Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


MissMarple posted:

However, the two things that are happening more and more as part of GaaS is that companies are trying to improve efficiency in two areas. Content Creation, and Willingness to Pay.
A big thing that's underappreciated is that obsolescence has slowed a lot too. "We've achieved photorealism!" is still a punchline, but technological advancement is deep in diminishing returns. Hardware advances have lost their ability to revolutionize games every 5 years and the bottleneck has been shifting to production costs. One of the big risks of maintaining a game 2-3 years after launch was that it would start to look and feel dated, but that process has slowed significantly, to the point that there are games porting to the next hardware generation and continuing their life. The expectation is reversing into expecting that games GAIN value over time.

OneEightHundred fucked around with this message at 03:48 on Dec 18, 2017

Adbot
ADBOT LOVES YOU

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


hackbunny posted:

so many non-Japanese games, even high profile games
Could you be more specific?

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


I'd guess that a lot of it just came down to the fact that the job market for being good at drawing characters was a lot thinner in the US, which only really had comics and cartoons that were almost exclusively marketed to kids. Game budgets were also tiny (especially for PC games), and there was also a desire to not use Japanese illustrations on the cover art because they looked too cartoony and too obviously Japanese.

Charles Zembillas worked on kids' TV shows before he worked on Crash, and most of the other ones listed were mixed-role.

hackbunny posted:

And I'll never understand the obsession everyone seemed to have for lovely CGI covers back then, to the point of American publishers replacing well designed, timeless covers with instantly dated CGI that didn't look good even at the time
Because Americans had a stigma against illustrations anyway (see above) and everyone was on this hot new "computer-generated graphics" craze like flies on poo poo. The dazzle factor of being high-tech was more important than being, you know, artistically good for a long, long time, which should be obvious enough from how aggressively 2D anything was abandoned when 3D acceleration hardware started hitting the market, and even before that when games like Donkey Kong Country, Vector Man, and Myst were being created.

OneEightHundred fucked around with this message at 01:48 on Jan 5, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Whistling rear end in a top hat posted:

Cover art is usually the domain of the publisher.
The extent to which it's the domain of the publisher has decreased though, there's much more focus on consistency and IP opportunities now, like any game shipping today is going to make drat sure that the main character has the same design in-game, on the cover, and in the marketing material, and they're much less likely to have different art styles in different territories.

Game art and marketing used to be much more hands-off like they were with books, and imports were often especially hands-off. The Phalanx cover is an obvious case, but publishers would happily do stuff like rewrite the entire backstory in the manual because nobody really cared and most games weren't developed with a mind toward establishing an IP.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


leper khan posted:

It’s just a hard problem, and academia isn’t terribly interested in game ai, which has timing constraints not present in general decision management or planning. Not many organizations are interested in a multi-year effort to improve on the state of the art with no guarantees of success for results pretty much no one will care about.
The problem of academic interest is that there are basically two completely different reasons to make a game AI:

One of them, semi-common in board games and strategy games, is to be as good as possible at beating opponents, or at least as good as whatever constraints are imposed to control its difficulty, and that type of thing is of interest to academia because it's useful elsewhere. Timing constraints are definitely not a barrier to academic research, there's been work on stuff like Starcraft and Doom bots and training self-driving cars in GTA entirely because so many useful applications of AI in things like robotics and self-driving cars have to be done in real time.

The other one is the one that you see in most game NPCs, which is not to make an effective AI, but more of a systemic performance, and the goal is to be interesting and fun to play against. A lot of being fun is things like giving enemies a distinct identity, and making them seem fair (which means, more than anything, giving the player a huge information advantage), but actual intelligence is often counterproductive. Most of the interest now is in group and emergent behaviors so that the AIs are more interesting and do more things.

OneEightHundred fucked around with this message at 23:12 on Jan 7, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


As long as the user can tamper with the operating system, install third-party drivers, and run in a VM, there's no way to get rid of cheats.

A distrusting server can prevent a lot of cheats, but it can't prevent any cheat that mimics superhuman reflexes. Hiding information from the client is also easier said than done, like yes you can stop sending information about players that should be blocked by LOS, but then they stop making sound.

