Cyrodiil Scaled x 2 : IV - Oblivion

» Tue Feb 15, 2011 10:59 pm

Hi folks, it'll be a few days before there's more progress as I'm heavily rewriting TES4Scale to cope with multiple re-referencing of duplicated records. The old code was getting too hacky and messy so I've mostly trashed it and started a better rewrite. This will make it easier to preserve other CELL record details such as exterior cell water heights, ownership and regions as well as duplicating other worldspace specific records, scaling them and fixing FormID references. But it'll take a bit of time. Just scaling the region (REGN) record points themselves wasn't enough for me to test region generation as I need them assigned to the correct CELL records too.
Testing region generation will be the next task, but I've no reason to think it won't work.

Given that mods plus an "old" character with a fair few hours on it can make load times significantly longer I am thinking that 4x will simply be unplayable and 3x will be pushing it. Just my thoughts, which do you think is the most plausible solution Lightwave?

On my system any poor FPS is usually down to huge AI lag as the test files I put up use the original NPCs with AI packages specific to the Tamriel worldspace. Unfortunately you can't disable AI until after you've loaded your first gamesave to test whether this is the only cause on your system. But once disabled in the console this should no longer affect test load times and in-game playing performance - so it's worth trying that. Apart from the AI problem and the initial load I haven't personally noticed any real FPS differences between 1x, 2x and 3x. Btw, how long does it take your PC to get to the initial main menu with a 3x landscape loaded?
Psychologically 3x shouldn't feel bad, 3rd time lucky, everything happens in 3s.

Also the Vvardenfell:Cyrodiil map surface area difference is probably closer to 3x than any other (though this is a guestimate - I haven't tried counting the actual surface area). Splitting the world into 2 or 3 smaller ESPs is possible, though does complicate the management a bit and trying to create an AI package to walk from city to another might force people to load all the ESPs anyway to avoid errors. On the plus side, all the LOD (land and objects) would be visible even if just one of the files is loaded, so it's still possible to test smaller mod chunks and get the feeling you're in the bigger world.

(many suggestions)

Thanks but they're definitely amongst the hardest ways of going about the task and with lots of complications - it would almost be simpler to just start with a completely new heightmap with more intricate features, but that would be an unrealistic amount of work. This method is really quick and is working pretty well.
Vality7 is correct in pointing out that any non-linear horizontal:vertical combination will not work well for any the placed content as it's is either going to be spread out or meshes will end up overlapping one another. The landscape would be OK though. The content is possible to fix in the render window - if we are going to region generate then I can just import, say, architecture type records which will make it easier to locate settlements and fix the layout in the render window. But that will mean more work in the CS.

The critically flawed assumption is this: Morrowind is the first Elder Scrolls game (which it isn't).

I'm not sure how you arrived at that assumption. TES3/TES4 are at core the same game engines, at an absolute scale to themselves (no imaginary vast wilderness) and with the same length measurement systems. With players able to travel back and forth between Cyrodiil and Vvardenfell in both game engines it's then distinctly obvious to the player that the relative size of both lands is greatly disproportionate to what might be expected in written lore or with a map and it's worse when viewing the distantland from the other province in either game. TES1/TES2 are from different generations (they're DOS) and not directly comparable.

Most importantly and even in isolation on it's own game engine, Oblivion's Cyrodiil is too small, in the way it looks, plays and feels, with tiny villages called towns that fail to capture the feeling of a sizable town or a diverse world - more ambient sound would probably help for a start, the streets seem dead without it, but the IC is a mere fort on an island. As the capital province in Tamriel, Cyrodiil is supposed to be large, but it doesn't feel that way, so I'm hoping to move towards capturing that in the most productive, sensible and achievable way.

Also I've already pointed out that movement speed with the same stats is virtually the same in TES3 and TES4, it's the traversability of terrain that's the biggest factor, but there really are a multitude of importany factors and realistically we're not going to make it look any bigger without actually making it bigger.

