September 27, 2017
SF Metrics: OpenTracing & Metrics, and Time Series Databases at Scale
On September 13th, Heavybit member Librato held their monthly SF Metrics meetup at our San Francisco Clubhouse. The event featured a talk on...
In episode 9 of Demuxed, Matt, Steve and Phil are joined by Matt Fisher, Lead Video Playback Engineer for Vimeo to discuss the past, present and future of online video and how formats like 4K and 8K could change the way content creators capture and edit footage.
About the Guests
Matt Fisher is Lead Video Playback Engineer for Vimeo where he develops ways to deliver content on the platform that specializes in high definition video. Matt was previously a Video Playback Engineer at Twitch, the world’s leading social video platform and community for gamers.
Matt McClure: Hey everybody, welcome to Demuxed again. I'm excited to say we have Phil finally back in the studio.
Phil Cluff: Back in the studio? I'm literally in a different continent.
Matt: So yeah, huge thanks to Nick, obviously. We'll have him back on because he's been fantastic. But we're always glad to see your face, even if it's via video screen.
Phil: I miss you, too.
Matt: So, Demuxed news. It's early in the year, obviously, but we do officially have a location, as we've said, and now we have dates. It's October 17th and 18th at the Bespoke in Westfield. And yes, you heard that correctly, that's two days. So we'll look forward to seeing all of your amazing submissions. We're going to need more of them than ever so we can make sure that we fill the schedule and make use of all the time that we have.
Let's go ahead and dig in. But today we have Matt Fisher from Vimeo. Known Matt, I think, since I moved out to the Bay. I think back then you worked at Twitch, when we first met. But why don't you tell us a little bit about yourself and why you're the one to talk about 4K video today.
Matt Fisher: Yeah, well as was mentioned, my name's Matt Fisher. I'm the Lead Video Playback Engineer at Vimeo. I've been there for just over a year now, I guess almost a year and a half. And as you mentioned, I came from working at Twitch.
Why am I the guy to talk about 4K, that's a great question. Aside from the fact that at Vimeo we definitely pride ourselves on high quality video, it's something that we try to enable as much possibility around video delivery and quality and good consumption experiences for our creators.
Really understanding the current landscape and where we're heading is extremely important for us, and it's something that I'm truly interested in.
Matt: I guess before we dig into 4K and 360 and VR and all that sort of stuff, which we've talked about very briefly on the show, or at least one episode, right? Yeah, light fields, which digs into VR a little bit. But yeah, I guess let's talk a little bit about HD's history online, or more generally, just video quality online.
Phil: Video resolutions.
Matt: Yeah. My favorite old-school online video example is eBaum's World. Let's go from the beginning of downloading DivX files from eBaum's World to today. Yeah, let's talk about what that journey has been like.
Matt F: I guess the real start would be, especially when you look at it from today's perspective, it seems so rudimentary, it seems so simplified. But in all honesty, there's a lot of similarities to what we do today still.
Video delivery is still relatively similar to what it was. Maybe we took a big jump from progressive to segmented delivery, but at the end of the day, it's always just constant iterations, constant upgrades. Looking back that far, you're talking back in the 360p, 240p days.
Steve Heffernan: Yeah, I remember. Even in the late '90's you had Real Player. And that was like, 1-something-p or something, but it was beautiful. It was like videos playing in my web browser!
Steve: Cutting-edge technology, as Nick would say.
Phil: To be fair, your screen was probably 800 x 600.
Phil: Looks a lot better.
Matt F: I guess that's what it was, because when you go back and look at any of that content now, there's some great internet videos that have stood the test of time and you go back and look at them now from the eBaum's World days. You look at them now and you're wondering how you even made out what was playing in the video because it looks so bad. And maybe it is just all relative to the screen size. But at the end of the day, that wasn't too long ago, right?
Matt: No, not at all. I wonder if that's part of the reason why, what was the Flash site that was so popular?
Matt: Newgrounds! Newgrounds is what I was thinking about. I wonder if the fact that it was just Flash cartoons and the shittiness of the video. It wasn't video, so it's fine. The quality looked fine.