The upshot is that detection only needs to work once to ban a cheater, and going to higher privilege levels in the OS or even hardware to stay hidden always means a drastic increase in complexity.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Chernabog posted:

Hearthstone: They have some weird stances like refusing to buff cards that never see play or nerfing cards out of existence rather than making them reasonable. And sometimes they justify those stances with even weirder explanations like " it's not the soul of the card." (TBF they have gotten better about this but it still happens occasionally.)
Just in general, balance is difficult, especially in a game that has a lot of hard-to-quantify effects (disables, movement, range, etc.), and sometimes the thresholds for differences in effect are really small. Burst damage is an obvious case: The overwhelming majority of things in games are 100% effective until their health is zero, so just the difference between getting a kill and not getting a kill on something with poor follow-through can be enormous.

A lot of things are just not good designs and a fun place for them can't really be found. If something is amazing but only works in specific situations, then more has to be dedicated to producing those situations and/or trying to prevent disadvantageous encounters, so that creates incentives for extremely defensive compositions. Conversely, some things become more effective if you run more of it (usually because there are severe drawbacks to running enough counters to deal with it), so then it's either something that dominates your entire composition, or isn't worth it. Some things are more effective at different skill levels, and a lot of players just don't really care if something is ineffective and just want to play the fantasy (*cough* sniper rifles).

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


SupSuper posted:

Why suffer through cheap CD-ROM load speeds when you've got a perfectly good hard disk right there?
Well, there have also been significant stretches of time when optical storage was a significant fraction of what hard disks could store, and a lot of games made during that period were heavily seek-optimized, which eliminates the main drawback of optical. I do think that it's kind of unusual that running off of the disc with a minimal install hasn't really even been an option since before DVD-ROMs were out, especially given the number of games that required the disc to be in the drive for copy protection anyway.

That said, a lot of early CD-ROM games did run off of the disc.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


shame on an IGA posted:

I think the tipping point for that was around 2001-2002 when RPGs outgrew a single CD with heavy compression and DVD-ROM hadn't penetrated the market yet.
Well, I'll give some historical points from things like archived Best Buy flyers because there were several things changing over at different points in time.

A typical new computer sold in 1996 had a hard drive capacity of 1.6-2GB. In 1997, Riven was released on 5 CDs, so at that point, there were still some games that were too large to even install.

It depended a bit on the type of game though. The two main things that caused a game to use large amounts of CD space were using red book audio (a.k.a. uncompressed CD tracks) for music, and FMVs. Games that didn't do either of those were typically much smaller and could be installed to disk, even if they were on CD.

In 2001, a "gaming PC" had a hard drive capacity of about 80GB, so at that point, there was more than enough capacity for installs to disk to be the norm.

DVD-ROMs weren't available until 2000 or so and it took a while for them to penetrate the PC market. There are a few reasons for that, but it was partly cheapskatery and partly because it was often a choice between a DVD-ROM or a CD-RW drive. The iPod didn't come out until 2001, 99% of people listened to music off of CDs, MP3 piracy was the hottest poo poo in town, and CDs were by far the easiest way to move large amounts of data between PCs, so CD-RWs were often the more attractive option. Most PC games were shipping on CD in the mid-2000's even as the Xbox 360 and PS3 were launching, although that was about the time that games were starting to change over to shipping on both formats or occasionally DVD only.

That problem also coincided not so much with RPGs but with games shifting away from pre-rendered and FMV-heavy content towards game designs that were much harder to split up across discs. There's a reason that there are a lot of multi-CD games but not very many multi-DVD games.

OneEightHundred fucked around with this message at 18:06 on Mar 6, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Commute 15-20 minute mostly-highway drive from burbs to burbs.

Lunch = Aside from BYOL, there are some sandwiches+salads for sale in the break room, a cafe in a shared area outside of the office (which unfortunately was used as an excuse by the property owner to ban food trucks, boo), a few shopping centers about a 10 minute drive out. If you want something fancy, then you can drive 20 minutes drive to downtown Durham and have a tarantula burger.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


leper khan posted:

Circle is the button typically used in Japan for accept on PlayStation controllers. Also on a lot of ps1 games. It’s where A is on Nintendo hardware.
I'm still not sure why the reversal happened. The US/Japan schemes used to be the same, up until the middle of the PS1 generation when they diverged. It's not like there's some legacy reason for it... the legacy was to leave it the way it was.

