Search Unity

  1. Check out the Unite LA keynote for updates on the Visual Effect Editor, the FPS Sample, ECS, Unity for Film and more! Watch it now!
    Dismiss Notice
  2. The Unity Pro & Visual Studio Professional Bundle gives you the tools you need to develop faster & collaborate more efficiently. Learn more.
    Dismiss Notice
  3. Improved Prefab workflow (includes Nested Prefabs!), 2D isometric Tilemap and more! Get the 2018.3 Beta now.
    Dismiss Notice
  4. Want more efficiency in your development work? Sign up to receive weekly tech and creative know-how from Unity experts.
    Dismiss Notice
  5. Improve your Unity skills with a certified instructor in a private, interactive classroom. Watch the overview now.
    Dismiss Notice
  6. Want to see the most recent patch releases? Take a peek at the patch release page.
    Dismiss Notice

Are streaming games just about to start happening...

Discussion in 'General Discussion' started by Arowx, Mar 27, 2018.

  1. EternalAmbiguity

    EternalAmbiguity

    Joined:
    Dec 27, 2014
    Posts:
    2,133
    For what it's worth my comment there was about the "Netflix of gaming" aspect, not about streaming. I agree with most of the sentiments here that streaming isn't ready for primetime yet, if it ever will be. The game pass however is pretty much as relevant now as it'll ever be. It's not really waiting on a technological leap.
     
    zombiegorilla and Teila like this.
  2. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    Yeah, there are quite a few subscription services.

    I can't comment on the MS one, but Origin has some significant differences from the Netflix model. For starters, quite a few games on there have alternative monetisation built in, such as Battlefield 1 pushing its Premium Pass quite agressively. As a publisher it's also a way to get people into their "ecosystem" (ie: using Origin) so that they can better sell you the new releases that aren't available through the subscription. Neither of those apply to Netflix.
     
    EternalAmbiguity likes this.
  3. EternalAmbiguity

    EternalAmbiguity

    Joined:
    Dec 27, 2014
    Posts:
    2,133
    Good point about EA. And Microsoft has moved that direction with some of their games as well. Forza Motorsport 7 has loot boxes, though I don't know if they can be purchased with real money or not. And I know Sea of Thieves will not have loot boxes, but will have microtransactions (http://www.ign.com/articles/2018/02...otransactions-around-3-months-after-release-2).

    So there is indeed a difference from Netflix. But the majority of games in the program do not have these other monetization methods, so I don't know how much they're depending on that.
     
  4. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    The other comparison I was going to draw was with bundles. Games in bundles are often great quality stuff that's simply past it's primary / high sales period, which I presume would also make it a good candidate for a going in a subscription.

    But anyway, my original comment wasn't that these can't be profitable, it was that I don't think it's in our or the audience's best interests in the long term. The catch is that when I say "our" I mean "the industry as a whole", whereas the people trying this stuff are interested only in their part of the industry. A big publisher wants to grow its profits and probably doesn't much care what impact that has on other publishers - they're the competition, so why would they?

    I haven't looked up stats, but I strongly suspect that in the movie industry the per-audience-member income figure has dropped thanks to Netflix and co.* Netflix is still doing really well out of this because they did things just right and at just the right time that they've got nearly all of the audience, putting them in a dominant position. So it's good for Netflix, but is it good for the industry as a whole? I don't know enough about TV and movie production to know, but I'd have thought that more large production companies trying to compete in ways other than low pricing would be a good thing. And applying that same logic to games is the basis of my earlier comment.

    * This is based on assumptions, though, so it could be wrong. My thoughts are that since Netflix became an option my spending on TV shows and movies dropped by a few hundred bucks a year at least because for a few dollars a month I already get more than I can watch, and I suspect that this is the case for many, many people. On the flip side, it's also possible that they got the income of a bunch of people who previously spent zero dollars - free-to-air-only viewers, or people who previously pirated stuff. My point of view is clearly not a particularly robustly tested one, since it's based on a sample size of one.
     
    Last edited: Apr 3, 2018
    EternalAmbiguity likes this.
  5. EternalAmbiguity

    EternalAmbiguity

    Joined:
    Dec 27, 2014
    Posts:
    2,133
    This is also a good point. And a big problem is that it's probably very difficult to know the effects of including ones content into such a service. I'd like to think that a company could compare their revenue before adding games to the service, and then afterwards, and see if they're making more or less money, but obviously that's unrealistically simplistic and can't hope to be examined isolated from all other variables.