I'm just wondering, how could one do this for a completely new landmass? Is there a special program, or is it all done in the CS?

These scaled landmasses are exported and imported from the original heightmap using a program called TESAnnwyn. It's not possible to do the same in the CS without masses of work, re-texturing and splitting it in to smaller mod files to get around the CS active-mode memory limitation. If you want to create your own landmass though there are many ways to do it, and there's always something going on in the Construction Set forums.

Lightwave

» Wed Feb 16, 2011 2:19 am

Creating a map on the same scale as Cyrodiil, then stretching it in the same way wouldn't work?

» Wed Feb 16, 2011 12:23 am

I haven't really been keeping up with this, but is the plan to create a program that will work alongside the user's existing mods and scale them up with the rest of the world?

I'd love to have the world be bigger, but unfortunately when running 200 some mods it's NEVER going to be compatible with the mods I run and I'll have to make do without

» Tue Feb 15, 2011 10:49 pm

I haven't really been keeping up with this, but is the plan to create a program that will work alongside the user's existing mods and scale them up with the rest of the world?

I'd love to have the world be bigger, but unfortunately when running 200 some mods it's NEVER going to be compatible with the mods I run and I'll have to make do without

No, I am pretty sure that we are worrying about getting Cyrodiil complete first, then we move onto mods

Plus, the majority of mods should still run correctly. Most of OOO would still work flawlessly. Harvest would work flawlessly. The only kind of mods that I think would not work are Dungeon mods and some armor/weapon mods that place their weapons in exterior cells.

» Tue Feb 15, 2011 6:59 pm

No, I am pretty sure that we are worrying about getting Cyrodiil complete first, then we move onto mods

Plus, the majority of mods should still run correctly. Most of OOO would still work flawlessly. Harvest would work flawlessly. The only kind of mods that I think would not work are Dungeon mods and some armor/weapon mods that place their weapons in exterior cells.

Plus, to be completely believable, city meshes would have to remain the same size, and the cities would have to be expanded, e.g. more buildings, not bigger buildings, because the actors remain the same size.

» Tue Feb 15, 2011 6:41 pm

A fundamental flaw with oblivion is down to the original project that Bethesda began when they chose Cyrodiil for TES4: they tried to create the enitre provice for the game, rather than a smaller part of it at a greater scale (like the worldspace used for Cyrodiil Scaled x3), and the result leaves the player with the sense that they are part of a much smaller world... especially if they use a mod like AEVWD (Almost Everything Visible When Distant). It might have been in some ways less ambitious, but had they perhaps tried to create a small number of fully realized cities (2 or 3, including the Imperial City) and the area around them, they could potentially have produced a much more immersive experience, with a much higher level of detail and interactivity (so a relatively small world that always has something interesting around the next bend rather than miles and miles of terrain with nothing more than a couple of monsters to fill the void).

If anyone is going to take the concept of an upscaled Cyrodiil mod and run with it, it might be a good idea to produce even a single city as a testbed to see just how far this concept can be taken.

Of course, that's just my two drakes.

» Tue Feb 15, 2011 7:16 pm

Lightwave, how goes the rewriting of TES4Scale?

» Wed Feb 16, 2011 2:26 am

Don't tell me THIS is dead...

» Tue Feb 15, 2011 9:00 pm

Uh - it's a pretty busy time of year for most people. Give the man some time.

» Wed Feb 16, 2011 1:19 am

In playable area, Cyrodiil and Vvardenfell are almost the same size (Cyrodiil being only 4% larger). This is obviously a major disproportion, and I think you should find which of the scales (X2, X3, or X4) would make Cyrodiil more proportionate to Vvardenfell's size from TES3. As far as the overall lore is concerned, Cyrodiil is 30% larger than Morrowind, and Vvardenfell is about one-third the size of Morrowind. This means that Cyrodiil (when properly scaled) should be four times larger than Vvardenfell. For this reason, you should stick with the X2 scale (because at X2 scale, Cyrodiil is four times larger than it currently is).