Matt F: Well, it was all vector-based, too.
Matt: Right, exactly.
Matt F: So it scaled beautifully.
Matt: Yeah, I wonder if that has something to do with that being more popular. I'm going out on a limb here, but I'm pretty sure that had more views.
Matt F: I never thought of it like that, that's actually a great point.
Steve: You had Homestar Runner and Rascal.
Matt F: Which looked beautiful at the time, right?
Matt F: Especially when it was displayed in its native vector format through a Flash player.
Matt: Even Homestar is funny because you go to the website now and it's this tiny little square at the top of your browser window. And back then it was like, wow, this fit in my browser? Do I have enough pixels to show this?
Even then, though, I remember trying to watch videos on there and those damn loader bars would take forever, even on Homestar or whatever. And then you start to have something like eBaum's World and you're oftentimes either right clicking a file and downloading it and watching it locally or, if you're lucky, you have the browser plugin that allows you to watch it in your actual browser.
But these things were slow. I feel like it took a while for us to even get to the point where we could reliably watch, what is SD, technically?
Matt F: I would say anything sub-720p.
Matt: Okay. So there's no lower limit there?
Matt F: Yeah, I guess.
Steve: One by one.
Matt F: Go all the way down if you really want to.
Phil: That's really low definition, really SD.
Matt: Because I remember when I was legally getting videos online occasionally in college, one of my favorite resolutions was 576p.
Steve: Yeah, that's a thing.
Matt: I think that's native DVD quality, didn't take quite as long to download, legally, and it looked fine on my crappy 720p TV. It was fine. So at what point did we make this transition into 720p feeling kind of standard? Because at this point, if a video's not online, and it isn't 720p, then are you even videoing?
Matt F: Going back to the brand of Vimeo, this is why it was birthed in the first place.
Vimeo prided itself as supporting HD video very early on, right out of the gate. It's crazy to think that this was only years ago, not decades ago.
Steve: It was probably 2006, 2007? I remember there was whole slew of companies that wanted to be the HD YouTube. I could probably name 50 at one point around 2007, 2008.
Matt F: Absolutely.
Steve: And then Vimeo really owned that space, YouTube eventually brought out HD, but there was definitely a wave there of people jumping on board like, yeah, HD's the thing.
Matt F: And it's similar to the landscape we have today, where you have the battle of, just because there's a user-generated content platform to deliver this content, you still need people to create it. You still need people to generate the content and put it up there for consumption. And then on top of that you need to be able to consume it.
So whether that means having hardware capabilities or internet capabilities as a bottleneck to overcome, that's just more and more variability. I think when you compare and contrast it to today's ecosystem, there's a lot of similarities.
Matt: Let's talk about 4K. I think most online service providers would cap out at 1080p. I don't think I've seen anything higher than 1080p, and you're lucky if you get that online today. I feel like most HD, if you're clicking HD you're getting 720p, maybe you're getting 1080p.
Steve: You even have the phone plans now, they define SD as 480p and under. And they'll try and cap you. It's based on bit rate but they'll try and cap you at the resolution sizes that 480p is SD and 720p is HD and I think they don't let you go above that. If you're on one of these unlimited plans where you can turn on the smart video capping or whatever, then yeah.
Matt: In air quotes, "unlimited".
Steve: Yeah, exactly.
Matt F: I'm curious as to how big of an issue that could potentially present down the road, where we're not really decoupling bit rate from resolution. And as we strive to lower bit rates and increase resolutions, how that manifests itself from especially a cell phone plan concept.
Phil: I actually found in testing something the other day, a British internet service provider that was man-in-the-middling HLS manifests and stripping out higher bit rates.
Steve: No way!
Phil: I'm not going to name and shame, but it's out there.
Steve: Would SSL help there? If you're an internet service provider today and you want to actually deliver HD...
Phil: Yeah, this is essentially how we found it. We found it because we realized manifest was not behaving the same way over TLS as it was over Clear.
