover from one module to the other is done at it's best ability. At the same time, not causing any full dependency. The whole reason we build modular is so that we can actually change parts easily if there's a newer and more efficient way to do things. So if you design with different modules so tightly coupled, then we have a problem when times comes to change because when you change one module, you have to affect so other many modules. For a tech team, it becomes very stressful. And we know when you are stressed, you don't produce good work. I think that's always the key.

Zoe: So basically, I think you really look for kind of independent modules that could be potentially easy to be, relatively easy to be replaced.

CP: I wouldn't say the word replaced because it's not a very good word to use. I would say, I would say in terms, I always look at technology that is other parts, it's easy for us to evolve our tech stack. Because I think very importantly from a technology standpoint is that we always want to deliver better and better over time. So whatever we look at, we must have the opportunity for that component to be optimized, okay? To be enhanced over time.

Zoe: Yeah, to be upgraded, for example. Right. Yeah, so then it's just because, I think there's several goals you're trying to get, right? Because you want them to be upgraded. So relatively, you want each module to be relatively independent of each other, but you also look for the overall performance optimization so that they can work each other as seamless as possible.

And on the other side, you also, previously when we met, you mostly mentioned you want also the whole system to be able to scalable. So there's several things coming together. They got to be conflict against each other. So how you handle that kind of things?

CP: Yeah, that's the fun part of tech, right? If there's no conflict, we don't have things to do. I always say. Yeah. The technology side of things is always re-looking over and over again. The job is never done. Today, we can actually launch a new platform, running fine, scalable, but tomorrow, we always find, hey, the small little part that we think we might have overlook it, and then we reoptimize it again. And when you look at all this optimization point, you'll find more and more it is just like a string, a thread from a shirt. The more you pull, the more you actually get from it as well. I think that for me is always the most fun part of things is always really looking. One thing about the tech group that I run is to always tell people our work is never done. If we think that the work is done, that means we are not doing things correctly. Our work is milestone driven. We have completed one milestone, but it always important for us to re-look at what we are doing to find out how we can even do it better.

Nathan: So I was gonna say, I love your optimistic outlook. That if there isn't challenges, we don't have any work to do. I love that. That's a t-shirt right there.

Zoe: Yeah, I just wonder-You can give us some example. Like, for example, some modules, you really want to optimize it, but then sometimes you may have to let this module to at least work with another module. Sometimes you have to compromise something, right?

CP: Yeah. I think, yeah, it's true. For example, if you look at, just from the OTT perspective, from the subscription to the playback, rather to the encryption of VRM, all these modules are literally independent in a specific way, even the end coding and the transcoding. But if you look at the whole flow of chain, even from the upstream before it hits the platform to a media supply chain, there are always the intracity of the connectivity and is dependency.

Like codec, for example, coming from a codec as heavy as big as ProRes, it doesn't make sense for that to get into our OTT platform directly, and then do the encoding into the streaming format. So what we did was to introduce a distribution master format that allow us to actually have smaller, lighter, and more efficient way to distribute this into the platform. While we host our original masters in the archive for any further edits for any promo works that needs to be done, right? And that's how the whole handshake between the different modules needs to be looked at, not just with the module itself, but with the whole encompassing components that is supporting it. And this allow us to then streamline certain flows, right?

On the other hand, look at just player, for example. There are so many native players on the market, there are also commercial players. So the decisions whether to use a single commercial player, okay, or to use multiple native player is another tricky as well as challenging things to look at. It also depends on how the organization wants to go forward. So what Mola has been doing previously was to really branch out into native player because of its ease of control in individual component, right?

But right now, we are doing real fundamentals in the new platform, and for us to work more efficiently, I think the key would be efficiency. If we were to move a pure native player, we will find that every single team needs to have its own set of issues and resolutions. So we decide to adopt a singular commercial player that we can work with so that we come back to a baseline.

But if you look at having a baseline concept, it's not necessarily the best as well because what it means is that you cannot fully utilize every single platform's uniqueness. IOS, Android, and Ryzen, all these players will then becomes a baseline that you will spread across, and you cannot really enhance the user experience based on the platform that you're running.

So this is always the, I would say the chicken eggs part of it, and the dilemma between any technology decision that we made. It really depends on what is the end game? So for us, the end game is about user experience and stability. So coming to a fundamental allow us to build the stability and the consistent user experience to every one of our users.

Yeah, and I think that is one of the key challenge for any OTT platform provider like us. Number one, from the regional perspective, different countries has different capabilities as well as segment of users in terms of their endpoints experience, right? Just from the handset perspective, countries like Singapore, Malaysia, which are more developed, you'll find that it tends to have the bigger players where have the handsets that's more powerful that can deliver more experience. But countries like Indonesia, Vietnam, there comes more anger where you have slightly, for short of a better word, cheaper handsets that's not as powerful, right?