There of course are other benefits of sticking with the X2 scale:

1.) It would be much easier to work with
2.) It would not cause a mountain height problem (or a horizontal/vertical disproportion)
3.) It would render, load and run MUCH faster than X3 or X4
4.) And, it would be much more practical for modders

Another thing I think you should consider is placing cities. I know the desire is to ultimately expand the cities, but for now it would be best (in my humble opinion) to simply copy the normal size cities and place them in the X2 world map. I understand that some of the cities would not fit properly (mainly Bravil, Skingraad, and Anvil), but you could copy and paste them rather quickly and it would give us modders a really good base to work with. Anyway, I hope you are still working on this, and I thank you for attempting it.

...And again, X2 scale is not just easier, but IT IS THE CORRECTLY SCALED SIZE compared with TES3.

» Tue Feb 15, 2011 8:10 pm

Sir Nappyhead! Hear! Hear! I agree completely!

Even just longer distances between vanilla cities in a x2 world map would feel immensely more immersive and add leaps and bounds to the feeling of being in Tamriel. And that project stands very good chances of succes and immediate rewards as opposed to the more oppulent undertaking and humongous task of creating a x3 or x4 worldmap with expanded cities. A x2 world map with vanilla cities spaced out proportionally could act as a proof of concept and be the first wave upon which more daring endeavours could be initiated.

Sir Lightwave! Thank you very much for your time and hard work on this marvellous concept. I sincerely hope that all is well and that this mod will be fruitfull.

Thank you and all the best
Frans

» Wed Feb 16, 2011 3:43 am

To me this is the single most promising mod out there. I would hope more modders would catch on to this and run with it.

Even placing the vanilla cities of Bravil and Skingraad wouldn't be to difficult (you would only have to add a few more walls and maybe slightly change the outer layout of Bravil to fit on its islands. Consider that the vanilla Chorrol, Kvatch, Chedinhal, Bruma, Lewayiin, and Imperial City could be copied and pasted to the new X2 world without alteration. For example, with Kvatch, you could simply center the city on its hill, and similarly do the same with the Imperial City.

I think this relatively simple mod has the potential of being looked back on as the Oblivion modding community's greatest achievement.

» Wed Feb 16, 2011 4:50 am

:goodjob:

x2 is the scale which seems both most lore-reasonable and most attainable.

In an appropriate world-conversion, what would be the first steps for modders to tackle?

Cheers,

Michael

Edit:

I think this relatively simple mod has the potential of being looked back on as the Oblivion modding community's greatest achievement.

Absolutely no doubt on this point.

Size matters. A lot.

» Wed Feb 16, 2011 5:14 am

Everyone you do realize in order to get the city imported into this would be a huge undertaking bytself...Plus all the pathgrid that would have to be rebuilt..I can keep on listing a few things ...

But this project really needs a team of dedicated individuals who know how to do this stuff in order so that things could be accomplished ...Lightwave is only one person...and can only do so much...

» Wed Feb 16, 2011 2:23 am

I'm pretty sure you'd have to create four times as many cells as well, to prevent massive loading issues. That would require remapping every spawn point as well as all the path grids.

» Wed Feb 16, 2011 1:58 am

lore-reasonable

Okay, drop lore. Reasonable, yeah.

Morrowind was greatly reduced in scale compared to what it "should" have been following lore. The only game of the series with a realistic scale was Daggerfall. Of course it meant that it was mostly filled with procedurally-generated garbage. A whole team working for ten years couldn't take the heightmap from Daggerfall and detail it to the level of Oblivion or Morrowind, with hand-placed and QA-reviewed stuff for each cell. Twenty years wouldn't be enough. Thirty years wouldn't be enough.

So, what we get is, as King Arthur's lackey Patsy would say, only a model. Cyrodiil is about the same size as Vvardenfell because both models are about the same size. What they represent isn't, though. So the Cyrodiil model appears considerably more condensed than the Vvardenfell model.

