|
Thx for posting this. However, this video occurred 50+ hours ago. The video I posted is 12 hours old and the result of Hero Marine playing and testing the game. The video I posted are HeroMarine's conclusions. The video you posted is his brief cursory look at things.
I'd have to say the video I posted offers Heromarine's deepest insights into the game.
I gotta say Nina is really holding it down for Stormgate on Twitch.tv. if Frost Giant creates a fictional UNN like they did in SC2 they should make Nina the anchorwoman of the News Station.
The interplay between Kate Lockwell and Donny Vermillion was epic.
|
i gave it another 5 hours after the patch. it definitely is better but it's just not good enough to take my time from other games and there are plenty of great games
|
I think the new patch is a very good improvement over the old and for the most part also the health of the game.
The new lighting helps a lot and maybe it's only due to my poor eye sight and the new lighting / etc helping but it honestly feels as if the unit movement feels a little smoother also.
Overall I think it is definitely a positive improvement for the game.
|
Northern Ireland25342 Posts
On September 22 2024 11:36 Fango wrote: The performance problems are the only thing that makes me really worried about the game. Everything else is improving with each update. I think it's absurd that I can enjoy triple A titles like Cyberpunk and Baldurs Gate 3 without issue, but a Stormgate 1v1 can't even approach a max out on the lowest settings. The instant units shoot eachother it effectively stops working.
You just can't have a f2p title, making effort to bring RTS back to casual players, and then only elite PCs can run it. You're eliminating entire demographics. So much for younger players or nerdy students playing in their dorms.
Additionally, even RTS diehards will not necessarily have a powerful PC. As a genre it's very light light on spec requirements. The same could be said for competitive game fans, Dota, LoL, Rocket League, Counterstrike, SC2 etc all run on just about anything.
Absolutely in terms of specs. Plenty of diehards play on lower specs for visibility and performance anyway. In general RTS games need quite low ones, I certainly haven’t found many that push things.
I don’t have things like a 4K display or a high refresh rate monitor so even playing more graphically intensive modern games, my rig is fine I don’t really need to push it.
Hopefully they can make big performance improvements and soon, it’s really limiting their potential penetration, and I’d wager quite a lot of the demographics they’re targeting don’t have high end gear
One of my Alma maters threw down a good 3500-4000 per PC for a full eSports cafe, if they’d consulted basically anyone who actually runs eSports events we’d have told them it’s insane overkill for the current crop of competitive games. Just as a ridiculous aside
|
As has been mentioned (by Frost Giant), apparently true performance optimization can't be done until the game is mostly "feature complete". As I guess a lot of underlying code is still actively being developed, it can't or it doesn't make much sense to do optimizations at this stage (unfortunately).
When you watch that SC2 Alpha development video, the last 3 years are optimizations and campaign. 12,000 builds of the game from most of the design being done in 2007 until the launch in 2010. Video here for those who have missed it (optimization discussion starts around 5:50):
+ Show Spoiler +
HeroMarine's video is basically: the graphics look nice, the game needs fully customizable hotkeys, and many units seem copy/pasted and need more coolness/creative factor.
He's right that from a tech tree point of view, the one unit that doesn't really "fit" in with Starcraft Terran is the Lancer. (Although you could say it's similar in function to a firebat/hellbat - melee that does a small aoe. Thematically it looks/feels unique).
Of course, most of the vanguard units aren't "direct' copy and pasted, as they all have some functionality that is different. (medtech's spells are different than medic, a hedgehog sieges for anti-air, atlas can leave a burning ground effect, etc. etc.).