Matt: So, we've gotten to 720p, 1080p. Some people can see it, some people can't. Where does 4K fall into that, in terms of the support spectrum today? What is 4K in your mind, watchable 4K?
Reasonably today, if a website says, "we support 4K", at what point do you say "bullshit", and what point do you say, "yeah, that's legit"?
Matt F: Again, I think you have to really look at it from two perspectives. There's the bit rate and then there's the resolution. Anybody can deliver a 4K resolution with very low bit rate and I don't think the viewing experience would benefit from it whatsoever. The landscape as it exists today is a wide spectrum but narrow usage, I would say.
Again going back to the idea that when you look at it from just the 1080p perspective, and when 1080p first came out. It wasn't too long ago when having a 1080p TV was all the rage. It was a big deal to jump up even from 720p to 1080p, in my opinion, at least.
And then the down-scaling factor came in as well when you had 1080p content being down-scaled to the 720p display, it looked alright as well. So the landscape today in terms of content is growing, and I think 4K is still in its relative infancy.
Once you start seeing more and more content down to the handheld devices, cell phone devices start really pushing more and more 4K video, that really helps drive a lot of the ecosystem as well.
But it's end-to-end, it's about the creators and it's about the consumers. So without easy access to 4K televisions and an internet service provider that can provide you with enough down to support 4K, and I'm talking, let's say, 30+ megabit downloads just for that one video, without all those pieces to the equation, I don't think it can ever necessarily be solved across the board.
But that being said, I think the spectrum, as I've mentioned, just keeps getting wider. There's people that couldn't download 1080p today in some parts of the world, they're going to have trouble doing that. They'll probably favor a 720p or even 540p stream. But there's 2K video out there right now and there's a fair amount of it. There's 4K video out there now and there's growing numbers of that as well. From an engineer's perspective, it creates more challenges.
I think delivery and smart delivery is going to be a challenge that increases over time, especially with the landscape we're in right now.
We've only really been talking about television screens and computer screens from this perspective, too. So we can even just start scraping the surface with the idea of VR and spherical video and things like that, where I think the conversation completely stems off into a new direction. But I think it only continues to validate the need for these high resolutions over time.
Matt: Let's talk about why you think this is necessary, and do you think it's necessary in 2018? Earlier, before we started recording, we were talking about why this is needed. When you say that, do you think this is needed in terms of, as an industry we need to be thinking about this for next year, or is 4K just a gimmick at NAB and IBC in 2017, 2018, and it's not something that we should really care about for consumers?
Or do you think this is something that we need to start worrying about now? Or do you think is is something that we need to start worrying about in 2020? Do you see what I'm saying?
Matt F: Oh absolutely, yeah. I think it's not a deadline-driven investment. From our perspective and a lot of, and I say people working in the video industry, and even from a consumer's perspective, I wouldn't say there's any deadline. There's no prioritization around it. If it's accessible and it's not expensive, people will flock to it, I don't see why they wouldn't.
But I think we're reaching a new realm where you're going to have those people that will continually argue that having a 60-inch display, which is a relatively large display even in today's standards, you're not going to see any benefit between 4K and 8K, just from a perception perspective.
And the argument that anything about 4K requires an 80-inch display, I think looking at it in that perspective is a little bit narrow-minded. There's a lot of areas, as we've just mentioned, like the whole VR tier that I think that's a huge conversation to have in terms of putting validity behind any of this.
Steve: Just talking about VR as higher resolution itself, because it's essentially just a bigger video.
Matt F: Right, I don't think anybody in today's landscape is having absolute jaw-drop factors with spherical video. I think spherical video is great, I think it's a step in a very cool direction.
Whenever I look at spherical video and VR integrations with video, I think just from an educational perspective. I couldn't imagine how cool it would be to be in school and have VR integrated into the classroom or something like that.
But there's a lot of other use cases that I think a lot of people ignore, thinking from a content creator's perspective, just because they're filming in 8K doesn't mean they necessarily want people to consume it in 8K. They can be down-sampled or cropped and edited, which is, I think, a huge benefit from a creator's perspective.