Canine Blues Arooo posted:

The difference between these games (Save Payday 2...) and the Mobile ones can be summed up in one word: Integrity.
There's a talk somewhere or other that compared the mobile/F2P situation to syndicated television, talking about how the business model really sets the parameters of what you can do. The whole model of syndicated TV is that the same show could be running on multiple channels and people have to be able to turn it on and watch it even if they were watching a completely different season the day before, so everything had to be chopped up and it was much harder to run long story arcs, etc. whereas the "golden age of TV" with things like Breaking Bad was only possible by changing the business model.

The thing with games that are fully paid up front is simply that there's never a point where the developers have to think about how to convert non-paying players to paying players. F2P always does, and most of them answer by trying to repeatedly annoy the crap out of players that aren't paying, and many of those also figured "hey, we can just never stop annoying the player and we'll keep getting more money!" It resembles coin-op in a lot of ways.

Canine Blues Arooo posted:

I think most Western Games, especially ones with larger budgets, lack any kind of cohesive direction or vision.
I think the bigger problem is that western games have become largely about creating a comfortable experience. The Souls series was a bit of a wake-up call, but there's more to the problem than just difficulty, it's also a lack of willingness to defy convention, a lack of non-mechanical challenges, and a lack of good motivators outside of the core gameplay loop.

GC_ChrisReeves posted:

Yeah DLC and MTX are more of a thing now because of massively increased development costs, dev time, higher quality bar and static box prices over 40 years, so margins have been getting ever smaller.
DLC is just a modernized way of doing expansion packs and MTX exists because people value their money differently.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


I dunno if I'd call them trolls, but I've seen reviews on Glassdoor that are almost definitely fictional.

There are also a handful that are literally outside commentary from people that have no idea what goes on inside the company.

(I don't work for Riot.)

kznlol posted:

One of the FPS games I played in early testing phases recently used UE4, and had a console command that allowed you to specify your mouse sensitivity by first inputting your mouse DPI and then inputting a desired distance (in inches) of mouse travel required for a 360.

This was super convenient for me (and for a lot of other players who played the game) because I really dislike having different sensitivity in different FPS games. Is doing something like this actually difficult? Why do I not see similar options (at least, for instance, displaying the required mouse travel for a 360 in the game somewhere) in other games?
There are a few complications:

You'd need to get everyone to agree on a standard, it would probably needs to be kicked off by a third party because the incentive for every developer is, if anything, to match the sensitivity behavior of their own games.

Physical distance to turn distance isn't necessarily linear. There can be acceleration curves, contextual modifiers (zoom-in), and pitch/yaw might have different sensitivities (usually less pitch sensitivity).

Ideal sensitivity isn't necessarily the same for different games anyway. The whole point of sensitivity controls on gaming mice is that it's not necessarily the same in different parts of the same game.

OneEightHundred fucked around with this message at 05:51 on May 30, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


For games at least, using 32-bit numbers everywhere isn't usually too big of a deal in practice because it's extremely rare to have buffers or pointer diffs that span over 2GB. Most of the problems from 64-bit porting are from union types breaking, and from schemes involving bytewise compares/hashing/etc. breaking due to the pointer size change introducing padding that contains uninitialized data.

I've also seen at least one case of "let's serialize this structure containing a pointer to disk, and we'll just fill in the pointer after it loads!" introducing a compatibility break.

OneEightHundred fucked around with this message at 17:45 on Jun 27, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


A huge complication is dependency trees, caching, and reuse. If you know exactly what assets you'll have when you start loading and exactly how much when you finish, then maybe you can time it reliably. But let's say you drop the player into a random location on the map in an open world game, and there are like 25 world chunks to load. A bunch of those are going to have assets used in multiple chunks, but they only need to be loaded once (and might already be loaded!), so the cost of loading each chunk is variable, and the only way to actually figure out the full set is something like per-chunk manifests, and then you have to spend time and memory de-duping the manifests and looking up all of the file sizes.