Hopefully more stuff is continually added into the game over time. I am loving the patch, but was hoping for a new unit or two or something being redesigned/added functionality. Hoping that's the case in the patches after the October 3v3 release.
|
Northern Ireland25342 Posts
On September 22 2024 22:43 SoleSteeler wrote:As has been mentioned (by Frost Giant), apparently true performance optimization can't be done until the game is mostly "feature complete". As I guess a lot of underlying code is still actively being developed, it can't or it doesn't make much sense to do optimizations at this stage (unfortunately). When you watch that SC2 Alpha development video, the last 3 years are optimizations and campaign. 12,000 builds of the game from most of the design being done in 2007 until the launch in 2010. Video here for those who have missed it (optimization discussion starts around 5:50): + Show Spoiler +https://www.youtube.com/watch?v=z4lrG1QD68M HeroMarine's video is basically: the graphics look nice, the game needs fully customizable hotkeys, and many units seem copy/pasted and need more coolness/creative factor. He's right that from a tech tree point of view, the one unit that doesn't really "fit" in with Starcraft Terran is the Lancer. (Although you could say it's similar in function to a firebat/hellbat - melee that does a small aoe. Thematically it looks/feels unique). Of course, most of the vanguard units aren't "direct' copy and pasted, as they all have some functionality that is different. (medtech's spells are different than medic, a hedgehog sieges for anti-air, atlas can leave a burning ground effect, etc. etc.). Hopefully more stuff is continually added into the game over time. I am loving the patch, but was hoping for a new unit or two or something being redesigned/added functionality. Hoping that's the case in the patches after the October 3v3 release. There’s probably a reasonable middle ground between having it run smooth as butter on a potato rig, and choppy performance even on decent rigs. Indeed, some better rigs on paper may run it much, much worse.
This isn’t an issue on a game that’s internally developed until it’s oven-ready, then released.
Stormgate is seemingly reliant to whatever degree, and opinions vary, on building hype and some revenue in Early Access. Which, if performance is a pretty widespread problem currently, needs at least some redress
They chose or were forced to choose this business model, they have to adapt to current realities to some degree, rather than pointing to other products with a different development lifecycle.
Let’s say hypothetically we don’t see notable performance increases in a year, or two years. How many folks realistically are going to come back excitedly for a 4th, 5th time or whatever now that performance is finally decent?
And I’m not an engine optimisation guy, but the core gameplay does seem pretty feature-complete it just lacks optimisation. If they want to add a bunch of graphical bells and whistles, that’s still not really going to impact on what’s killing performance, which is the CPU-bound game logic
They’ve got pathing issues, that needs tweaked. So maybe they’ll have to redo more than us plebs assume
But overall, responsiveness is good, pathing is good (except when it isn’t). Netcode in general is pretty good, and rollback also works well (when people can use it), but performance is bad.
That to me screams ‘basically feature complete but needs optimised’
Your 1v1 module works pretty well, your 3v3 PvE suffers from performance problems. You’ve cut the supply limit for your upcoming 3v3 PvP mode to try and mitigate this.
Which again screams out that features fundamentally work, but their performance is the problem, so it’s time to optimise.
The only upcoming features that relate to the game itself that aren’t yet in the game, or coming next big update are the editor, more campaign stuff and some units for PvP
Now to me, and I stand to be corrected by more knowledge folks, I’m not seeing huge roadblocks, or big upcoming features that mean you have to hold off improving performance.
|
United Kingdom20285 Posts
That to me screams ‘basically feature complete but needs optimised’
* Campaign design and interface (it's currently full of placeholder stuff, hacky stuff between levels, missing key features like meta-progression and multiplayer)
What i mean by meta-progression is stuff like choose-your-own-adventure unlockable unit upgrades, tech, mercenaries and mission order in the WOL campaign. They are a critical part of campaign IMO because they allow you to customise things to your specific wants, which helps the fun aspect a lot. As you can't do everything all at once it also adds greatly to the replayability of campaigns - for example you can say that i'm gonna save money and rush to Bel'shir and The Dig, then upgrade your goliaths + tanks and use them as your core while on a different playthrough you might go to other planets, research different techs, and buy different mercenaries and upgrades. SC2 really nailed this. The more the merrier, it makes great and popular youtube content as well for creators like GGG which draws clicks from inside and outside of the communities.
It is not really an issue with the campaign being feature complete and there just only being 3 missions - the 3 that are there and the framework for them is really hacky and missing features.
