The Star Citizen Thread v5

[Zyll Goliat]

The major netcode overhaul comes with the 3.0 update. It's been known for quite a while now.

Well I ask same question 3 years ago....your friend Mr.Nowak said back then that this will be a kid´s play to fix and it´s a matter of weeks not month´s when we going to see hundreds of people per instance....

Guess what 3+ years forward and we are in the same place as we are been at the begininng .....16 players,not stable,laggy mess..with almost no improvements at all from the vanilla netcode and you want me to believe that 3.0 going to bring some"MAGIC" right?That we going to see such a huge improvement in the engine that it´s known for problematic netcode for ages?All games that are FPS based and required huge data-transfer between players(aKa netcode)in the CE didn´t pass that magic number of 24 ppl. per instance....to be honest 99% of games are been on 16 players per instance and even then we all know how bad they act in that multiplayer environment....

Just ask yourself why no one is using CE to build multiplayer FPS action games???The very same CRYTEK engineers could not fix or improve this problem for AGES and now Croberts magical skills will suddenly do that part???

 

[szopin]

Not really, no. It's all stuff we've seen a bajillion different times in a bajillion different shovelware games and movies. I mean, it's not really surprising considering how much we've seen them rip off existing stuff in the concept art that drives those designs — of course the final product will look like everything else since that's where they got it from.

Quite a few. Hell, there's an entire secondary marketplace on Steam that lets you do that these days. Not to mention games that are continuously being updated — games like, say, EVE online.

But again, that's not really the point here. The point is that SC is simply not at an alpha state yet. Even CIG says as much. The question is when will it be, because at the current pace, a proper alpha seems a very far way off. There's just too much missing, not just in terms of content and foundation, but in actual tools needed to do any of the testing that gives an alpha its purpose.

It's CryEngine 3 graphical quality, which is decent enough, but it's definitely starting to show its age. Not even CIG is dumb enough to fiddle with the single thing that CE3 has going for it: its rendering pipeline, so short of an complete engine replacement, it's already behind the top performers out now.

That wouldn't be much of a problem in and of itself though, since most of visual quality these days comes in the form of having intelligently made assets — the stuff people made the last-gen engines do at the end of their lives was still pretty spectacular compared to what the same engine did just a few years later. But that's where Chris' lack of experience starts to show, what with his reliance on pure geometry and similar brute-force methods, and his explicit distaste for the graphics tricks that every good engine makes use of to look good. He has picked the worst way possible to create the detail he's longing for, and it just not a good solution for where he wants to end up. It is pointless busywork for the artists; it doesn't scale; it's far more costly to un-approve (which he's fond of); and it's far more problematic to make modular, which they desperately need it to be. It's shades of Strike Commander all over again…

So he's not actually “sticking with that portugueser”. That's good news, at least.

What are the current expectations for the release date? Would like to know both from believers and heathens. Will be interesting to dig out screenshots of 'top of the line' games from as many years ago and compare to current. They probably should go with custom engine, even now, porting assets is not that big of a deal considering how many times they will have to refactor them. MS needed 3 years to get freelancer into a working game, dropping content but also redoing all the assets if you look at the early vids vs released, them coming up with: we change engine, we have to, see you in 3 years might be better for them, lying never works in the end, someone will capture your lifeless body being flung into a van

 

[ErikMaintainer]

10 REM The Best Game Ever
20 CLS
30 PAUSE 0
9999 REM it's still in alpha

The fidelity! The immersion! I just need a few more backers!

Take my money.

 

[1500]

So yet another major rework of the netcode. You would think they would have tried to do it right the first 3 times. Or at least not lied about it.

But CIG comes across as a company that

starts cutting wood
finds it does not fit
cuts some more wood
finds it does not fit
measures what they need
slap the first two peaces of wood in place
waves hands around
label it WIP, alpha-construction
move on to something else.


So yet another major release of the netcode, from a demo that only had 4 people playing in it at a time.

