NAB 2019: Epic Labs Talks LightFlow Per-Title Encoding
Jan Ozer: Jan Ozer here from the show floor at NAB. I'm with Alfonso Peletier from Epic Labs.
Alfonso Peletier: Hi Jan, it's good to see you.
Jan Ozer: They are debuting a product, or I guess to the NAB audience, debuting a product called LightFlow, that's one of the most exciting things that I've seen here on the NAB show floor. It's going to combine per-title encoding with network modeling, as well as per-device-type encoding, and I wanted to give Alfonso a chance to tell us about it, so.
Alfonso Peletier: Yeah, thanks for that. Yeah, we built LightFlow. It's an AI algorithm, by which we're optimizing the quality of experience across the workflows. What we're doing is that we're gathering a lot of information about the context of the user, such as what type of content they are perceiving, what type of user, of devices they hold in their hands. And we're feeding all of that information, for LightFlow AI Engine to come up with recommendations.
Jan Ozer: So how do I control that as a user?
Alfonso Peletier: Yeah, so we have three main areas where you can be controlling. The first one is what we call Content Aware Optimization. So what we're doing is that we're ingesting the content, and we're analyzing it to understand what is the complexity of it, and LightFlow is able to understand what is the optimal way to prepare that content. It's able to understand what is the optimal bitrate ladder and the encoding parameters that should be used in order to, for that content to have the maximal visual quality, with an optimized bitrate. That's what we call Content Aware Optimization. LightFlow, so that's what we call Device Aware Optimization. So since we're able to understand, we're looking after the analytics of the player, we want to understand what type of devices the users have. What is the format that they are able to play, and what type of codecs they want to play back. As a result of that, LightFlow is able to come up with recommendations in terms of multi-codec, multi-format and the players will be receiving a very rich manifest with the higher rate type of bitrate ladder, H.264, H.265, for them to be able to pick up. And then we're doing what we call Network Aware Optimization. We're able to measure the quality of experience metrics on the player, to understand what is the optimal path for the users to go on and fetch the content, in order for them to improve their startup time and reduce their buffering ratios.
Jan Ozer: Okay, what stage is the product at now?
Alfonso Peletier: Yes, we're production-ready. We have it rolled out in a few customers already. We're also partnering with some major players in the industry. We have a pretty meaningful relationship with Haivision. In fact here at NAB, we're exhibiting at their booth for a low-latency solution with an ingest in SRT, playback with CMAF over Akamai. We're also partnering with Akamai. We've been very proud to be a partner that is able to absorb all the VOD transcoding that they are doing. We're helping them with some customers as well.
Jan Ozer: Okay, so what do I buy from you and what do I pay for it to implement this system?
Alfonso Peletier: LightFlow's going to be able to make a recommendation every time your user presses play. So the more that you use LightFlow, the more optimized their bitrate is going to be, so you can--
Jan Ozer: What do I buy? What's the product? Is it a service, is it a software program? How do I use it, buy it?
Alfonso Peletier: It is a service. It is running in the cloud. Every time a user presses play, LightFlow is going to come up with--
Jan Ozer: Presumably, I have to buy it first, so what do I do to sign up and what does the cost structure of that look like?
Alfonso Peletier: Yeah, it has a cost structure based on viewed minutes of the audience, so it is similar to what an analytics company charge for. So the OTT players are used to this type of model.
Jan Ozer: Do you, I mean is it a soup-to-nuts kind of thing, where that includes all the bandwidth for delivery or is this something that sits on top of my AWS account and I pay all those costs and you?
Alfonso Peletier: Yeah, so if you have already have a relationship with a CDN for instance, LightFlow is going to be piggybacking into that. So LightFlow will come up with a recommendation. We're making good use of the workflows now that they are based on APIs, so LightFlow is making the recommendation, we integrate with your workflow and we provide our recommendations on how to optimize it.
Jan Ozer: Okay, and the exciting piece for me is A) you can build device-specific encoding ladders so you don't send a phone 1080p video, but you can send a phone HEVC video or a tablet HEVC, so you know whether they can play and you can really optimize your parameters to match the viewing platform. So it's pretty impressive stuff. Thanks for taking the time.
Streaming Media's Jan Ozer interviews Phenix Technologies' Kyle Bank on the show floor at NAB 2019, and hears why low-latency CMAF isn't good enough for Phenix's demanding customers.
Streaming Media's Jan Ozer and Encoding.com's Greg Heil discuss findings from the 2019 Global Media Format report, such as why better codecs don't always find wide adoption.
Streaming Media's Jan Ozer and NETINT's Ray Adensamer discuss NETINT's Codensity T400, which is aimed at companies that need to do large live video encoding jobs at scale.
At NAB, NPAW showed off its new smart ads service, which not only measures whether or not an ad is delivered but looks at its streaming quality and how it impacts overall viewer engagement.
See you later, per-title encoding. There's a new system in town and it's called context-aware encoding. Learn which companies unveiled CAE solutions at NAB, as well as other streaming breakthroughs.
Companies and Suppliers Mentioned