* Custom map editor and ways to explore and join games in the client
* 3v3 hero competitive mode
* Friends list, block list
* Report functionality and some kind of moderation system (right now it's 100% manual, if somebody is spamming n****er in the chat you have to just watch it the whole game and send pictures to a staff member who will go into their backend and ban the player - that can't scale, at least without some changes to make it require less work). Yes you can disable all chat, but if the advice to everybody on a feature-complete "first truly social RTS" was to disable all social features indefinitely then they might as well pack it up and go home. It's obviously mandatory for the criteria.
* Clans
* Tournament systems
* Onboarding experience
* Live-replay spectate, rather than having casters and viewers ingame as a player and slowing down the player's game as they have to right now - this exacerbates performance problems
* Replay / game update system that can preserve replays across multiple versions - right now every time they change anything at all in the game, every replay becomes unwatchable forever which is not really acceptable. SC2 broke past replays once IIRC, i think you can still view them with an old copy of the game (while you can't in stormgate) and i think it was a big mistake to do that one break.
* Probably the ability to handle performance declines gracefully e.g. smooth time dilation. Performance declines on the computer of somebody in the game WILL happen no matter what the performance level is and you wouldn't want random lurches in time and dropped inputs when this happens.
All of this is on the list for feature complete, some of it doesn't have dates and some isn't due for a year. In my opinion it's pretty much all more important than competitive 1v1 for both the health and monetisation of the game.
|
On September 22 2024 22:43 SoleSteeler wrote: As has been mentioned (by Frost Giant), apparently true performance optimization can't be done until the game is mostly "feature complete". As I guess a lot of underlying code is still actively being developed, it can't or it doesn't make much sense to do optimizations at this stage (unfortunately). I understand that focusing on optimisation is not ideal when you're still early on development.
The problem is that the level is beyond just feeling not optimised. Systems that run any triple A title, even the heavy ones, breaking when trying to handle a single max out army, is a foundational problem.
As far as I'm aware, this is not a demo version of the game for them to work on gameplay and design while the real engine is being built. It needs to run 5-10x better for 1v1 to be popular, and probably triple that if they want 3v3 and custom maps to be popular. Can a game optimise itself that much?
On September 23 2024 01:39 Cyro wrote: All of this is on the list for feature complete, some of it doesn't have dates and some isn't due for a year. In my opinion it's pretty much all more important than competitive 1v1 for both the health and monetisation of the game. None of what you list a gameplay/engine feature. And that's the issue in terms of performance. Obviously there are countless UI, gamemodes, and social features yet to be added
And yes, performance is more important than everything you've listed. If all that was added but there wasn't serious performance changes, I still wouldn't be able to play the game
|
Northern Ireland25342 Posts
On September 23 2024 01:39 Cyro wrote:* Campaign design and interface (it's currently full of placeholder stuff, hacky stuff between levels, missing key features like meta-progression and multiplayer) What i mean by meta-progression is stuff like choose-your-own-adventure unlockable unit upgrades, tech, mercenaries and mission order in the WOL campaign. They are a critical part of campaign IMO because they allow you to customise things to your specific wants, which helps the fun aspect a lot. As you can't do everything all at once it also adds greatly to the replayability of campaigns - for example you can say that i'm gonna save money and rush to Bel'shir and The Dig, then upgrade your goliaths + tanks and use them as your core while on a different playthrough you might go to other planets, research different techs, and buy different mercenaries and upgrades. SC2 really nailed this. The more the merrier, it makes great and popular youtube content as well for creators like GGG which draws clicks from inside and outside of the communities. It is not really an issue with the campaign being feature complete and there just only being 3 missions - the 3 that are there and the framework for them is really hacky and missing features. * Custom map editor and ways to explore and join games in the client * 3v3 hero competitive mode * Friends list, block list * Report functionality and some kind of moderation system (right now it's 100% manual, if somebody is spamming n****er in the chat you have to just watch it the whole game and send pictures to a staff member who will go into their backend and ban the player - that can't scale, at least without some changes to make it require less work). Yes you can disable all chat, but if the advice to everybody on a feature-complete "first truly social RTS" was to disable all social features indefinitely then they might as well pack it up and go home. It's obviously mandatory for the criteria. * Clans * Tournament systems * Onboarding experience * Live-replay spectate, rather than having casters and viewers ingame as a player and slowing down the player's game as they have to right now - this exacerbates performance problems * Replay / game update system that can preserve replays across multiple versions - right now every time they change anything at all in the game, every replay becomes unwatchable forever which is not really acceptable. SC2 broke past replays once IIRC, i think you can still view them with an old copy of the game (while you can't in stormgate) and i think it was a big mistake to do that one break. * Probably the ability to handle performance declines gracefully e.g. smooth time dilation. Performance declines on the computer of somebody in the game WILL happen no matter what the performance level is and you wouldn't want random lurches in time and dropped inputs when this happens. All of this is on the list for feature complete, some of it doesn't have dates and some isn't due for a year. In my opinion it's pretty much all more important than competitive 1v1 for both the health and monetisation of the game. Which all becomes moot if the performance is crippled. They’ve already altered supply caps for 3v3 PvP’s upcoming launch, and it’s not for any game design reason.