I think this presents a very strong challenge to any player like us because how can we keep the experience consistent to the users? And to be extremely frank, there is no way you can have a full consistent experience. It's about finding the best fit for that user in their country, right? For example, in a high bandwidth environment like Singapore, the experience will really be towards super high quality audio video experience, as well as interaction to the users.

But in a very mobile driven country like Indonesia, then it might be actually towards efficiency of delivering the stream so that they get to see what they're supposed to be seeing. A very interesting example that I always put is that, because football used to be, or soccer used to be one of the key component within the Mola platform. If you look at any matches, a lot of people always just look at the players, and maybe the ball itself, whether it gets into the net, right? But if you look at how codecs have been engineered and how it actually has it been evolved, look at the grass field, right?

 In the original days, with the quality science VGA component, literally you don't see any patches, or rather, you rarely see patches instead of the leaves of the grass. But today with good quality codec, you can really see the grass field nicely, including the lines painted on the glass itself. I think those are the elements that, from a provider perspective, to me, it is important we can get to that state on the devices. But again, due to the limitation of the bandwidth in the in-country as well as the handset, we might need to moderate that.

So some users will get the super crisp, clear quality of video. Some users will get a slightly, not as clear, but most important for football is they can see the ball. That's the most important as well as the jersey number so you know which player is actually kicking the ball. But it's also about the latency as well as the streaming quality. So you don't have a player suddenly appearing from the left of the field and it jumps the right. You really can see the player running across.

So the challenge in terms of delivery is always how to maintain that kind of smoothness versus the kind of clarity. There is always a balance point that we need to compromise depending on the environment, depending on the handset of the users as well.

Nathan:  CP, can you share with us some of the latest technologies that you guys are exploring that are gonna help you do this dance between clarity and smoothness and any of the other variables that you guys are using?

[00:20:03 Latest technologies are exploring]

Sure. I think from a technology angle, we are doing quite a lot of things. When we first started in the organization, the first thing for us was to re-look at the whole delivery environment or the delivery network, right? There is no equal provider of the CDNs in the world. We have the big players, we have the boutique players, and we have home brew players. I think everybody has a part to play in this world depending on how we want to actually get this to the end user. And this is where from the technology team in Mola, we are very focused on making sure that we have all the ways to deliver the extreme to the end users at the best rate as well as the least latency.

So we work with almost every one of the CDM providers that we know of to ensure that we have that route, you know, to the users. Then the challenge come is how do we determine which route to take, right? It's just like we are. I mean, David, Zoe, we are world travelers. For us to travel from Singapore to San Francisco, I have 101 airlines to take. So which one is the most efficient for me, right? The same thing for CDN. And we do have to have all those things as well. And this is even before looking at the business metrics. We're just talking about technology metrics.

And that's why things are selected. Things like logic as well as design comes into play to ensure that we have that route taken care of. Then the next one is the video itself. How is it different rendition, okay? Because for different countries, as I mentioned, different countries, different handset, they have different kind of capability. So the renditions of the profiles that we actually send to the devices also plays an important part. Of course a very basic profile is you just have that six different rendition and that offer to every devices, right?

It is a very base way to do things. It works, but it's just not give you the best experience on the handset that your users might have. So that's from us to also move forward is how to discover as well as how to determine what of rendition to be sent to what devices. So these are the things that Mola has been working on to make sure that, ultimately, when the users gets into our platform from whichever endpoint, be it the mobile, be it setup box, be it on the web, our systems will try to deliver the stream to them using the kind of technology that their endpoints can best tap on. And then we look at the delivery route to them. So that we actually has the delivery end to end being taken care of.

So is that, obviously first you're gathering data on all the different devices and the path so that you know which airlines are available, right? Is that AI running in the background making all of these decisions? I guess, millions of decisions I would imagine. Or how is that, is that third party services? Is it your own AI?

CP: This is the part that, for me personally, I have never liked to use the word AI because it has been overused, right? So if you look at the base of AI as artificial intelligence, it means that it just, it makes logic decision based on data. That to me is AI. No matter how people are beautifying it, making it sexy, you know, kind of things. But to me, AI, the base is that it's really giving the correct data what it needs to do, right?

In that case, yes, that is what we are doing. We are collecting every single data that we can from our endpoints so that we know the route to the endpoint, And then based on that, we set up different logics so that we can then determine what to do for the particular endpoint based on the country, based on the user's network. And that over time gives us a better delivery to the end user.