The alternative is to say "each world chunk is 4%."

My favorite is how City of Heroes did it for a long time (it was eventually patched to be more accurate). It would fill the loading bar by a fixed amount for each asset it loaded and then just speedily ran through the rest of the bar when it was done, and it was just spaced in a way that the worst-case scenario would only fill up like 2/3 of the bar.

ChickenWing posted:

machine learning-integrated loading bars
This will only be a joke for so long.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Reducing load times is kind of hard because of annoying hardware bottlenecks. If you want a game to look nice, you want to fill up memory with as much high-resolution stuff as will fit. If you want that memory to be used efficiently, you want stuff that hopefully has a fairly compact in-memory representation, and doesn't compress well, so that means that "ideally," load times are about as long as it takes to pull ~8GB of data off of the hard drive.

A 5400 RPM 2.5" magnetic drive only reads like 100-130MB per second though, so that's like a minute.

This is also one reason that unskippable intro movies can exist (to act as a flashy loading screen, essentially).

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Re: Linux, IIRC another complication is that most Linux users that play games dual-boot Windows already, so that reduces the benefit of porting even further.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Another factor is that a bunch of first-person games flat-out don't have third-person animations for a bunch of player actions, or even a third-person character model (i.e. Bioshock), so they'd have to create all that just so you could look at yourself in a mirror.

It's also kind of a difference of context, I guess. Like the reason there were a bunch of mirrors in 90's games and are fewer now isn't because rendering tech moved in directions that make mirrors less viable, it's that it was a new feature then, so games were jamming in mirrors and reflective floors to show off the tech. (Kind of like the ray-tracing demos are doing now with puddles and polished metal.)

OneEightHundred fucked around with this message at 17:06 on Aug 27, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Chev posted:

Browsers don't, but the private servers they link to might. And they might do a ton of other things they aren't supposed to. The problem with browsers isn't browsers in the first place, it's private servers. Any guarantee you may have from the publisher or devs isn't guaranteed anymore. You might say as long as it's private servers it's none of their business but they're still the ones who are gonna have to explain to the press why it's fine that their game's the one with private servers full of pictures of giant sausages, or servers that collect IPs for hacker use or whatever.
That can be mitigated by using a rental model, though you still have to worry about bad admins.

Stick100 posted:

Having a server browser lets people see how many (or few) players are online.
They actually don't, and that's one of their subtle but important problems. Even in popular games with less-popular modes, regions, times of day, etc. they're prone to a problem where there might be far more than enough people to get a game going, but nobody actually in a server playing, so everyone looks at the server list, sees a bunch of empty servers, and logs off.

Above that point is a couple servers being heavily queued, and the rest empty.

OneEightHundred fucked around with this message at 17:23 on Oct 13, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Star Warrior X posted:

GTA 5 doesn't cost $60. It costs $60 minus whatever steam discount, plus whatever number you feel like spending on funbux.
This is really the biggest factor. DLC and microtransactions have completely overhauled the revenue model of games, not just in terms of getting more money out of the same box price, but also allowing highly-variable amounts of money to be spent on it.

OneEightHundred fucked around with this message at 00:23 on Oct 22, 2018

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Red Bones posted:

With modern remasters of classic games (e.g. the recent Spyro and Crash Bandicoot series remasters, the 2011 Halo 1 remaster, the upcoming Warcraft III remaster) how does the cost, the development time, and the development process compare to developing a game from scratch? For clarity, I'm asking about the remasters where the graphics and the engine are overhauled or remade but the game design, level structure etc etc remain the same, rather than a total remake/reimagining that changes a lot of features.
It depends massively on what the code looks like, and how the tech has changed. Code written for the PSX is tough to port because the PSX didn't have hardware floating point or real 3D support, so games written for it weren't even doing their most basic math and graphics the same way as a modern game would. For something like that, it might make sense to just figure out a way to port the useful portions of the assets over to new code and mostly redo it.

That wasn't even an unusual approach at the time, like the FF7 PC port uses very different data formats from the PSX version, which AFAIK is partly because some of the PSX data (like the music) was designed to be interpreted by PSX-only libraries supplied by Sony.