    Over time though, it should indeed come to light if there's a significant decline in per-user revenue, which as you say may have an effect on the industry.
     
  6. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    One argument for this style of monetisation is that per-user revenue drops but user numbers increase more than enough to compensate. Generating enough new customers could definitely be a win-win. The bit I don't understand is where that volume of new users is going to come from.
     
  7. EternalAmbiguity

    EternalAmbiguity

    Joined:
    Dec 27, 2014
    Posts:
    2,133
    I've heard the same. One possibility is that people who are only interested in one aspect of a company's output might sign up for the service, and thus the company gains someone essentially paying in universally rather than only within their niche of interest.

    The nice thing about such a service is that a lot of the risk of new experiences is taken away, so someone might wind up supporting more experiences than they would have had they had to pay $60 per game. In one sense the company is getting less than them paying $60 for each game, but in another those people wouldn't have even tried these other games with the traditional model.

    Of course for the above to be successful, a person really needs to be paying more for the service than they would for their current exclusive niche. Otherwise the company is still making more with the traditional method. But I think when someone sees the wealth of games offered by such services, possibly including games they were interested in in the past, they'll be willing to pay a bit more for the access.


    I actually just checked out the Origin Access FAQ page and re-realized something: new games aren't in Origin Access. You get a 10% discount, but by the time you break even you're buying 5 games, 5 games that will probably wind up in Origin Access so you're paying for them twice.

    Another thing to point out along similar lines - such methods may not be targeting "new" customers, but customers who'd otherwise buy these games at steep sales. If the games are only entering EA Access after 6, 12 months, well the kind of person who would wait that long is more than likely the kind of person who will wait for steep sales.

    Mass Effect Andromeda is $20 (and man, as a former ME fan that hurts. Why, Bioware...). It could easily be on sale for $15 or even $10. But if someone has Origin Access and just waited for the game to be placed on it, they'd have paid $30 for the year since release (or somewhat lower, since it was placed in the program a while back, but you get my point).

    In addition, by building a service like this they're also cutting out retailers. Even if you're buying digitally you may be buying from other places like Best Buy, who's also getting a cut of the sales. This is an ancient article, but it was the first that popped up: in retailer situations, the publisher may be getting as little as $30 on a $60 game. Given such situations, you'd understand why it's beneficial to cut out brick and mortar stores, other digital storefronts (worth pointing out that the two companies doing this, EA and MS, are two who've pulled all/most of their games from other storefronts), and even cd key resellers.
     
  8. duisti

    duisti

    Joined:
    Nov 29, 2017
    Posts:
    52
  9. Aiursrage2k

    Aiursrage2k

    Joined:
    Nov 1, 2009
    Posts:
    4,717
    Asset bundles are very easy to use, just select the objects and then assign it to an asset bundle, build the bundle, and upload.
     
  10. Murgilod

    Murgilod

    Joined:
    Nov 12, 2013
    Posts:
    3,967
    I know this is an old post and you haven't even visited the forums since August, but why are you leaving this here? Why on earth would anyone stream these games? It says the games are ad free and, yeah, that's true, but instead the ads are in the streaming app itself. Also who's going to stream games like Angry Birds when it's probably going to take more bandwidth to play the game twice than to download the app? Who is this even for?
     
    Joe-Censored likes this.
  11. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    3,660
    Meh streaming is already there and will only grow up until a formula take the world by storm, maybe a until dawn type of game where latency isn't an issue and experience is more key.
     
  12. Joe-Censored

    Joe-Censored

    Joined:
    Mar 26, 2013
    Posts:
    3,832
    I still think streaming gaming is a solution looking for a problem.

    I could see it taking off if there were very expensive graphics chips, far too expensive to be reasonable for many consumers to buy, far more capable than anything a consumer would buy, and games that take full advantage of that hardware that aren't affected by network latency. If that were to happen then streaming would actually have a product to sell that would be unavailable from other sources. Kind of like the arcades of the 70's and 80's which had all the best games and gaming hardware, where the versions for home consoles were significantly inferior (if even available at all).

    But the reality right now is consumer gaming hardware keeps getting more and more capable, and prices for near high end consumer hardware are not climbing. Graphics cards marketed to professionals may be more expensive, but aren't actually much more capable than what any gamer can buy off newegg by saving for a week on a fast food job salary. Games themselves are also not increasing in price. AAA game prices adjusting for inflation haven't increased at all since the NES days, even though the costs to create them have skyrocketed. The sizes of games have increased, but other than the transition from HDD to SSD slowing the pace of increasing storage for a few years, storage space on PC's or consoles have been rapidly increasing and is generally in abundance and very cheap.