 

[Tippis]

Out of interest (I haven't been following it too closely recently) what professions are in 3.0? I'm guessing trading will be in, but is there any news on mining & bounty hunting?

More interestingly, how do they actually define “professions”? Are we just talking about them offering a couple of different mission templates or are they actually adding anything remotely resembling game mechanics and world dynamics?

 

[szopin]

More interestingly, how do they actually define “professions”? Are we just talking about them offering a couple of different mission templates or are they actually adding anything remotely resembling game mechanics and world dynamics?

They had 4-5 listed on a slide during gamescom, though pretty sure they consider bounty hunting as already in game (and probably exploration too, you can find grim hex after all)

 

[Asp Explorer]

Thing is - it's genuinely difficult to get networking right for large numbers of entities on a realtime basis. Like, really hard!

Especially with the consumer junk that most people consider internet connections. It's further complicated by all the tertiary services that are involved in large scale games these days in order to offset costs, but as more and more dependencies on 3rd party services happen, the less "real-time" stuff can be delivered.

 

[Osmanthus]

Well I ask same question 3 years ago....your friend Mr.Nowak said back then that this will be a kid´s play to fix and it´s a matter of weeks not month´s when we going to see hundreds of people per instance....

Guess what 3+ years forward and we are in the same place as we are been at the begininng .....16 players,not stable,laggy mess..with almost no improvements at all from the vanilla netcode and you want me to believe that 3.0 going to bring some"MAGIC" right?That we going to see such a huge improvement in the engine that it´s known for problematic netcode for ages?All games that are FPS based and required huge data-transfer between players(aKa netcode)in the CE didn´t pass that magic number of 24 ppl. per instance....to be honest 99% of games are been on 16 players per instance and even then we all know how bad they act in that multiplayer environment....

Just ask yourself why no one is using CE to build multiplayer FPS action games???The very same CRYTEK engineers could not fix or improve this problem for AGES and now Croberts magical skills will suddenly do that part???

Note that in the past a lot of people loved saying "weeks not months", but after many instances of CIG failing to deliver that once-popular saying is now completely ignored. It's a great way to hedge.

 

[Soliluna]

That wouldn't be much of a problem in and of itself though, since most of visual quality these days comes in the form of having intelligently made assets — the stuff people made the last-gen engines do at the end of their lives was still pretty spectacular compared to what the same engine did just a few years later. But that's where Chris' lack of experience starts to show, what with his reliance on pure geometry and similar brute-force methods, and his explicit distaste for the graphics tricks that every good engine makes use of to look good. He has picked the worst way possible to create the detail he's longing for, and it just not a good solution for where he wants to end up. It is pointless busywork for the artists; it doesn't scale; it's far more costly to un-approve (which he's fond of); and it's far more problematic to make modular, which they desperately need it to be..

Biggest point to take from his design and implementation choices for geometry and visuals in general is that he is limiting the potential consumers for the product. ED has the luxury of scaling down quite well - CryEngine has never been known for scaling well and looking at the assets in Star Citizen it won't scale well either.

He's unnecessarily hindering the potential sales of the game by not only focusing on the niche space game audience, but also the niche high-end PC audience.

 

[SkUnimatrix]

It's the common game development trap of not looking out of the window and seeing that stuff is actually coloured.

It's an odd philosophy, since an unrelenting colour scheme can make you feel fatigue and a desire to play something else with a livelier palette. (also applies to ED to be fair with it's orangyness, and darkness in general.)

What's wrong with Tandy Orange?

 

[szopin]

Biggest point to take from his design and implementation choices for geometry and visuals in general is that he is limiting the potential consumers for the product. ED has the luxury of scaling down quite well - CryEngine has never been known for scaling well and looking at the assets in Star Citizen it won't scale well either.

He's unnecessarily hindering the potential sales of the game by not only focusing on the niche space game audience, but also the niche high-end PC audience.