Halo:CE's remaster was a fork, there have been a decent number of interviews/articles with details on it.

Some PS remasters, the remasters of the PS2 God of War games for certain, were done by forking the code and using only retail disk data (i.e. nothing was rebuilt from the source assets).

Xun posted:

If this isn't too politically charged, how do you guys feel about reddit mobs getting game employees fired like in the Arenanet thing? Is this something you're worried about or were the company(ies) who gave into the mobs just unusual pushovers? Does this affect your opinions on unionizing?
I don't think a Reddit mob was really the main factor there. There was a Reddit mob over having a trans character in the same game and they ignored it. The problem is more that responding to a player giving a suggestion (even a stupid/obvious one) by retweeting them with a comment that amounts to "get a load of this guy" is an extremely bad idea.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Hughlander posted:

Yes but that happens in every industry, it's not games specific. Best you can do is keep track of game mechanics you love/hate and if someone comes up with an idea like that be prepared to show them where it was done in the past and people hated it. Youtube clips of it from LPs, reddit posts disparaging it etc...
I think the best route is try to hammer out some quick proof-of-concept of the way you think it should be. It helps cut through a lot of the reservations.

... unless they're convinced that, say, waiting 5 seconds to open a door is better because it's more "visceral" or some other thing that they think is way more important than it really is, at which point you have to wait for fresh eyes testers to tell them it sucks.

Hughlander posted:

I remember calling that out at some places saying, "We just need an A-Frame where the elevator doors open to saying "This quarter all design decisions will be copied from..." and then a spot for a box where we can put CoD/Halo/Battlefield/PUBG. It'll save a lot of time in meetings."
ftfy

OneEightHundred fucked around with this message at 18:57 on Feb 26, 2019

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


RossCo posted:

It's also interesting to see peoples reactions when the controversial position of "I don't object to a union in principal, but I don't see the benefits given current evidence" is stated.
You're going to see a lot of "don't see the benefits" because framing it as primarily a remedy for abusive working conditions is a mistake. They're white-collar desk jobs with decent pay and benefits. Despite the headlines, working conditions and job security have been steadily improving, mostly because of consolidation and risk dilution. Talking about how much studio heads make has to contend with widespread profit sharing and bonus plans. A LOT of the workers accept the long hours as worth it for the fat bonus or to work on something awesome. Many of them could quit and easily get a job that pays more with fewer hours at a non-game company.

That isn't to say that the problems aren't real, but if the pitch is "join a union because you're being abused" then it's going to run into a lot of people that think "well I don't feel abused, and it's not like I'm trapped in this job anyway."

The more relevant thing really is that it's simply leaving a lot of money and benefits on the table by not doing it.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


About the frame rate stuff, there are a lot of ways that it can happen beyond just collision detection.

Framerate dependency mostly happens because the intuitive way of writing a game loop is to advance the game by whatever timestep, then figure out all of the various things that can affect the game state that happened since the last frame, and apply those to the game state according to the game rules, and repeat.

The problem with doing things that way is that, by default, it's not dealing with the fact that things happen between frames and in a particular order at all and there are a lot of ways that can lead to discrepancies.

Of those things that affect the game state, the order that they're processed is almost never determined by what sub-frame time the event theoretically would have happened, it's determined by something else like what order the objects were created, which means that the order that things happen can change depending on whether they happen in the same frame or different frames. It doesn't matter if the physics system is good enough to handle subframe interactions if it all goes into per-object contact lists that then get processed in a completely different order and aren't interleaved.

Failing to integrate acceleration and gravity when making physical-behaving stuff that doesn't use the actual physics system is a another common mistake.

Subframe time carryover is rarely handled correctly. By that I mean, if a game is stepping once every 100ms (just hypothetically), and some event is supposed to happen in 150ms, then it will really happen at the 200ms frame, but most game logic is written in a way that any responses to that event are treated as if they happened at the current frame time, so they start at the 200ms mark too and now that's all running 50ms behind. I think most shooters by now are handling this correctly with rapid-fire weapons, and sometimes it's done correctly by accident, but it's still a problem for state transitions (like animation states) and especially timers.