    So I just don't see where the demand for streaming gaming is supposed to be coming from. Putting games you can already install on your computer onto a streaming service is never going to be more than a niche market. Playing games designed for PC hardware streamed to an iPad or mobile device sounds interesting, but these users haven't shown that they are particularly interested in GPU intensive games. To the contrary, they seem to be far more interested in bright colors, cute basic characters, simple particle effects, and being told over and over that they are doing great for doing next to nothing.

    I think people look at the success of Netflix, Amazon, and other video streaming services and just assume that is the next logical step for games. But I think those people fundamentally misunderstand the success of streaming video, because the alternative was having to go to the store and buy or rent a physical disk during store hours. If you were up at 2am and wanted to rent a movie, you were just completely out of luck. With games the alternative is having to wait just 20-30 minutes for your game to download from Steam, any time of day or night. That isn't anywhere close to the same situation.
     
  13. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    I think Netflix et. al. solve other problems, too.

    From a financial perspective I got a Netflix subscription because it saved money, significantly. We used to buy an armload of movies every few weeks, now we only occasionally buy things not on Netflix. On to of that, since there's no additional outlay to start watching something new, we can try stuff that we otherwise wouldn't.

    From a short-term consumer perspective that's two things that would indeed give good value if the technical issues could be solved. I say "short-term" because from an industry perspective it's a terrible idea, giving people more stuff for less money and setting up an environment where engagement duration, as opposed to quality, is the primary measure of success.
     
    Joe-Censored likes this.
  14. Kiwasi

    Kiwasi

    Joined:
    Dec 5, 2013
    Posts:
    15,831
    It really depends on the customer. For me a video game subscription at $20 a month would probably end up with me spending more per year on video games, not less. Especially given the current climate with Steam sales where no one pays full price for a game anyway. Done right subscription services could actually increase the overall pool of cash available to spend on games.

    Netflix is dramatically cheaper then going to your corner video store was. To the point that all corner video stores are out of business. But most of the funds that went to corner video stores didn't even make it back to the publisher, let alone to the artists doing the work. I've yet to see any evidence that the average actor or director is making less money due to Netflix.
     
  15. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    Sure does. According to a quick Googling, it seems like "the average gamer" spent a bit over $1000 per year on average, 10 years ago. More recent sources didn't seem to answer that question specifically, unfortunately.

    The "if done right" argument never seems to be backed by math with the assumptions explained, though.

    After accounting for growth that's happening anyway, where exactly is this massively expanded audience going to come from to account for the potential drop in average spend? Publishers, developers and platforms are already trying to grow audiences as much as they possibly can, why would this change so dramatically impact their ability to do so?

    Genuine question: Did Netflix et. al. cause a dramatic conversion of non-watchers to watchers? What about stuff like Kindle Unlimited, has that convinced a bunch of non-readers to start reading?
     
  16. Kiwasi

    Kiwasi

    Joined:
    Dec 5, 2013
    Posts:
    15,831
    I really didn't realise I was that atypical.

    If your numbers are right, there isn't a way the math can be made to work.
     
    angrypenguin likes this.
  17. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    2,093
    Exactly. Using a streaming video service like Netflix or Amazon is vastly more convenient than going to a store and purchasing physical copies of a video. Streaming video lets me instantly watch a video. Steam already gives me near instant access to games that I want to purchase and play. And running a game on a server and streaming video and audio would merely offer an inferior alternative compared to downloading games through Steam and running them locally.

    The other big difference between video and games is the number of times I might re-use the same product. With most movies, I will stream it one time and then not watch it again even if I enjoyed it. So there is very little reason to have a local copy of the movie. If I enjoy a game, I could play it a bunch of times. Some games will be dozens of hours, while others (such as competitive multiplayer) could be replayed for hundreds or even thousands of hours. If a game will be played a lot, it definitely makes sense to download it and run it locally instead of streaming each frame.
     
    Joe-Censored likes this.
  18. LurkingNinjaDev

    LurkingNinjaDev

    Joined:
    Jan 20, 2015
    Posts:
    1,242
    They contain all the hardware... I guess all the PC parts too... so it's a bit higher number than in the reality. But hey, Kotaku...
     