To be able to film in 8K and not even have to pan the camera but be able to do that in post, I think is very enabling for editors and directors and content creators.
Steve: How interesting. So they could crop down into the frame, so they have a much wider resolution so they can move the camera around within that wider frame.
Matt F: Absolutely.
Steve: That's interesting.
Matt F: Right. The idea of having more ability to do things in post and keep things as high quality as you expected, I think that's very cool. And again, an area of interest a lot of people don't necessarily think about when they're on the consumption side of the equation.
Phil: It's one of those things that suddenly came up when we talked weeks ago about light field cameras as well, making a lot more decisions in post rather than it really being about high resolution immediately. It's more about giving the directors and producers and the editors more decisions later down the production chain, and I think we get the same thing here.
Matt F: But I think the next step past that is, okay, I have a film that I'm working on and it's shot in 8K. That's fine, we're probably going to release it at a 4K max or even a 1080p video online. There's still the question about shareability factor of that content during its creation, and the feasibility of using online tools to do that.
If I had a 4-plus-K video and put it online for my team to all view and put notes on and play back on their home televisions to do test views of it and such, why shouldn't we enable that? Why should the rule be that, oh, it's not worthwhile for the grand audience, so you're just going to have to plug in a hard drive or figure out some kind of high-capacity vast solution for storing your content and playing it back on devices for your unique use case.
From our perspective, I think it would be a little ignorant not to pay attention to it and not to give it the interest that it deserves. But I think the big unanswered question that everybody would ask is, when does it become a reality? And I think that's a tough question to answer. I think the only concrete date or time or expectation I've really seen around it is the 2020 Olympics in Tokyo, which will be broadcast in 8K.
Steve: Whoa, really?
Matt F: Now, who can consume that outside of Japan? Because at the end of the day, the Japanese are really leading the way and they have for quite a while. But the NHK has done some really cool stuff and research and they're always ahead of the curve around this and I think what they're doing, especially around their involvement with the 2020 Olympics, or so it seems up to this point, will really help solidify the need for this. And where the need is more evident, I think we'll find out as we travel down that road.
Phil: It's interesting because the first demo I saw of 4K was Olympics content when I was working at BBC and we actually had NHK come over and demo a load of the stuff that they were doing for it. It's going down a load of BBC backbones, experimental for the time of 2012 Olympics, so it's not surprising but we can make sense, I guess, of their pushing the 8K agenda for the next one.
Matt: When we talk about accessing, maybe you could tell us from Vimeo's perspective, when you're looking at user data for watching content, what we've heard, for example, from folks at YouTube is that surprisingly enough, people will actually select a higher bandwidth and then be okay with just waiting for that to load. They're okay with a much longer startup time, but they'd rather see the higher bit rate even if it takes them longer to get there.
With Vimeo's perception as being high quality, I would assume that people would do the same thing, especially if you're watching art house on there or something like that you're not going to want to watch at 240p. I assume you guys are already seeing this in some way play out in your data.
Matt F: It's funny you mention that in relation to initially speaking about eBaum's World. Because I think that mentality of the "play, pause, walk away, come back in 15 minutes to watch your content", that's a very archaic way of looking at things. I think it's reminiscent of progressive delivery, where you're going to download the file from start to finish either way. If you kick off the download, you know that the more time you wait, the more you're going to be able to consume when you come back.
That's not necessarily the case in today's world. You could easily start a video and the buffering algorithm within the playback mechanism is going to allocate the buffer it sees adequate to allow you to start playing back the video and hopefully not rebuffer down the line, but might not allow you to allocate any more buffer than is configured in that mechanism.
So I don't know, when I look at the idea of people play-pausing, walking away, and then coming back, it's very circumstantial whether that's actually going to benefit them or not. But to that point, I think if I was to point a finger at a demographic that would favor that behavior more than others, it would be less of the casual consumers and more of the professional consumers. People that, again, are working with creators or giving feedback to creators or have a reason to kill that time to watch content.
Generally speaking, I think a lot of the use cases of people discovering content and playing it back, they still really want that start time to be as minimal as possible.
And it goes many ways, but again, I think the main point I'm trying to make is that it's a funny mentality to have and I think it is really based off of people's experience playing back progressive content, not segmented content.
Steve: Yeah. You touched on something there where there's a lot of user expectations built into this. That's going to be a major driver, I think, of adoption of these formats. One of the things that we've heard, and maybe you guys have experienced this, when you increase the video resolution for the general audience, most people don't notice. That's not going to have a huge impact on your watch time, right?
But if you then brought it back down, that's when you'll see the impact. People will get used to the higher-quality resolution and not like the lower resolution and start to complain. So I think that's really interesting, it almost means that whoever gets there first is going to be pushing everybody else to be upgrading their resolution.
For instance, if Vimeo or YouTube started doing 4K as a norm, and assuming that there were devices that supported it, it would start to become the user expectation and start to pull everybody else to do the same. And these consumers would go to other websites and say, "Oh, the video here is crap."
Matt F: I think that brings us back to that topic of really understanding how bit rate impacts resolution. Because again, it's one thing to encode or transcode to 4K, but if you don't have the bit rate to back it up, it's going to be heavily dependent on the content you're viewing as to how good it actually looks.
I think people will always strive for more clear, beautiful looking video with the best quality of experience they can possibly have.
Unfortunately, that's not always feasible, and I think that spectrum of ability, I guess you could say, to play back and consume content continues to grow. And I think that's one of the interesting parts of this continual evolution down the line of higher, higher resolutions, more bits being shoved down the pipe, whether the ISP or connection can handle it or not.
The demand will continue to grow. Unfortunately, the ecosystem isn't growing at the same pace as the demand, and when you look at it from a worldwide perspective, having the ability to have 10 megabit download speed across the entire plant is still a dream. That's a pretty crazy realization to have when you start talking about delivering video content that could be well over 50 megabits per second.
Phil: When we talk about the ability to consume and the desire for content, where are we now in terms of hardware support? I'm sure I read a stat somewhere that said 15%, 20% of TVs that are being sold now are UHD, airquote, "ready," 4K ready. Where are we, realistically? The mobile market is getting there, but there's not that much that's anywhere close. What was the feeling at CES this year?
Matt F: I think the 8K was generally a big talking point at CES this year. But again, whether that's the hype train rolling along and keeping people aware that these hardware manufacturers are in this space and they are thinking forward is very reassuring. But whether or not that is something that you'll see in your living room next year I think is pretty doubtful.
You mentioned cell phones as well, which I think is another thing to consider because when you look at the lifespan of the average consumer cell phone as opposed to the lifespan of the same person's television, they're dramatically different. We're on 6-12 month cell phone lifespans now for a lot of people, and people don't do that with their televisions.
There's probably still a lot of people that are comfortable getting home from work and watching Netflix on their 1080p, 40-some-inch television and it works completely fine. And there's no real need, until the price drops so much that you need a new TV anyways and the de facto solution is just a 4K display. I think that is a major driver in terms of when creators are considering how they want to deliver their content or how businesses want to deliver content, whether it's worthwhile.
So you have these cell phone screens that are beautiful and they're getting better and better, but in terms of pixel density, how far do you go before it's just not worthwhile because of the size of the device?
Steve: When you're talking about a phone, there's got to be a max. We're talking, Apple sells retina displays...
Matt F: Retina displays look beautiful until you put a magnifying lens in front of it and put the screen two inches away from your face in a VR headset, and then all of a sudden you can start noticing pixels again.
As we continue all these new mediums for playback, the equation grows, and I think the considerations increase.
The idea of having 8K on a 60-plus inch television screen, when you pull that into the mobile perspective, it doesn't necessarily make sense. Does it make more sense in some other playback mechanisms than others, well if it's VR as opposed to just watching a YouTube or Vimeo video right on your handset in your hand, I think it does.
Does it mean that you deliver some types of video for some playback options as opposed to others? Maybe. But I still think when you think about the whole landscape, there are areas that have a higher priority or a higher validity of use around some of these resolutions.