Using a fixed time step doesn't make these problems go away, but it does make them behave consistently, and it also allows a lot of subframe carryover problems to be eliminated by doing things like making all of the animations last for multiples of the frame time.


Also just a fun personal anecdote related to the carryover problems, I was making a TAS mod of Quake and found out that if the framerate gets high enough, it hits an internal 72Hz limiter, which conflicts with nearly all of the animation in the game being driven by chained 100ms timers, and causes everything to animate 10% slower than it should.

OneEightHundred fucked around with this message at 07:10 on Apr 29, 2019

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Big K of Justice posted:

Feature animation is pretty good about giving everyone and their dog a credit, but games? I can't think of any restrictions on not giving everyone a credit.. even if its under one heading/card.
I don't think there are any legal requirements. There are a ton of games, mostly non-AAA games, that don't have a credits listing at all.

Hollywood AFAIK is driven by union rules, both in who must be credited, and also who must not be credited (i.e. see the Robert Rodriguez DGA incident).

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


leper khan posted:

Specifically placing and filling out props in environments is one of the roles on the chopping block. AI based approaches are getting much better. Eventually the role will transition to working with a tool that automatically does 90+% of that work and then doing some touch ups.
For general world fill-out, yeah, but there are still a lot of cases that it's very difficult to make procedural generation work well, especially tighter areas where you want it to look detailed, but you don't want the detail crapping up player navigation.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


mutata posted:

(A) Clearly explaining why they're making the deal, including the financial reasons, survival reasons, "the end result will be a better game than without this funding", etc. Basically telling the truth.
Problem with this is that it's near impossible to change anyone's mind once they're convinced that EGS is bad and the extra money is just going into the yacht fund. It's hard to actually demonstrate that the money is making the game better, and going into more detail is just going to invite more armchair expertise and disbelief that things cost as much as they do.

If the root problem is doubt over how much the company cares about its customers, then the only answer anyone's going to accept there is "here's all the stuff we're going to do to show how much we love our customers," so may as well just skip straight to that, do a good job of making and maintaining the game, and let that do the talking. Get the conversation away from business/financial/store stuff that really has very little direct effect on the player experience anyway unless there's some bundle/bonus deal or a sale.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Hyper Crab Tank posted:

To be fair, it's not exactly an easy problem to solve. Steam's efforts for years have been geared towards finding a way to get games in front of people that are likely to buy them that doesn't require spending countless man-hours on curation. There are millions of people with Steam accounts with diverse tastes, budgets, available time each week to play video games, computer hardware, what have you. Ultimately Steam's goal aligns with that of both customers and developers - getting games into the hands of the people who want to play them - and there really isn't another platform today that offers what Steam does with the same availability to developers. It would be great if there were for competitive purposes, but there isn't.
Algorithms and curation can't make good games more visible through the chaff without someone's game being the "chaff" that gets buried, and then the people on the receiving end of that will complain about that instead. Steam gets a lot of hate for what's really a problem of a hyper-competitive market caused by the cost of making a GOOD game coming way down.

Ironically, the 30% issue is largely related to that. Steam offers some nice stuff, but they could ask 30% because of their massive dominance of the PC distribution market. As games rely more on other promotional avenues to get noticed, Steam becomes less important to their success and it becomes easier to pull the game off of it. Launcher/storefront fragmentation has set in now and is probably irreversible.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


There's kind of a combination of good reasons and bad reasons. The good reasons are that it can achieve some artistic effects. They can make things seem less in-focus if that's what you want. Achromatic lenses are kind of interesting, you can control what color fringes you get depending on the wavelengths that you intersect, and wind up with more of a certain pairing of colors in the frame, and they don't produce as much of an out-of-focus sensation as a simple ("rainbow"-producing) lens.

The bad reasons... video game graphics have a very long-running "more is better" mentality, and a tendency to adopt new visual effects before they've really been refined, with later games dialing them back to more reasonable levels.

OneEightHundred fucked around with this message at 06:42 on Nov 2, 2019

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Shaocaholica posted:

I used to work at a pretty big studio doing 360 and PS3 dev but that was forever ago. Now I work in a totally different but related industry.