I imagine there is overlap here with core engine stuff. Replays and live observing would seem two to me. But most don’t seem to have a huge overlap.
Others seem moot if the game isn’t performing well, especially given they made a point to go 64 tick, and implement rollback.
Unless FG are planning major engine overhauls, which they have not said, it would seem now is the time to make some performance improvements. Even if it’s a parallel process and they’re also allocating resources to add the kind of content the game needs
It doesn’t have to be Earth-shattering, merely some noticeable improvement, sooner rather than later
|
|
I'm not trying to defend the performance, only just listing what has been said, and also showing SC2's development as a past example.
If only 10% can run the game smoothly the game will never grow.
I'm watching Chickenman's VOD of his tournament last night and his PC can barely run the game. Like 10 FPS in mid-game fights and major sound issues. Here's a link for an example (wait till the spriggan attack starts). Too bad, as this match itself was pretty great!
I am certainly in agreement that there's no point for 64 Hz and rollback if performance is bad on the majority of machines.
|
SC2 ran on a potato from day 1 of their 4.5 month beta test. SC2's game engine is an impressive piece of software engineering. I am sure UE5 is also an impressive piece of software engineering. THe problem is UE5 is not made for RTS games.
If anything they should've gone with UE4. I've seen small teams perform miracles with UE4. Perhaps this is why ZeroSpace is on UE4?
BTW, Rise of Nations Extended Edition is on sale on Steam for $4. That's less than a cup of coffee at Starbucks.
|
There's certainly a benefit to 64 tick as can be seen in other games. Higher tick servers are so much smoother and responsive. That being said, if it causes low FPS, input lag, and wacky time dilation, you've flown too close to the sun. You've just made it less smooth and responsive than if you stayed on 16 tick.
Rollback seems to work well, at least it's optional so those playing cross server or with weak connection can choose to do so. I don't see any downsides to rollback in this implementation.
On another point, HeroMarine's comments on hotkeys are fair. They'll enhance hotkeys with future update, moving off a grid system and adding things like Steal+Add. Hell, even a preset option for "RTS classic" or something, basically Starcraft hotkeys, would be useful for some.
His comments on identity are very subjective. But a lot lies in Starcraft having a massive history and campaign. There's no way you can make a spiritual SC2 successor and not have a race with marines, tanks etc. But there's no reason why various Stormgate units can't feel as cool as a zealot or siege tank if you played with them for years. There's a few I'm already a big fan of (Weavers, Imps, Spriggans, Kri) and that's without any lore material at all
|
United Kingdom20285 Posts
Snowplay is where the performance problem is, and it's not a part of UE5. The ingame metrics prove it
|
I managed to get Stormgate into a manageable state by using 'undervolting'. Try it at your own risk.
If you go into a custom and lower the graphics and just sit and look at the Shroud detail, you'll understand why the game runs so poorly. It has particle effects and has constant animations going on with it. Even trees and grass are continuously swaying , which I suppose is normal for other games... but in an RTS you expect low settings to be low settings.