Anyway, Morrowind's scale isn't lore correct. Oblivion's scale isn't lore correct. They're merely practical.

And Arena's scale wasn't lore-correct either: Arena's wilderness was infinite, randomly generated garbage. You couldn't travel from one city to another without using fast travel. Redguard was a special case, since it was an action-adventure game rather than a free-form RPG. And Battlespire was just a giant dungeon, without any part in Tamriel. I won't bother talking about the "Elder Scrolls Travel" games.

» Wed Feb 16, 2011 1:23 am

As well as the grass problem, the player's height visual always stays the same. If you do a 'player.setscale 2' or 'player.setscale 0.5' then you only see the difference when you go to 3rd person mode (then NPCs around you move their heads to look at your new height, then move them back again when you revert to 1st person mode). Also the viewing distance is the same ...

So if I change the scale of my character then I won't have to worry about mini horses and mega chairs?

» Wed Feb 16, 2011 6:23 am

Anything still happening here? Sounded really promising.

» Tue Feb 15, 2011 9:33 pm

Since it looks a little busy this time of year (without lightwave gracing us with his presence) I'm not going to ask if this is dead.

What I will ask, is how goes the progress? Do we think this will work well? I downloaded and ran around the 2x version for a while and like the feel of it a bit. I'm about to replicate a couple towns and travel between them and see how it feels that way, and also how it feels walking inside them. Should this 'mod' become more than a testbed of ideas, I would like to extend my services of rebuilding the cities.

Once I have something more, like the two cities for a test, maybe I'll upload it and see if anyone likes my ideas. Currently, I'm planning on copying the vanilla IC and Chorrol. Should be a good way to see how it feels. If it's good, I'll redesign Chorrol, and upload that. See if anyone likes.

Since I haven't said it yet; :goodjob: !

» Tue Feb 15, 2011 9:27 pm

Thats cool, does it add more trees ect. to fill in the extra land, if not I offer my own landscaping skills to help improve the feeling of it? Oblivion just wasn't big enough and the regions not deep enough, it always seems like just when I feel like im getting deep into a forest, theres a town or a trail or all too often ruin with nothing new in it. Also I feel like the map and quick travel break the immersion, you can get around to easy. I feel you should have to memorize the landscape to travel, not follow the little arrow on the map. If you have a map, it should appear drawn out, and without the 'you are here' arrow, mapping skills should be required.

» Wed Feb 16, 2011 3:28 am

So in high end PC Tamrielx2 not run correctly - Tes Cs not support so big land mass((
base esp 213mb not loaded in CS as active plugin -anybody did??

» Wed Feb 16, 2011 4:35 am

So i guess no one took up the challenge?

What if cities, settlements, ruins and whatnot could stay in original scale, with their immediate surroundings, so only a circular void around them was left for landscaping? Not that i'd know how to do that... just thinking.

» Tue Feb 15, 2011 10:17 pm

oh man i was just looking for this, thanks Hippie for reading my mind

since this is (if i understand correctly) a separate worldspace everything custom made would need to be moved by hand, which is really nothing more than editing the new landscape to fit and a tedious copy-paste job from the old worldspace. regions could be re-generated with a minimum amount of work.

» Wed Feb 16, 2011 6:55 am

oh man i was just looking for this, thanks Hippie for reading my mind

since this is (if i understand correctly) a separate worldspace everything custom made would need to be moved by hand, which is really nothing more than editing the new landscape to fit and a tedious copy-paste job from the old worldspace. regions could be re-generated with a minimum amount of work.

Well the problem is quests isn't it. Objects and items are more or less easy to move (it'll take a while though), but the quests are much more difficult, having to go through every one of them to make sure they refer to the right worldspace and all that.

Otherwise I really like the idea of this.

» Wed Feb 16, 2011 11:14 am

yea, i hadent considered quests, dialog, pathgrids, and pretty much everything else NPC related. then again, it is my personal opinion that if this was to be done it should not be a direct copy of oblivion but instead a complication of all the pwnage that is mods in a single worldspace.