Anyway when we were working with 360/PS3 we were always up against the memory limit that by todays standards is laughable at 512mb.
PS3 was split 256MB main and 256MB GPU which was even more annoying, since you couldn't use more of one at the expense of the other like you could with the 360's 512MB unified.

quote:

My question is, are developers still working right up to the limit of consoles even if consoles end up going to 16GB/32GB?
Yes, but not exactly with the same implications.

Are games using most of the RAM they're given? Yes, of course, they're always gonna run up RAM because any RAM that you're not using is RAM you could be using to boost texture detail or something like that.

Are they still having to tune things around the RAM budget? Yes, because see the previous answer.

However, a lot of things that could take a pretty big bite out of main RAM on the last generation and were hard to reduce at all (like executable size, and various types of game data) have not scaled up as much as the RAM limit, so more of it is being used on fewer things that have more room to be scaled down if needed.

OneEightHundred fucked around with this message at 19:12 on Jan 10, 2020

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


mutata posted:

Tools and artistic capabilities have tended to outpace an engine's or console's ability to push the art so every new generational shift tends to just accommodate what artists seem to already want to do. We also have paradigms of cg art creation that reach all the way into animated film and film special effects, so there's plenty of knowledge on how to push things harder higher faster etc.
We also know what's obviously an approximation of more complex things. Bumpmaps/parallax, hair, transparents, area lights, soft shadows, formula-based BRDFs, TAA, indirect lighting, etc.

(I think it's a bit of a red herring though. Rendering tech is deep in diminishing returns, content really matters far more for driving visual quality now, so content creation improvements, a.k.a. better tools, are really where the action is going to be going forward, and the rendering tech breakthroughs that do matter are the ones that make content creation easier. That's not even getting into how rendering tech has been losing its ability to differentiate games from each other thanks to feature set convergence and engine reuse.)

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Shaocaholica posted:

I think human animation face and body have a long way to go in games. Deformations still look gamey. Clothing simulation still looks gamey. They are much better than before I guess but not as good as what's actually possible which just shows the tools and pipeline need work.
Spider-Man bulk-streamed verts for its facial animation and it worked well, so I think that'll become more common.

A lot of the problem of things looking "gamey" is tough to solve though because even with a highly complicated and perfect simulation, there simply isn't enough information about player intent available from 6 gamepad inputs, and players want to do things that are physically absurd. How do you accurately animate a player sprinting straight into a knee-high immovable object that doesn't involve them tumbling on to the floor? "Well obviously they'd decelerate," yeah but they're supposed to move the stick to decelerate, and they didn't, so now what? Same with insisting that they go from idle pose to doing a 180 and shooting a specific pixel on the screen and that pixel had better have a bullet in it next frame. Anything that requires an anticipatory action that isn't mapped to an input is just SOL.

Cloth physics have a similar issue: While there are some obvious improvements to be made, cloth clipping through models happens in cases where the accurate thing would be to tie in knots and likely restrict your movement.

Even with all that though, content is still kind of king because once extra simulations get tacked on to things, it creates a bunch of content work to hook it all up, and to tame it to the point that it's not doing anything screwy, especially if it affects gameplay.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Coffee Jones posted:

Those are all CPU comparisons, Mark Cerny was talking about a dedicated decompression ASIC built into the PS5 which is getting me hot and bothered.
ASIC decompression's actually been a thing for a while. PS3, PS4, and Xbox One all have Zlib decompression ASICs, but some games are choosing software decompression anyway for technical reasons.

Big thing that happened with lossless compression is asymmetric numeral systems were invented a few years ago, which made it possible to get arithmetic/range coder efficiency (i.e. provably optimal) with much better performance, so that kicked off the race to crank out new algorithms incorporating it. zstd is one of them, pretty sure Kraken is too.

TBH I'd be happier if we got ASTC support so we could start knocking texture sizes down, it'd probably have a bigger impact, and it's really disappointing that newer desktop/console GPUs don't support it. The only sub-8bpp option is still S3TC, a 1998-era format that's been improved on multiple times.