Hindering? Guess who can afford high end pcs? It was deliberately aimed at high spenders, again, marketing is not their weak point

 

[Osmanthus]

What's wrong with Tandy Orange?

I wouldn't mind if they allowed you to change HUD colour in the ingame options since it's right there in the config file anyway

 

[CMDR_Orlando]

Not really, no. It's all stuff we've seen a bajillion different times in a bajillion different shovelware games and movies. I mean, it's not really surprising considering how much we've seen them rip off existing stuff in the concept art that drives those designs — of course the final product will look like everything else since that's where they got it from.

Quite a few. Hell, there's an entire secondary marketplace on Steam that lets you do that these days. Not to mention games that are continuously being updated — games like, say, EVE online.

But again, that's not really the point here. The point is that SC is simply not at an alpha state yet. Even CIG says as much. The question is when will it be, because at the current pace, a proper alpha seems a very far way off. There's just too much missing, not just in terms of content and foundation, but in actual tools needed to do any of the testing that gives an alpha its purpose.

It's CryEngine 3 graphical quality, which is decent enough, but it's definitely starting to show its age. Not even CIG is dumb enough to fiddle with the single thing that CE3 has going for it: its rendering pipeline, so short of an complete engine replacement, it's already behind the top performers out now.

That wouldn't be much of a problem in and of itself though, since most of visual quality these days comes in the form of having intelligently made assets — the stuff people made the last-gen engines do at the end of their lives was still pretty spectacular compared to what the same engine did just a few years later. But that's where Chris' lack of experience starts to show, what with his reliance on pure geometry and similar brute-force methods, and his explicit distaste for the graphics tricks that every good engine makes use of to look good. He has picked the worst way possible to create the detail he's longing for, and it just not a good solution for where he wants to end up. It is pointless busywork for the artists; it doesn't scale; it's far more costly to un-approve (which he's fond of); and it's far more problematic to make modular, which they desperately need it to be. It's shades of Strike Commander all over again…

So he's not actually “sticking with that portugueser”. That's good news, at least.

Ofc you've seen it, it's sci-fi purposely made with very multiple recognizable mundane stuff, it's been done in every great sci-fi movie from Star Wars to Alien or Blade runner and it helps the viewer/player connect with it.

Yeah steam greenlight games aren't quite the same scope or development process as Star Citizen. This is a crowdfunded project which is quite different way of developing an Early Acess title.

Cryengine 3 is not the same engine that is being used for Star Citizen now. They started with 3.5 but they've already overhauled more than 50% of it, they quickly added PBR and multiple other stuff to improve graphics and performance. Remember that most of the things they are doing were never done with the engine and with every Cryengine upgrade they can hand-pick the pieces of code they want/need and integrate it, they have the technical expertise to pull that off if they want.

I was wondering when the Chris Roberts obsession would surface [rolleyes]. He is the CEO and the one that leads the vision but there's no need to blame him for everything, it's up to the graphics/technical programmers/engineers to decide what are the best ways to give the best graphical visuals within the performance limitations and you can bet that the original Cryengine programmers/engineers with multiple years of experience are doing just that.

They talk about it in their weekly shows, worth the watch the ones bellow:

[video=youtube;ko1vLmaqgqA]https://www.youtube.com/watch?v=ko1vLmaqgqA[/video]
>https://www.youtube.com/watch?v=WvlnG45Kois
>https://www.youtube.com/watch?v=j02lrXo3fXA
>https://www.youtube.com/watch?v=hp5AlLytcn0
>https://www.youtube.com/watch?v=1HYQADNOFyQ
>https://www.youtube.com/watch?v=0AzGT_ZcMHg

So he's not actually “sticking with that portugueser”. That's good news, at least.

Well considering that he was brought in 2012 to compose music for Star Citizen and keeps doing so I think "sticking with that portugueser" is quite apt.