    Kiwasi likes this.
  19. ShilohGames

    ShilohGames

    Joined:
    Mar 24, 2014
    Posts:
    2,093
    I would not read much into those numbers. It was a statistic 10 years ago regarding game enthusiasts in North America that filled out a specific survey.
     
    angrypenguin likes this.
  20. angrypenguin

    angrypenguin

    Joined:
    Dec 29, 2011
    Posts:
    11,052
    Remember that averages can be skewed upwards by a small number of big spenders. More importantly, it's 10 years old... who the heck knows what has changed in that time? In 2008 mobile gaming was only just starting to catch on, for instance.

    I was playing with the math earlier.

    Under certain circumstances the math can work out very much in favor of a developer. For instance, if people got access to something like LoL or DotA though the service the owners would make bundles of cash, because of the whole duration-of-play thing.

    If you have a game like Skyrim which a dedicated audience plays through many times, you could end up in a similar situation.

    However, if you make games which are a few hours long and are generally played only once... well, that just isn't going to work out in your favour here, for multiple reasons.

    This is all assuming a pay-based-on-played-hours model. There are other models as well, but of course the overall system can't sustainably pay out more than it takes in.
     
  21. LurkingNinjaDev

    LurkingNinjaDev

    Joined:
    Jan 20, 2015
    Posts:
    1,242
    I don't stream my music (Google Play allows local copy anything I listen to), I don't stream my games. Period. I stream movies and TV-shows when I'm watching, because I watch those infrequently (one particular movie or one particular TV-show, although I keep local copy of the James Bond 50th anniversary Blu-Ray collection, all the Buffy tVS and all The West Wings ... :D )

    Edit: especially because the Internet here, in California either abysmal or freaking expensive, you can choose. At least comparing to Hungary (which is surprising, because Hungary isn't really good at anything else...)
     
    zenGarden likes this.
  22. Antypodish

    Antypodish

    Joined:
    Apr 29, 2014
    Posts:
    2,050
    I don't see feasible nor economical for streaming games on mass scale yet. Tell me if I am wrong.
    But at given time, just I am pessimistic about the tech.

    On server side that is
    You need lots of computing power, to be able satisfy rendering for each player, who streams the game.
    Now consider 1000s players per game. Probably need 1 GPU for each player. Or even one CPU per player.
    In contrast, in network based games, you can run multiple players on one server. Which is much more economical.

    Lets say CoD, or Battlefield, or other modern game, with high graphics details. Not to mention high 4k resolution.
    If subscription per month cost 30$, for that price in 2 months you got typically new AAA game.
    Of course you may not have strong enough PC / console to play such game otherwise. Hence streaming seams appealing and my argument will fall apart.

    30$ for a month however, seams good deal, at least from point of few of game servers. For consumer only if you play multiple games, and a lot of. But as was mentioned, you don't own game anymore. Technically you don't own game on steam, or origin either. But at least you can re download it at any time. Even 5-10 years later. Providing steam - internet will exist.


    However, as someone mentioned that this is happening but in different form, at least since early internet and later flash, then HTML5 and WebGL. Most potatos rigs could / can run these games. You can technically stream to cache and play games asap. And these already can be anything from 1 to 100 MB size these days. Yes they are not 1 to 100GB size games yet. Nor 4k res. But it does happens. And mostly free of charge, for exchange of watching ads.
     
    Joe-Censored likes this.
  23. Player7

    Player7

    Joined:
    Oct 21, 2015
    Posts:
    1,052
    If they are going to carry on bundling garbage storage hardware like 5200 RPM hdd's in consoles then yeah I could see streaming being an alternative for the end user with super fast internet speeds living near to there countries major internet exchange points.

    maybe next gen consoles drive down prices of SSD's and NVMe with an increase in capacity as these games aren't getting smaller.
     
  24. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,314
    To stream a triple A game in 1080p , this means each player is allocated a virtual PC power to render the game at 1080p , plus the server that manages multiplayer.
    When there is million online players on peak on a game i don't think such PC farms will be enough to stream and deliver 1080p or 4k for each player lol
    Gamers will always want no internet latency in solo games, no streaming.

    This is what i skip instantly, movies or games with ads, it's really for casuals lol

    Me too. I think streaming will be like fast food low price and bad quality, play a game anywhere for low price in less quality with bandwith issues.
    Perhaps once they'll have enough streaming customers they will add ads before or after you start game LOL.