Consequently, very few PCs can handle the load in 200+ Infernal fights. A bit surprised it made it past the base development stage. They need to redesign their low graphics settings to make the game more accessible.
|
On September 23 2024 12:24 Sameday wrote: ... A bit surprised it made it past the base development stage. ... Just another case of major mismanagement, nothing new with FGS i guess.
|
On September 23 2024 12:24 Sameday wrote: I managed to get Stormgate into a manageable state by using 'undervolting'. Try it at your own risk.
If you go into a custom and lower the graphics and just sit and look at the Shroud detail, you'll understand why the game runs so poorly. It has particle effects and has constant animations going on with it. Even trees and grass are continuously swaying , which I suppose is normal for other games... but in an RTS you expect low settings to be low settings.
Consequently, very few PCs can handle the load in 200+ Infernal fights. A bit surprised it made it past the base development stage. They need to redesign their low graphics settings to make the game more accessible.
I'm no expert but that shouldn't need computing power. That's like a endless gif playing , nothing more
|
Northern Ireland25342 Posts
On September 23 2024 18:39 Harris1st wrote:Show nested quote +On September 23 2024 12:24 Sameday wrote: I managed to get Stormgate into a manageable state by using 'undervolting'. Try it at your own risk.
If you go into a custom and lower the graphics and just sit and look at the Shroud detail, you'll understand why the game runs so poorly. It has particle effects and has constant animations going on with it. Even trees and grass are continuously swaying , which I suppose is normal for other games... but in an RTS you expect low settings to be low settings.
Consequently, very few PCs can handle the load in 200+ Infernal fights. A bit surprised it made it past the base development stage. They need to redesign their low graphics settings to make the game more accessible. I'm no expert but that shouldn't need computing power. That's like a endless gif playing , nothing more Such particle effects to my knowledge are calculated on the fly, rather than just playing some stored animation, most of the time anyway. I’m actually unsure of it that is sometimes, it feels you don’t always need dynamic particle interaction and a much less taxing stock animation might do the job fine. Perhaps someone may educate me!
I love Pillars of Eternity, but on my ancient rig that game would perform smooth as butter into slideshow once I advanced a bit further and there was a lot going on in terms of particle effects.
Although to my knowledge they’re kinda the domain of a GPU, I had a shit one ergo bad performance.
That said I haven’t heard anyone who’s more technically savvy than me find the GPU be the culprit for bad performance, it’s always the CPU that’s the bottleneck
|
United Kingdom20285 Posts
Animated and dynamic models (e.g. stuff which is not always present) are often heavier on the CPU than static ones, but depends on engine stuff. There must be some load with the destructable trees because so many of them can be destroyed individually and their state influences actual gameplay. In SC2 for comparison a destructable rock was basically just a single fat neutral unit, and graphical effects like things being on fire were usually entirely client side and not part of the gameplay simulation at all.
Sim time is the biggest problem for high end systems and for other players on weaker PC's impacting your gameplay experience. The other stuff is quite heavy but playable - you can get a good baseline for how it runs by playing on an empty map with a few units. The sim time will increase as you build more units, and that per-unit cost is what really gets out of control.
|
On September 23 2024 21:48 Cyro wrote: Sim time is the biggest problem for high end systems and for other players on weaker PC's impacting your gameplay experience. The other stuff is quite heavy but playable - you can get a good baseline for how it runs by playing on an empty map with a few units. The sim time will increase as you build more units, and that per-unit cost is what really gets out of control. 100s of dead bodies created performance issues for the Borderlands games built with UE3 and UE4. They mitigated it by having a setting that allowed dead bodies to disappear after a user configured # of seconds.
On September 23 2024 08:35 Cyro wrote: Snowplay is where the performance problem is, and it's not a part of UE5. The ingame metrics prove it If UE5 requires additional tech like Snowplay then I question FG's decision to go with UE5. No one is using UE5 for RTS so there is no pool of super experienced outside people who can parachute in and help them out.
These things make Stormgate a difficult project to rescue.
|
|
|
|