Zoe: So how do you collect the data from the, for example, from the receiver where outside from the player side? How, so what kind of a tools?

CP: So today, okay, when I joined Mola on board, that was one of the key challenge that I had because I don't have visibility on actually what is out on the market from our end user handset. And that's a big challenge. And without those information, that's really literally, you are really firing the duck, right? So the first thing that I actually went to do was to work with, okay, should I mention company names here? This is the question.

Nathan: If you're comfortable.

CP: It's okay, right? Yeah, for me, just I need to know because I've been doing a lot of all these things, I just need to be curious. Sometime they do not ask to mention organization in that sense. Otherwise, they say the organization is that, once they mentioned that name, they will go and find that company and say, hey, can you sponsor?

Okay, . Yeah, so when I started joining Mola, one of the key challenges actually about making sure that we know how is our users experiencing our stream. And that part is critical to me from any technology angle as well as in the future. And in marketing recommendations, all those datas are critical.

So the first thing that I actually went to do was I went back to a well-known, dependable partner called Conviva that I have been very familiar with. So with assistance of Conviva, we build up a path that I can then collect all the endpoint information from the stream perspective. And with this information, it helps me to make the correct decision of how my CDN needs to be retrieved, how my players needs to be built, how is the codecs of the streams that I'm looking at, how efficient it is to reach my end users in the different part of the world?

So if I'm starting one, as we move along, this technical data helps me to realtime decision making using the system. As we start to see CDN's performance dropping, we can immediately switch, okay, the delivery path to a much more non congested CDN, for example. At the same time, this data and collection allows the team to look at where is the bottleneck that we need to fix from a technology angle so that we can actually eliminate that right from source as well?

So datas like this are really critical to our operational as well as our roadmap planning. I think after we have implemented Conviva, the key thing that actually help us to reemphasize as well as to reiterate the concept of the differences between the different countries, right? For example, if look at Indonesia, per se, majority of the streams are really mobile, which means that we're looking at anything maxing out maybe about one Mac, okay?

Whereas from these high quality streams that goes onto broadband, the number is significantly lower. Of course that, again, depends on the time of the day, but I'm talking in general over 24 hours, right? The mobile space innovation market is so much bigger as in compared to the broadband side of things where you have the Android TVs and so forth. But over the year, we also see that changing. I think as a country mature, as a market mature, these statistics will start to change. So if you compare this with Singapore, you find a majority actually are on the last screen setup where we then deliver highest bit rate streams, okay? As huge as five to 10 meg where it's needed so that you actually fulfill the 60 inch TV running on the smart OS.

So this is the kinds that also then help us to, or rather, reminded us that not everyone is equal. If I deliver a stream that is suitable for handset, you'll find that that stream does not do justice to the Android TV version running on the 65 inch TV set. But if I were trying to fully push the quality to the 65 inch TV set, I will face problem with the person running on a typical Android mobile on earth 4G network. But I think over time with technology, like now with 5G coming in place, all this, things starts to change again. You'll find that we can then deliver much more high quality stream to the same user even with the same handset.

I think there's always these three parts. One, the handset itself, which is what the users are holding, okay? And second is what we are actually delivering and capable to deliver of. But the most important is actually the route between these two, right? The network itself. And these are the times that, in different environment, different countries, there are different stage of maturity, but everybody is due to catch up or leap from each other depending on which state they are in today.

So when we do from a technology standpoint, the easy way is always do what is needed to satisfy today, right? The hard part is do what is required to make what is needed today. At the same time, build enough flexibility for the systems to evolve. And that's why, for me, it's important doing the second part and doing it correctly, doing it well because that gives you much more stability as well as the longevity in the business.

Nathan: So CP, as we wrap things up here, kind of moving, looking forward, we've talked a lot about, you're always very thoughtfully developing and designing a tech stack, we'll use that as kind of the broad term, to, like you just said, do what's needed, but also do it's required, but also thinking about the future. Let's say either for Mola or for OTT delivery in general, what does the future hold, do you think? What's next that you guys are looking at down the road when it comes to video delivery? Is it the codec? Is it the path? Is it all of the above? What's your thought? What are you looking forward to in the future?

[00:30:59 End user experience is the most important element ]

I think from a technology standpoint, it's all of the above. But as any business goes from OTT delivery standpoint, I always says that it is the end user experience that is the most important element because there's always two ways to look at it. One is you look from the base, right? It's like you're building a Lego set. You get the base, you get the blocks, you start building it, and you start piecing them together. The other one is having a vision as well as a picture in place first. Whether you want to build a castle, a house, a horse, a car, you need to know what you want to do, and then you look for the pieces to start building towards that.