OneEightHundred fucked around with this message at 08:30 on Mar 21, 2020

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Chev posted:

Yeah, normal mapping was kinda pointless in games until we got better lighting models that'd come a couple years later with real time programmable shaders, everything would look plastic, plus game makers didn't really figure out what to do with normal maps until the idea of baking higher poly models came about
There were some other complications. They were introduced at a time when fully dynamic lighting was an extreme technical challenge given the available GPU power, and none of the approximations of complex multi-light environments (vertex shading, lightmaps) worked with normal mapping, so the path of least resistance was very harsh lighting environments with a small number of light sources.

OneEightHundred fucked around with this message at 08:25 on Jul 28, 2020

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


It's kind of funny that the two example cases are a matchmaking system and an airport because they're completely opposite in terms of business strategy.

A matchmaking system you basically want to gather as much info about how it's working as you can because there are a lot of weird human factors that go in to how well it works, especially in sparser markets, and you really want to plan for its fallibility and design everything with the assumption that you might have to go back and change it.

An airport is one of the longest-term infrastructure projects you can possibly do, and every single decision made is going to stick for an extremely long time, so you'd better get everything absolutely perfect on the first attempt.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


SerthVarnee posted:

Any WTF moment you throw up as a hook is spoiled ingame.
Nah, you just tease with a tiny fragment of the WTF moment so people that see it want to see the rest of it.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


You'll wind up seeing most games using it just because SSR's failure cases are just awful, but it is kind of a wash. It is definitely beneficial, but side-by-sides are deceptive because people playing the game aren't getting to see that alternate experience to make the comparison, and while it helps a lot in low lighting conditions in particular, it's also those conditions where it lacks a "hook" the most. The high-poly geo stuff that Epic demo'd has way more obvious benefit for instance.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


They're bigger because AAA is a production quality arms race above all else, and they'll use every ounce of hardware capacity they can get. Games are bigger because storage/bandwidth are abundant.

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


TooMuchAbstraction posted:

Look at the credits for a game you enjoyed, odds are there's something like 10x as many graphics artists as programmers.
That's a bit of a red herring though. The increase in staffing is mostly because of an increase in budgets, driven by increases in sales, and it doesn't really correlate well with disk usage. Textures are the biggest space hog, but the production cost of a single texture has very little to do with its resolution, for instance. The biggest driver of disk usage is the expand-to-fit nature of AAA. I will use a 32k texture for a thread sticking out of your shirt seam if the hardware lets me get away with it.

(That's not really THAT much of a joke either, one of the things making the UE5 demo tick supposedly is massive stream-in rate from the SSD, in which case some napkin math suggests it's probably taking up a big chunk of the SSD just for the demo... which would be pretty normal for a demo, really, but if you're talking about scaling up to a full game and asking how much storage it could get away with using, if it had no limit? You're probably talking tens of terabytes.)

dads friend steve posted:

I thought, at least for consoles, huge install sizes were driven a lot by needing multiple copies of assets to reduce seek time. So since the next gen is all super-fast SSD, we can expect some decrease in on-disk size (possibly to be immediately gobbled up by increased texture and poly sizes?)
AFAIK that's not really that common. Most of the benefit to seek optimization, you can get through location alone, and the most well-known case of duplication I remember was Insomniac doing it when they were doing it on Blu-ray discs where there was zero downside to using up every drop of spare space.

OneEightHundred fucked around with this message at 20:32 on Nov 6, 2020

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


Coffee Jones posted:

Regarding those middleware logos (SpeedTree, Scaleform, Miles Sound System, etc) on initial launch of a lot of titles, are you not allowed to do anything in the background while they’re displaying? Loading, logging into a multiplayer server, compiling textures, always happen after the logos.
It's pretty common for there to be stuff loading during intro videos. If you're doing it on PC, go to Task Manager and watch memory usage.

Adbot
ADBOT LOVES YOU

OneEightHundred
Feb 28, 2008

Soon, we will be unstoppable!


chglcu posted:

Lots of middleware contractually obligates you to make the player miserable. It's awesome.
99% of vendor intro videos are because of either that, or because they paid for it.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply