I deleted TSW. DTG wake me up when you finally fixed those stupid blurry textures. I played TSW last time 2 months ago. So i decided to play it today. 1st service was a freight train from Dresden to Röderau with the BR145. After 30mins the textures turned into a complete blurry mess again. Instantly deleted it. I'm sick of this BS. How long should we wait till this is finally fixed???
I hear you I un-installed it at the end of March. The problem with the blurry textures I think has more to do with mounting issues. Why does it have to load up an entire route just for 1 train? So with the layers if you are using one that isn't in the core route its loading the entire 2nd route. Well there is why you are having so many issues why load 10gb instead of just 3gb. Makes no sense, the game was bound to have these issues with that setup after 3 or 4 years of DLC.
Whilst I haven't yet deleted TSW itself, I did delete a bunch off add-ons I barely play, and man it loads so much faster now You'll be waiting forever probably.
I can somewhat understand this decision. It mainly comes from the lack of a strategy for how we as users are informed about progress. In my view, a long‑term strategy is needed that outlines which steps will be taken and, most importantly, a clear goal on the horizon. Console players suffer a lot from this issue, and nothing is more frustrating than not knowing what to expect in terms of strategy like when we can expect a major step forward. So far, it’s mostly been a lot of “we’re working on it,” but noticeable progress is missing, even though DTG gives a slightly different impression. And yes, I understand that big steps don’t happen overnight, but we’re years down the line now and what do we actually know? “We’re working on it” is basically the phrase you could put on a tile and hang above the toilet. And no, this isn’t meant to bash DTG or anything like that it’s more of an “our” point of view, and the fact that DTG unfortunately shares very little is a valid point of criticism. “It could be better” is probably just as worthy of being printed on a tile for above the toilet.
ive uninstalled all but my UK DLC to help loading times and overall help TSW perform. i would LOVE to just uninstall it but what else is there to sate my desire to drive trains (on playstation, my PC isnt powerfull enough for running train sadly). DTG have our privates in a bench vice right now and will just use JT or Firefly to try and trick us into thinking TSW is better
What's wrong with supplying two .pak files for each route? One for the route abd scenery etc, and one for the trains? (This is a genuine question btw)
thats WHY the game is in the state its in, if you want the Class 484 on the GWE remaster, the entire island line 22 is loaded for the one train. you want the 350/1 on GWE then thats the WCML South for just one train. etc etc etc. Matt himself has explained that this is how the DLC mounting works, i think some other people have also spoken on it but it is indeed how most of TSW stuff works. a few exceptions being things like the Scotrail 158 and DRS Class 66, i also belive the Class 390. MOST of the layering system does need to call an entire route DLC to mount for a single train. Id say it was dumb, but this is dovetail were talking about. theyve done sillier things than this
The tricky part is that there is no core that includes the trains. Or only a small portion of it. That would save so much performance, because you wouldn’t need to load the entire route. Having all trains in one pool is, in my opinion, a solution but yes, you’re selling them separately, and that’s exactly the issue. You would have to separate trains and routes in the way they’re sold. So you either buy a train, which goes into your core game, or you buy a route, which then pulls the trains from the core. This is something many other games already do.
OR you have the route and train as a separate .PAK file, then the train and route can still be separate, wont need to bloat core by having stuff packed into it to the brim and when a route needs a specific train, it loads that trains PAK and doesnt need to load the entire routes PAK
Yes, that’s what you get when you separate the two. So the route is sold separately, and the train as well, and both would then indeed have their own .pak file. Or you have to seperate files in 1 buy that seems the same system yeah. I just don’t know whether loading many different trains would end up giving the same result. After all, you still need to load them into the route, which probably isn’t the perfect solution either. So it might only fix a very small part of the problem.
I'm no computer engineer but even this makes sense to me conpared to haow they do it. The placing of trains to the core is also an idea. I suspect though the latter may cause much longer load times but I'd be willing to make that sacrifice. But thats why it seems to only affect routes that the timetable has alot of different layers and not ones that don't. Look at Cathacart as an example if you run the default time table no blurry textures as majority if trains are in the route. But with the 380 timetable now you have to load the main route, the 380, the 158, ECML for the 801 AI, WCML for the 390 and 350 AI and Edinburgh Glasgow for the 385 AI. If you only had to load up the trains for those 3 routes I bet there wouldn't see as many blurry textures. And if you remove those 3 routes layers on the route it runs fine with no blurry textures. Same with BCC the original is fine but the 170 requires I think 3 or 4 routes and again blurry textures.
With the Birmingham Cross City as an example, you would need to load the Class 37, 43, 66, 170, 220, 350 and 390. Lets say each train is 1GB in size on its own, that would be 7GB for all the layered trains RATHER than loading the entire riviera for the 220, the entire WCML south for the 350, the entire GWE remaster for the 43, etc etc etc. Yes, you would still be calling up DLC to mount, but it would be WAY less than if the entire route was needing to be called. you might find with more individual files the game at large takes longer to first boot but i would take that over blurry textures ANY day of the week
some trains are already a separate .PAK to the route they come with, granted thats due to them being a separate DLC but still. You can use the Class 390, SR Class 158, Class 380, DRS Class 66 and i *think* one or two others without needing to have the route they go to installed. it is doable.
As far as separating routes and trains, that's been brought up before and addressed by Matt. Basically, the problem is that if a single version of a separate train is used as a modular addition to whichever routes it goes with, then any change to that train has to be tested in every route it touches. Keeping routes and trains together and self-contained means testing only has to be done in that closed ecosystem. As always, there are trade-offs either way, but DTG had to make a choice and they made one.
The problem with blurry textures started in the TSW5 era. While it's nice to see a lot of different trains on a route, it seems to be the game nowadays has become overloaded. Something not mentioned, are substitutions, with only an option to turn all of them off. I really want to see and option to turn off specific substitutions, good for performance and I also do not want to see some locos on specific routes. Another thing is the amount of versions of the same train you can choose from. I don't know how much this is an impact on blurry textures or mounting issues, but do we really need to be able to choose the dosto from Dresden - Riesa on the Koln - Aachen route or Salzburg - Rosenheim for example? Or choose all of the Class 66 versions on any UK route. On this moment I prefer to play TSW3 and 4 routes and some routes from TSW5. The GWE remaster gives me already blurry textures once I enable the class 350 layer. As the problems are still not solved after more then one and a half year, I just don't buy the newer stuff, unless I'm sure the product is not affected.
Why on earth would you need to sell them separately? It would just be one product containing two individual .pak files.
This explanation doesn't make any sense to me. Most trains appear as layers in other routes, so whether the train is included in the route's .pak file or in its own independent .pak file surely makes no difference! You'll still need to test it works with the other routes it layers into.
I'm pretty sure what Matt said was regarding having only one version of each train/loco instead of multiple ones coming with different routes. If you just split a route into route pak file train pak file this shouldn't be a point. Though I'm not sure what that would mean regarding the mounting issues on PS5. Would needing to load more paks there further worsen the issue?
From what I've been able to find reading about how DTG has implemented asset loading, it stems from how DTG package there assets, the wrap everything into one pool, like say the WCMLos has all the couches and locos in the same pak file, the engine has to go into every single pak file to get each layer, it does not load the scenery tiles only the loco or coach models called, the more layers the more files it has to access the slower it is, what DTG should have done (I'm not a software engineer, so this is my opinion based on stuff I've learned) is basically have say a global object pool where all stock assets would sit, so when route A calls for say a class 87, it only has to go there instead of having to look through several pak files to find the class 87 (example i know that route doesn't contain it) from blackpool branches to then place into WCMLos e.g. 1) [Low Memory Footprint] 100 services running the EXACT same commuter train --> Engine loads 1 set of textures and 1 physics model, then clones it. 2) [High Memory Footprint] 10 services running 10 DIFFERENT layered trains from 10 different DLCs --> Engine must load 10 massive texture sets, 10 cab interiors, and 10 audio banks. No 1 is the ideal, and no 2 is how DTG does it atm TSW treats every asset as unique even if it's the same train from a different DLC and UE4 can't share data easily between assets so it basically chokes itself, causing slowdown and stuttering, not even considering that we may see say the class 66 as the same loco, but each one is different based on what DLC it was released with, like different liveries, lighting and audio sets, to us it's the same to the engine it's a new asset so it basically chokes itself. Like i said this is just what I've learnt researching, I'm not a software guy I'm a mechanic who likes to read, so take what's above as my interpretation of stuff about how DTG use UE4 as speculation. Some links ive read (heavy stuff) https://gamedevtricks.com/post/origin-rebasing-space/ https://forums.unrealengine.com/t/world-partition-streaming-performance/2577446 https://www.xda-developers.com/tire...eres-how-unreal-engine-is-fixing-the-problem/ https://forums.dovetailgames.com/threads/playstation-5-add-on-limitation-explained.48903/ https://dev.epicgames.com/documentation/unreal-engine/world-composition-in-unreal-engine
When is Matt supposed to have said that? Because in TSW, we have exactly the opposite of what Matt is supposed to have said. Take a look at the German routes. I’ll just give you an example here. We have the German routes A, B and C. And the same locomotive is on each of these routes, because exactly the same locomotive has been assigned to all three routes. But all the locomotives on these three routes look the same, even though they're completely different from one another. But if I now want to drive route D and select that locomotive there, a list of all the locomotives installed on all these different routes appears on the right. And it’s actually quite difficult to know which locomotive to choose, as they all look the same but are completely different. One locomotive from route B has had an update, and the other locomotive from route C has been completely overhauled and modified in countless ways. And the locomotive from route A has never received an update. And this selection usually contains not just 3 locomotives, but 6 to 10 locomotives, so which one do you want to drive now? Because no user knows exactly which locomotive is the best. That’s why jivebunny suggestion is really good. DTG should start selling a route, but create a route .pak file for it, and a .pak file for each locomotive. Because then we wouldn’t have these problems with all these different locomotives that are all the same. But everyone is different anyway.