    I don't think the best solo games won't be included in streaming.
    Studios like Santa Monica or Naughty Dog won't sell their games in streaming, because they could not more sell them 60$.
    For example a player wanting to play Horizon 2 and Last Of Us 2, would subscribe one month for 20$ and stop paying subscription once he finished the games instead of playing 120$.
    (it's not exactly that , because for 20$ he would access thousand games not only two)
     
  25. Antypodish

    Antypodish

    Joined:
    Apr 29, 2014
    Posts:
    2,050
    This is an interesting point. However, while generally I agree, I have one counterargument.
    Since players are on subscription, there is chance game is reached to much wider audience.
    And here were income could flow in, rather than from single one-off payment.
    Of course this would need to be justified ...
    Multiplier games would be definitely on higher advantage here, since they probably will take larger chunk of subscription as a Return on Investment (RoI) for them self (if you know what I meant).

    Well, definitely streaming and real time online playing utilize different ways of transporting data.
    While video (i.e. youtube, vimeo etc) can handle some lag and buffer up ahead, you can not do that with for example online shooter. Or you will have mad customers ;)


    If I would see ads, I skip content automatically, going somewhere else. But that is only, if ad blocker is unable to handle.
    Now new waves of (I think EU) data collection consensus agreement popups floods, which annoys so much. Specially on mobile, when hiding halve screen. Generally I don't click anything, moving on. On stationary PC at least I can remove this popups.

    But I sometimes do spent few min on quick small game, rather playing something big and for long time.
     
  26. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,314
    Naughty Dog make big income, giving their game as subscription they would loose all income they control, the winner would be the streaming provider not the game developer studio.
    Ubisoft with their streaming, they are in good position because it's their games they make that return directly money to them, they control and can decide prices, money comes directly in their pocket.
    Ubisoft giving their games to Xbox game pass for example would be a big money loose for them indeed.
    I don't think all game studios would like to be paid with some percentage to something like Xbox game pass instead of getting directly money from their 60$ price title sales.




    Most successfull multiplayer game like Fornite does not need subcription, it's free also.

    I never play casual games , i select carefully which game i will spend some time on , perhaps because i have other things to do than spending days playing, while another activity can have lot more benefits lol
    (There is many indie quality small games that are not casual without adverts on PC)
     
    Antypodish likes this.
  27. LurkingNinjaDev

    LurkingNinjaDev

    Joined:
    Jan 20, 2015
    Posts:
    1,242
    The main problem with streaming games is:
    - one player games are one-shots usually, it's hard to get money out of it constantly on a subscription-model (plus the other one is true for this one as well)
    - competitive games on streaming are very bad, it's bad enough to deal with the random lag in MP games, but now you have another problem: the lag of controls themselves. Which also is very bad, especially in a competitive MP game.

    There is a reason why local approximation happens even when you have an authoritative server. Because of the player and the need of instant feedback. If you remove it, the laggy controls will kill your game.
     
    zenGarden likes this.
  28. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    3,660
    To be frank as a poor person, It's extremely tempting for me to look at it, at least for software not yet game, I'm fine with playing subHD game anyway. Abstraction of hardware is huge.

    But the big technical elephant in the room is deep learning, there is a case of looking at that tech and what it does right now and simply seeing what it does already that would help streaming. The first thing is details correction, it can uprez, remove artifact or reconstruct entire missing part of image in less that a few ms, that reduce the storage and bandwdith issue, it can probably do some input prediction accurately, so that reduce latency, and it can probably do all sort of prediction to simulate the game world at a cheaper cost, for the price of offline rendering at production time. It's like magic for real this time, and that's the thing it already does, the straightforward case, and the tech evolve by one magnitude of order every 6 month, and that's discounting clever uses we haven't thought about. Look at how quickly we got from can generate motion matching based on example to a commercial application (less than a year).

    If you are a programmer and you are not learning deep learning, experimenting with it, you probably miss the train.
     
  29. Antypodish

    Antypodish

    Joined:
    Apr 29, 2014
    Posts:
    2,050
    And who is suppose to run this deep learning for reconstructing missing parts?
    If customer, then whole the point of needing less powerful desktop in first place is smashed against wall.
    But if provider, then it means, they need even more powerful hardware, to run such algorithms. Which is bit pointless.
    Specially needing for 4k images at 60 FPS.
    So what is the point?
     
  30. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    3,660
    Not really, because the point is that efficiency is going up and many deep learning run already on mobile phone to supplement the camera. So I'm talking about proven tech, not pie on the sky academic.

    I mean phone ARE less powerful than desktop. Also if you aren't aware you can crunch a NN to 80% of it's size and compute by removing redundant node, reducing precision of some other and baking all the weight, which is a process which can also be optimized by deep learning.

    And forward looking dedicated hardware would be even cheaper, you only need 8 bit precision for baked and crunch NN and it's mostly a bunch of mad operations. It's ridiculously parallel too.

    EDIT:
    For comparison, my geforce 705 can handle 4k video at 150ish GFLOPS I haven't tried 60fps though, so maybe 300ish GFLOPS (a switch console) might be a tipping point for regular codec.

    And you assume the tech is directed to high end user, I dson't have a 4k tv, I don't have the money of one week of low wages burger flipping in one month (rents and food suck it up), and if the opportunity and convenience level is reached, it will be the VHS, not so great quality but good enough. I mean around me a lot of people are satisfied with pirated low quality streamed screener (lot of younger people without money), so quality of rendering isn't an issue when convienience strike. So there will be a market for it, just like there is market for people playing game on cheap rig that can't run the game and look at low spec gamer (me) lol.
     
    Last edited: Nov 9, 2018
  31. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,314
    I'm not sure the magic always work, true 4K will be always better, RDR2 gamewith checkboard is not as good as native Xbox Onex 4K.
     
  32. Antypodish

    Antypodish

    Joined:
    Apr 29, 2014
    Posts:
    2,050
    Thing is, your card is likely handle most of games anyway, on 1920x1080.
    If you can do that, probably wouldn't be bother about streaming games, but rather buy one, since can play on such rig.

    Don't know how NN is good for 4K screens.
    But problem with high res, is that every details become apparent.

    Is easier to cover missing pixel on smaller screens, or lower res.

    I would like expect, game streaming if anything, mostly for none gaming rigs owners.
    Anything over network, is at risk of reduced quality. From pick hour, when network is slower, to noise, to multiple internet users, to relatively slow access to internet. So people will likely stick to buy games. Unless quality can be maintained 24/7.

    I don't think we got reliable solution yet, to convince masses, to get tech. Maybe 5 - 10 years.Will see.
     
  33. BIGTIMEMASTER

    BIGTIMEMASTER

    Joined:
    Jun 1, 2017
    Posts:
    1,448
    I'll always go with the most convenient option first. Market will keep things at the optimum level of affordability.
     
    Kiwasi likes this.
  34. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    3,660
    Oh definitely in that timeframe, irrelevant of whether the tech is ready or not, as long the will is there.

    You would use a tile base solution where you increase a tile to bigger resolution or correct that specific tile. Also consider NN a kind of lossy compression but with detail made up on the fly based on learning the "semantic" of the sources. And that's right now.

    In the future, potential based on speed optimization:
    It can also do time interpolation "right now" to create slow mo video from low rate input, though as of right now the tech isn't at sub frame speed (hasn't been attempted yet), so that one mean in theory it can also supply missing frame on top of missing pixel for smoother framerate. In fact it can also correct light rendering, so it would be possible to have a low quality shader and then pass it to get a higher quality rendering.

    This is a good layman summary of where it is at now (this do not focus on how, but what)


    Doesn't handle 720p darksiders 2, 150 gflops is lower than the ps360 era of gaming (the previous era) and teh wii u. It chugs in unity editor with a single pbr character ... it's feature complete (cuda, etc ...) but mostly an entry level 2d card for desktop business application that are not 3d intensives.

    It's true for high end user, not for "vhs" level users like me, and we might be a sizeable market.
     
    Antypodish likes this.
  35. zenGarden

    zenGarden

    Joined:
    Mar 30, 2013
    Posts:
    4,314
    It's a very general presentation, he is just enumerating different software existing from years ( Substance, Houdini, procedural tools on triple A engines) :rolleyes:
    It looks like a presentation for people new to 3D or late about latest software.
    Procedural generation on Houdini , Substance Designer are not really AI, but procedural following rules input, it's not machine learning. Triple A games like Horizon or Ghost Recon using procedural world generation are also using rules and predefined textures and terrain rules.
    This is not already AI creating full terrain world from photos database, or creating 3D models and procedural textures from simple flat photos, but this could be the next step about procedural tools.
     
  36. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    3,660
    PCG is the first part, he does the same enumeration later for machine learning, which is why I proposed that, it's just an enumeration, therefore it show the results. I should have provided a time mark, but he does give a breakdown of topic order early on, don't stop the video in the middle