Right? A lot of technologies always falls into the trap of the first. Why? Because we are always being pushed. I need this solved, I need this clear. But I think real technologies needs to look at, I keep on mention, two things. One, we definitely need to solve today's problem because if we do not solve today's problem, there's no tomorrow. But we need to look at the day after tomorrow. Not tomorrow, we need to look at the day after tomorrow. So I mentioned that, from an OTT perspective, it is what would the end user wants, right? What will excite them coming on board to the platform, right? Today, if you look at the market, with all the players coming back into the global world, every OTT platforms come with a very similar base. Everything about it. The base is just deliver the video to the user so they can watch it, and get them on your platform continuously.

That is why the word is called content is key because content is the one that is driving. But content can drive to a certain extent. Having the user engagement back into the platform I think is as critical. And that is where I think, going forward, it is where the fight really begins. If we can find ways to have the users much more engaged with the content itself, that gives you the edge against the other platform, per se.

Zoe: Yeah, so then I have a follow up question really wanted to ask you. So how to guide, in your opinion, the final users to get fully engaged? Do you think, for example, somebody think you have to get a better quality? Quality means that the video should be more fluid. And even with high resolution, talking about HDR, 4K, or basically we may have to have some certain kind of format to really get the customer or the end users engaged? For example, if we do some extremely low latency, right? Then basically, the end user can really get some interaction with the distributor. And then this way, they can also got more engaged. So what do you see? I know this could be many factors. So if you like, like you make a justification. So which factor, what top one, top two, top three factors will mostly in influence the end user's engagement?

CP: I think what, and the things that you have just listed has become a fundamental. It is no longer good to have, right? When we first started in the OTT world, trying to get the lower latency, trying to get the higher video bit rate so that you get a better quality of video to the end users, getting HDR, Dolby, et cetera, those becomes, those are like add ons. But to me, those are actually fundamentals that it should have.

It's just like a differences between a Toyota, a Lexus, a BMW. They're all cars, right? Four wheels, engine, and so forth. The fact that the wheels are fundamental, engine is fundamental, right? The navigation system is fundamental. Then what I feel is, what is actually differentiating between them, right? From the platform, that's what I say about user engagement. How can users be more engaged with the platform?

I think some organizations are already doing today, like making group watch, content recommendations, all this. Some are passive, some are active, but this is the evolution, I think, from the platform perspective so that users becomes a part of the family. And to me, it is not just online. The integration between the online and the offline world will start to happen as well, right?

There have been, I would say a trend following what, the multiverse? I'm still trying to get in touch with what exactly is a multiverse today because different people, sorry, not multiverse, it's metaverse. Multiverse is Marvel's.

Yeah. So talking about metaverse, right? I'm still trying to get in touch. A lot of definitions of metaverse is actually out there. So I think those are the environment where any organization, even people like us start to need to understand how do all these new world can then be integrated and merge with what we have today? Even back to a very basic environment where, remember back in, what? In the past? It's not really past, maybe a few years ago, we are talking about the death of cable. Where the lean back experience of watching cable is just dying, right? And the lean forward experience of Netflix, HBO, et cetera is the new world. But if you look at what is happening today, it is changing again.

There is actually a tight coupling between the lean back experience and the lean forward experience into one. Right. So this is what I mean by evolution as well as things that we need to actually keep on looking at, right? One of the key between what we are also doing is to really merge lean back, lean forward experience to the end user so that we can satisfy, from an end user perspective, what is the best way for them to enjoy the platform. I think that's the key is to enjoy the platform. There's a lot of ways to look at it. There is no one size fit all at this moment. And the challenge for the technology team is to really identify which tech, which experience tool can help us to do that so that our users will feel that this is not just a view port, but it is a part of their lifestyle that they will be looking at.

Nathan: Hmm. Yeah, we hear social media platforms maybe talk about more in the past, but you're exactly right. Content is winning out, but it's the engagement, and it's the different types of engagement around the content. So that's fascinating. Very interesting. CP, thank you so much for letting us pick your brain and for a lot of wisdom that came out today that you shared with us. I find it fascinating how you tied the technology, the end user, and everything in between altogether. Every decision impacts the other one. So thank you again for sharing that. Thank you for being a guest on our podcast. And I hope we get to do this again and hear more about you and Mola and what you guys are doing. Thank you very much.

CP: Yeah. Thank you for getting me on here as well.

CP's experience with codec
Various tech stack and build modular
Latest technologies are exploring]
End user experience is the most important element