 

[VidarSnipes]

I wouldn't mind if they allowed you to change HUD colour in the ingame options since it's right there in the config file anyway

It doesn't work properly, and as far as I remember making it work properly is a surprisigly large amount of work. By the way I think the ability to manipulate HUD colours through the config file was enabled somewhere in the middle of 2015 and there were few iterations of the system - early on, the reflections of HUD elements were baked in on the interior textures and their colours were independent from the HUD itself.

 

[LeSabre]

Well I ask same question 3 years ago....your friend Mr.Nowak said back then that this will be a kid´s play to fix and it´s a matter of weeks not month´s when we going to see hundreds of people per instance....Guess what 3+ years forward and we are in the same place as we are been at the begininng .....16 players,not stable,laggy mess...with almost no improvements at all from the vanilla netcode and you want me to believe that 3.0 going to bring some"MAGIC" right?That we going to see such a huge improvement in the engine that it´s known for problematic netcode for ages?All games that are FPS based and required huge data-transfer between players(aKa netcode)in the CE didn´t pass that magic number of 24 ppl. per instance....to be honest 99% of games are been on 16 players per instance and even then we all know how bad they acting in that multiplayer environment....Just ask yourself why no one is using CE to build multiplayer FPS action games???

I just enjoy the "3.0 will fix the netcode, and then all the other features will simply switch on and everyone will see how great it all is" mentality. They've all forgotten that the netcode isn't holding back an awful lot of game, the lack of development of the game is what's been holding everything back

It's taken at least three netcode reworks to get it to the point it's at now, I'm betting the netcode rework this time around solves very little as per usual. 'Netcode' and 'rework' are this seasons "one hit fix for all" buzzwords, nothing more, much like how they missuse the word 'refactor' as a buzzword rather than honestly say 'rewrite'. Still, they might at least attain the heady, neigh lofty heights of 20+ players in an 'instance' with some actual stability this time around. Long way to go for 1000 though, and if they knew their engine at all they'd know it would be totally impossible to display that many in a single session in high detail anyway unless the end users were all rocking as yet undesigned biblicly powerfull pc's and a lot more than 200MB-up and 100MB-down bb lines, but gotta keep the public dreaming, and spending.

Network rework = Q4 2016 placebo. Be sure everyone takes their sugar pills to prepare for the incoming stupidity.

 

[Osmanthus]

Ofc you've seen it, it's sci-fi purposely made with very multiple recognizable mundane stuff, it's been done in every great sci-fi movie from Star Wars to Alien or Blade runner and it helps the viewer/player connect with it.

Yeah steam greenlight games aren't quite the same scope or development process as Star Citizen. This is a crowdfunded project which is quite different way of developing an Early Acess title.

Cryengine 3 is not the same engine that is being used for Star Citizen now. They started with 3.5 but they've already overhauled more than 50% of it, they quickly added PBR and multiple other stuff to improve graphics and performance. Remember that most of the things they are doing were never done with the engine and with every Cryengine upgrade they can hand-pick the pieces of code they want/need and integrate it, they have the technical expertise to pull that off if they want.

I was wondering when the Chris Roberts obsession would surface [rolleyes]. He is the CEO and the one that leads the vision but there's no need to blame him for everything, it's up to the graphics/technical programmers/engineers to decide what are the best ways to give the best graphical visuals within the performance limitations and you can bet that the original Cryengine programmers/engineers with decades of experience are doing.

They talk about it in their weekly shows, worth the watch:

https://www.youtube.com/watch?v=ko1vLmaqgqA

>https://www.youtube.com/watch?v=1HYQADNOFyQ
>https://www.youtube.com/watch?v=0AzGT_ZcMHg

I don't think you understand the difference between making a lived-in universe, like what they did with the early Star Wars films, along with Alien etc, and making everything look skullcrushingly boring

 

[McGlashan]

Well I ask same question 3 years ago....your friend Mr.Nowak said back then that this will be a kid´s play to fix and it´s a matter of weeks not month´s when we going to see hundreds of people per instance....