Let me try to explain clearer what I meant. There are three options: 1) The route is sold together with a base set of vehicles belonging to the route, all included in one PAK file. This is how it is done currently (I think some freight cars where recently moved to core, but that is a special case). 2) Again, the route comes with it's base set of vehicles, but they are delivered/downloaded as two or more PAK files (one for the route, one for the trains; or even separate ones for each train). But you would still have multiple versions of the same vehicle (class 66, BR185/146, ...) with the different routes. 3) Only have one versions for each loco, either directly in core or as a standalone PAK. It might still be sold together with a route, so if you buy the route you unlock the loco. But the important point is that there is only one version across the entire game. What Calidore mentioned that Matt has said regarding testing/updating was (I think) for option 3). Because there, when you have a new route which uses a specific loco, you might want to update it to include new features etc... But as this loco is already used by past routes (possible by quite a lot), you would then need to go back and test everything for those old routes again to ensure nothing in those older routes was broken by your updates (or don't test which will result in more broken stuff). Alternatively it might lead to those locos not really being updated with new features, as the effort for this becomes too big. For the other two options you do not have this, as there each route has it's own version of the loco (not looking at layering here). So if you add new stuff to the loco for the new route, the older routes still have their unchanged versions.
Why they don't use say a global install pool for routes, and a global pool for assets is beyond me, it would surely make things easier and help stop having 20 different versions of the same unit, also on a tangent RIP tree at the bottom of our field as lightning just removed that particular world asset
It completely baffles me that they have each loco within a route, instead of one core asset that gets updated with new variations and feature upgrades. The interface is a huge mess because of this, whether it's the scenario selection screen or the spawn menu (especially as there was a bug added into the mix about a year ago that causes duplicates to appear all over the place). In addition to the German locos mentioned by Emmy_MAN, we have the same thing with the UK Class 66 which exists in something like ten different "versions" in DB red livery. Stock gets updated for one route with suspension and other improvements, but the same stock in the same livery on a different route gets left behind (the Class 323 for example). Not only does it make for a terrible user experience and UI, but DTG are actually creating extra work for themselves (well, actually they're not since they never go back and upgrade the previous versions). It's complete nonsense and I don't see any justification for it, unless it's once again a PS or Xbox limitation that everyone else has to deal with. JB
I honestly don't remember if it was said whether layering works differently or what. So much good info is buried in livestreams, it gets frustrating. I'd love it if the CMs had the time and inclination to comb through past streams and post info of interest with searchable keywords.
I doubt I would fully delete the game, but as it ever eats into the HD space I'm taking a serious look at curating the amount of DLC I keep installed. The problem with a hard cull, as noted above, is the loco and train layers from stagnant routes that are still needed to layer into other stuff. I'm also taking a very hard look at what I buy in future. Was never going to buy Trent Valley Day One anyway as I just have a nagging doubt about it being rushed out and in a worse state than B2C at launch. However also thinking hard about whether Preston to Crewe will be Day One. But having reignited my interest in flight simming, the love affair with train simming in general and TSW in particular is waning for me.
I find climbing into my F4 in DCS and removing various buildings and vehicles plus the odd hapless pilot very therapeutic when TSW becomes annoying
I just can't believe how we are 6 months in since we were told its priority 1. If that was the case and surley you have had at least 1 person specifically working on said issue and in over 1000 work hours there is still no solution. Also on the December stream at about the 1 hour 3 minute mark Matt says "it won't take several months, it can't take that long."" Guess what we are 1 month away from it being several months.