Phil: It's super interesting because obviously, Netflix's post-production guidelines say for an original, you have to deliver it in 4K. You can't deliver it in anything under 4K. I think the minimum bit rate is like 240 megabits as well for what you're delivering into Netflix, regardless of whether they actually choose to deliver it at that sort of resolution. If you're shooting a Netflix original, you'd better be shooting it at 4K.
Matt F: Right, and it makes sense when you think about it. They're pushing a lot of 4K content. And especially around their originals, when you talk about uniformity and expectations around playback quality, if you're watching one Netflix original tonight and it looked gorgeous, and then you watch another one tomorrow that's a Netflix original as well, but of lesser quality, maybe it's noticeable, maybe it's not, that's a great question. But creating some uniformity behind it I think would remove a lot of potential bad feedback.
Phil: If it's Netflix content, all I know is it's going to be really grainy. Every original is grainy. I don't want film grain, I know you didn't shoot it on film!
Matt: So far we've been talking about this largely from the perspective of the consumer, what the consumer wants, what the hardware can support. Everybody in front of a microphone right now is more on the delivery side of that equation, basically making this stuff play back in a web browser or in some other sort of online delivery way.
A lot of this is already giving me a little bit of heartburn. So yeah, for people in our industry, why are people going to hate this, aside from just massive CDN bills and chucking the shit around internally in your pipelines. Why are people going to hate it, and are people going to not hate it enough? It's the customer love and our hate, at what point does that cross each other?
Matt F: That's a great question. I guess it depends how deep your pockets are. But I think something to consider to that point is, as you increase the resolution or your highest profile available, let's say, generally it's not just that. The transcode ladder is going to increase in the steps to get there.
Going from 1080p to 4K, if you're working on a good bit rate like a good ABR strategy, generally you don't want to make jumps from 1080p to 4K. That's going to be a pretty annoying viewing experience, if even possible. The circumstances to jump from 1080p to 4K directly back and forth would be very unique.
But nevertheless, if you're going to deliver 4K content, you would want the profiles to follow suit. So that would mean 1080p, 2K, maybe 5K, 6K, 8K, which dramatically increases transcode times, the amount of effort that needs to go into just getting the video into a format that can be delivered over the web. Then you have to deal with all the storage costs of storing all of that video, and then the delivery costs.
I think the bulk of the concern from the platforms that are delivering this is more from a transcode perspective, because it depends on how many people are watching it. It's all relative to how much you're going to actually pay for it to how much you're delivering it. If you're transcoding everything and only a small fraction of your consumers are watching the highest profile, I think that's the big consideration to make.
But at the end of the day, this is all doable now. It's just circumstantial as to if it fits the model at whatever you're working with and whatever you want to deliver, and how you want to deliver it and who you're delivering it to.
Phil: This all feeds into next-gen codecs. Because if you're a Netflix subscriber, you only get UHD, 4K content over HEVC. They don't deliver it any other way, there is no H264. And in my mind, I don't see anyone delivering 8K.
Do we really think people are going to deliver 8K over the internet in HEVC? Or is it something that's going to be something we don't realistically see until we've got AV1 out there and actually get some gains from that?
Matt F: Yeah, whether that's a physical blocker, it's not. It's not an absolute blocker because we can accomplish 8K video with HEVC right now. Whether that's the best solution and whether that drives it into a much higher saturation point amongst the industry and consumers and creators, that's a good point.
Again, I think the high priority ticket there in that question is the ISPs and what people can actually obtain in terms of bandwidth to their house. Again, looking at it from a worldwide perspective, and 10 megabits is the dream for everybody, then somebody else can have fiber to their door. So it's a very, very large playing field to consider.
I think it's going to continue to grow. Whether HEVC will get us there and get us to a saturation point that maybe what 4K or 1080p is now, I wouldn't want to bet either way, but I see it being a kicker, but maybe not the receiver that's going to take us all the way into the end zone. You know what I mean?
Phil: Do you ever feel what sort of bit rate we need to be talking about the HEVC 8K to be meaningfully pretty on, say, 65-, 70-inch display?
Matt F: I don't really like to think of that in the CBR mentality, more like the VBR mentality. I think the range is, I'm sure everybody has their own opinions on this. I would say anything above 50 megabits is good-looking 8K, it's circumstantial depending on what the video is.
If we're talking something that has beautiful, sharp edges or a lot of colors or a lot of movement, that's going to change. But you can get away with relatively good looking 8K at sub-50 megabit. Whether that is what we'll consider a standard, I don't know. I think right now anything below 50 megabits on an 8K video is just getting us to 8K in the environment we're in now. If you had the feasibility to up the bit rate, I think everybody would take it if they could.
Phil: So does that mean we need to start thinking about a new physical medium? Blu-ray's got us a long way. Blu-ray surprisingly got us to ultra high-def 4K UHD delivery. Do we think Blu-ray, even the high-storage Blu-rays, are going to get us anywhere with 8K? Or are we getting towards the end of physical media at all in that format?
Matt: Wow, interesting.
Matt F: Yeah, that is a great question.
Steve: We need a Blu-ray the size of a LaserDisc.
Matt: You heard it here first, online delivery is dead.
Matt F: Yeah, that's a great question though. As somebody that doesn't have a DVD player or a Blu-ray player in my house, I wouldn't be sad to not see that come to fruition. But that's not fair, that's not a fair way to look at that.
Going back to the fact that the majority of the world can't have an internet connection that can deal with 2K video, I think that's a valid concern.
Whether that manifests itself in a completely new medium or not, I think is the real question. I don't think we're ready for a new medium.
I think the concept of digital storage for video has ran its course. I knock on wood, I could be completely wrong in saying this. But if you have the capability on your internet connection to be able to download that quality of video, then I think you've pulled yourself out of that race.
So as that group size continues to get smaller and smaller, is it worthwhile for manufacturers to even create new devices, knowing that they're going to have a shorter lifespan than what would be considered that of a DVD player? Great question, I wish I had more insight into that but I'm the type of person that is out of that game completely.
Steve: Yeah, I feel like today it's a hard drive. How often do you want a physical device that can only store one movie?
Matt F: Right. Yeah, you would think that if we evolved into a new medium, we would just have solid state drives in our pockets and use that. You'd go up to rent a solid state drive.
Matt F: Right.
Phil: It's really interesting for me because I thought this and then I bought ultra high-def Blu-ray and I actually ended up having to buy the player to go with it, which actually ended up being an Xbox One. But I was blown away by how much better it was than Netflix's 4K, I really was.
Because I realized the bit rates for the dual layer is like 82-108 megabit, that's a lot of data when you talk about HEVC as well. I was blown away by how good it looked in comparison. I actually tried to find a movie that was out on ultra high-def Blu-ray and on one of the streaming services in UHD so I could back-to-back them and have strong opinions, but couldn't find any, annoyingly.
Matt: Question is, can you get Altered Carbon on DVD from Netflix's DVD service?
Phil: That show's pretty violent. That surprised me. I wasn't expecting it.
Matt: Quite enjoying it so far.
Matt F: I am really enjoying it.
Matt: That show in particular plays well into the next question of, how does HDR play into all of this? When we talk about 4K, HDR is the other word that's thrown around a lot in terms of new advancements and high quality viewing experiences. So, what is the relationship between 4K and HDR in your mind?
Matt F: In terms of HDR increasing the color gamut and actually seeing more colors in video content, effectively, I think it's a super important change that is going to be slower to manifest itself than I would like to see. In terms of its relationship to resolution, I've always been of the mindset that they're pretty independent from one another.
I think a lot of the comparisons are just being made in relation to resolution because of the time we're at right now, the move from H264 into the H265 realm, the support for it at a higher quality, but also on devices that support it, which are of higher quality now.
So I think all the pieces are just falling onto the same table at the same time, and I think a lot of those relationships or thoughts around how the two are related are just more circumstantial than anything. Nevertheless, as excited as I am about higher resolution, especially on different mediums and different platforms, I'm just as excited for HDR.
I think once you start consuming actual HDR content, especially on a big, beautiful screen, you realize how awesome it actually is. And how that represents itself with different types of content, like sports, would be awesome, more and more movies, feature films and such. It will really help benefit the landscape.
Coupled with high resolution, coupled with areas like virtual reality, I think only make it better and better. Especially in virtual reality if you're watching 360 video.
The more it hurts my eyes to look at the sun in a video, the better, in my opinion.
Pulling yourself into the virtual reality landscape that anything we can do to trick your mind, to keep you tricked for a longer period of time, creates a better experience.Doing that with games and such and full motion capture, I think is much easier to accomplish than just a 360 video with a fixed camera position, or at least what we're used to with 360 content right now.
So removing the screen door effect with increasing the size of the video, the resolution of the video, the bit rate of the video, the color spectrum of the video, I think all gets us closer to a world where that false reality can continue to trick your mind longer than, say, just watching a general 360 video where you can see the stitching and you can see the blurriness and you can't see things far away. Related, but I think coincidentally related, in my opinion.
Matt: Got it.
Phil: I think a lot of people would say that 1080p high dynamic range content looks better than 4K, especially when you think about on a screen, if you've got some bandwidth to spare but not enough to get a full UHD 4K stream down the pipe, a lot of people are going to get much more benefit, visually, from picking up a 1080p stream.
Matt F: Sure, yeah, can't argue with that. And on top of that, I think watching a 1080p that was down-sampled from 4K content looks even better.
Phil: Oh yeah, oh yeah.
Matt F: Again, going back to the argument that from a creator's perspective I think there's more validity behind super high resolutions now, especially in the online ecosystem, than the consumers, because of all of the hurdles that a consumer has to jump over just to get it, if they can even find the content.
From a creator's perspective, the amount of doors you open by overshooting or down-sampling that content is monumental.
That can be a huge win. But I'm looking at this as somebody who's not necessarily a creator, and looking at it as a fly on the wall. In my opinion it would be a much better circumstance to be in, to have overshot things and to have more options in the editing room.
Steve: How long until I can shoot 4K on my phone, do you think?
Matt F: It depends what kind of 4K, again. I'm sure a lot of phones have sensors that can shoot nice 4K video, whether that's a high bit rate, whether it has the frames per second that shooting sub-30 FPS 4K as opposed to 60 FPS 4K, at what bit rate, yeah, I don't think that world is far off at all. I think that world's coming to fruition already.
Again, you ask yourself how does that manifest itself when you go and look at just point and shoot content? Maybe the realm of very jittery content that is shot on 4K, does that open up a realm that allows video stabilization algorithms to work better?
They have a lot more room for correction, so that sure, you're still overshooting on your phone, but everything is going to come out looking like it was shot on a dolly. That would be great if you had the option. So from a creator's perspective, I think the option increase is great.
Then the big question is, do you use the internet to facilitate a mechanism to allow creators to share and create and modify and collaborate online around this content in its native form? Or do you just go ahead and run a transcode over that raw content so it can be dealt with online and then brought back almost verbally to an editor who's dealing with the raw, just that back and forth seems a little ludicrous.
It might be circumstantial in terms of how people are set up today, but in a perfect world, I think if you can facilitate all this over the internet, it would be the best case. I think this goes back to, why bother having a new hard physical medium for video past Blu-ray now, when and if you can facilitate plus-4K content over the internet?
Matt: Cool. Well, we're coming up on time. So I would like to say that I think the end goal that we've come to here is that we know we've made it once we need to wear eclipse sunglasses when we watch content on a television, right?
Matt F: That would be amazing, yeah. We'll all have welder's masks on our desks so we can have some longevity in our profession.
Matt F: There's a million-dollar idea for you right now.
Matt: I think that's also a million-dollar attack vector for the nation. Thank you so much for joining, Matt. This has been awesome.
Matt F: I appreciate it, it was great to be here.