Guess what 3+ years forward and we are in the same place as we are been at the begininng .....16 players,not stable,laggy mess..with almost no improvements at all from the vanilla netcode and you want me to believe that 3.0 going to bring some"MAGIC" right?That we going to see such a huge improvement in the engine that it´s known for problematic netcode for ages?

It doesn't matter, CitizenCon will come along and show off a bunch of new, purpose-built demos, of stuff that they'll claim will be in the next major release. Nonsense like farming probably. So that when 3.0 does the MVP dance and finally appears with the absolute minimum improvements possible while still technically checking the boxes for things like planetary landing and professions, and while still being just as vacuous and broken as the current version, the blinkered cultists will claim it's a fantastic success and attempt to deflect any criticisms by pointing at the shiny things lined up for the next update. It's never the next version that's the game changer in Star Citizen World, it's the one after that. When 2.6 hits (and why are they even bothering with 2.6 if 3.0 is supposed to appear in the next 3 months?) they need to have a 3.1, or 4.0 already on the horizon.

 

[szopin]

I just enjoy the "3.0 will fix the netcode, and then all the other features will simply switch on and everyone will see how great it all is" mentality. They've all forgotten that the netcode isn't holding back an awful lot of game, the lack of development of the game is what's been holding everything back

It's taken at least three netcode reworks to get it to the point it's at now, I'm betting the netcode rework this time around solves very little as per usual. 'Netcode' and 'rework' are this seasons "one hit fix for all" buzzwords, nothing more, much like how they missuse the word 'refactor' as a buzzword rather than honestly say 'rewrite'. Still, they might at least attain the heady, neigh lofty heights of 20+ players in an 'instance' with some actual stability this time around. Long way to go for 1000 though, and if they knew their engine at all they'd know it would be totally impossible to display that many in a single session in high detail anyway unless the end users were all rocking as yet undesigned biblicly powerfull pc's and a lot more than 200MB-up and 100MB-down bb lines, but gotta keep the public dreaming, and spending.

Network rework = Q4 2016 placebo. Be sure everyone takes their sugar pills to prepare for the incoming stupidity.

Ppl expecting a magic fix are in for a ride, 2.6 will have a few enhancements, 3.0 will increase the number by a bit, it IS a major factor that they need to address, half expecting them to end up buffing their recommended specs to 1Gbit/s and anything lower being unsupported/expect weirdness, they already said console peasants are unwelcome, strayans will be next (nothing to do with customer protection laws guys)

 

[lysan]

Biggest point to take from his design and implementation choices for geometry and visuals in general is that he is limiting the potential consumers for the product. ED has the luxury of scaling down quite well - CryEngine has never been known for scaling well and looking at the assets in Star Citizen it won't scale well either.

He's unnecessarily hindering the potential sales of the game by not only focusing on the niche space game audience, but also the niche high-end PC audience.

Draw distance in ED is really good too, I can see the bases from very far away ! That is impressive.

 

[LeSabre]

Cryengine 3 is not the same engine that is being used for Star Citizen now. They started with 3.5 but they've already overhauled more than 50% of it, they quickly added PBR and multiple other stuff to improve graphics and performance. Remember that most of the things they are doing were never done with the engine and with every Cryengine upgrade they can hand-pick the pieces of code they want/need and integrate it, they have the technical expertise to pull that off if they want.

Except that if you look at anything that uses Cryengine 3.5 it looks 10x better than anything CIG have released at gameplay level, in fact the only things they've managed to make look 'like' or slightly better than 3.5 have been either still images or very, very heavily railed movie-esque sequences for lightening wallets. I don't get why people think they've done anything visually better than other CE projects, the PU is a great example of "doesn't look anywhere near as good as they claim" purely on a graphical side, maybe I'm being very picky or have been looking at too many other high detail projects lately, but the graphics to me in the PU are dated as heck.

Until they can match their own very earliest ship comercials (2012?) for graphical quality (nope, they've produced nothing playable that matches them) then they haven't made full use of the only thing CE is really good at.