Jump to content
Welcome to the virtual battlefield, Guest!

World War II Online is a Massively Multiplayer Online First Person Shooter based in Western Europe between 1939 and 1943. Through land, sea, and air combat using a ultra-realistic game engine, combined with a strategic layer, in the largest game world ever created - We offer the best WWII simulation experience around.

dropbear

Presagis creator to unreal

Recommended Posts

dropbear

While looking around the latest unreal engine I fell upon this independent conversion utility.

https://www.presagis.com/en/product/openflight2unreal/

It uses datasmith as an intermediary so should in theory be able to import ALL the games modelling info into unreal, and therefore be able to tap into the thousands of hobbyist modeler's out there, especially in the Covid-19 times.

 

Thoughts?

Share this post


Link to post
Share on other sites
jwilly

In the past, Presagis's perspective on Unreal conversions was that Unreal was a very limited system compared to Creator, so conversions from the latter to the former would be for temporary demos and the like and significant data loss and capability loss were OK.

Epic's publishing of their DataSmith toolset may help with this. I do note though that everything here is v. 1.0 at the latest, published three weeks ago, and the release notes are both sparse and liberal in listing fundamental things that aren't, or aren't yet, supported.

It still looks to me as if Presagis's corporate perspective is that no one in their right mind would re-publish something developed in Creator in a gaming engine, so supporting the use of Unreal for easier demos has no downside.

As to importation of hobbyist models...sure, there's lots of stuff out there. Scotsman for one has been telling us for years that he has access to lots of American, German, British, French and Italian stuff. The question in regard to such models is whether their availability will make a difference to the game, or whether instead the need is for stronger gameplay mechanics and better terrain.

  • Like 1

Share this post


Link to post
Share on other sites
madrebel
3 hours ago, jwilly said:

or whether instead the need is for stronger gameplay mechanics and better terrain.

would this work to quickly make a terrain for the Utech engine via say voxelfarm and then import that to openflight/creator for our uses? say ... for better terrain? or is it creator > UTech only? last i read voxelfarm had some killer procedural generation tools for very large scene scapes. there is a plugin to export work done in voxl farm to UT format, if you can then take one more step to creator ... idk.

updated terrain is probably the thing that can most heavily influence the entire game for the better IMO.

Share this post


Link to post
Share on other sites
jwilly
1 hour ago, madrebel said:

would this work to quickly make a terrain for the Utech engine via say voxelfarm and then import that to openflight/creator for our uses? say ... for better terrain? or is it creator > UTech only? last i read voxelfarm had some killer procedural generation tools for very large scene scapes. there is a plugin to export work done in voxl farm to UT format, if you can then take one more step to creator ... idk.

updated terrain is probably the thing that can most heavily influence the entire game for the better IMO.

There have been several technologies over the years for procedurally generating wide-scale terrain. 

The challenge, I think, is that most ground spotting and fighting takes place at distances no greater than a few hundred meters, or maybe a little more. So, most of the most important aspects of the terrain need to have their highest performance at that scale at the high end, and...for infantry gameplay...at much tighter scales.

I'm not aware of procedural approaches compatible with Creator that work at those small scales.

The WWIIOL implementation of Creator involved building a set of 800m x 800m tiles...I think about 64 different designs...containing (in various combinations) roads, bridges, airfields, railroads, forests, rivers, shorelines and ocean-water. Creator provides an engine that starts by laying out a grid-array of "basic" tiles, with none of the features listed above. Then the shoreline of the ocean-water is defined...that involves particular tile types. Then the terrain developers can remove "basic" tiles, one at a time, and replace each one with one of the custom tiles at any of four compass orientations, arranged so that roads, railroads, forests and so forth all connect and line up.

There's a whole set of technology and mechanics related to tile corner elevations, tile "stretching" and seam matching, easiest to set aside for a separate discussion.

Much of the terrain developer's work is just beginning at that point. Buildings and most other terrain features and details are either placed individually, or are placed algorithmically and then adjusted individually to remove conflicts and achieve the local design goals.

It's also possible, but a lot more work, to do a localized fully-custom terrain solution. 

The limitation of the existing terrain technology IMO mostly comes from the 800 meter tile size, which handles rolling hills OK but is way too big to allow good modeling of rough terrain.

It's also too big from the perspective of texturing.

It might be possible to integrate a second tile-set, 50m x 50m. That would allow for modeling terrain 256x rougher, but it would require 256x as much developer work. Doc and others have told us many times over the years that the existing 800m tiles are right at the edge of being non-viable in terms of getting the work done at all. 

In any case, I don't think there's a viable terrain solution in the procedural-engine direction. The output of such an engine could be beautiful for an air-only game, but I don't see how it could work for an infantry game.

 

 

Edited by jwilly

Share this post


Link to post
Share on other sites
madrebel

right, what i'm saying is create 'Europe' using voxel farm, then export it to unreal engine as high definition 800x800m tiles. or larger if we can take it. from there, import them to creator. you'd still likely need to map speedtree stuff onto those tiles but i'm not sure how you get around that anyways as that's how it is done now yes?

as long as the tiles themselves are high definition. not sure how you'd map buildings and what not. hell, with high def terrain lets start over from the original map. i'd rather have a smaller much more detailed map than a big map with terrible terrain.

Share this post


Link to post
Share on other sites
blggles

You want to fill out a new map quick, an intuitive mapmaking app the community could easily use would likely be necessary. The game does have the PPO tech, which is easy enough. If PPOs could be compiled into the terrain data, and you made most objects PPOs, well there ya go. 

  • Like 1

Share this post


Link to post
Share on other sites
jwilly
35 minutes ago, madrebel said:

i'd rather have a smaller much more detailed map than a big map with terrible terrain.

That was the exact marketing conclusion that CRS came to, that led them to the Rapid Assault project. 

Quote

right, what i'm saying is create 'Europe' using voxel farm, then export it to unreal engine as high definition 800x800m tiles. 

I think the limit of the Creator terrain engine technology is a set of 64 tiles. I don't actually know if two tilesets at different sizes can co-exist. Rapid Assault, since it had no requirement to be integrated with or based on the existing terrain, was using a single all-new tile set at a size unknown to me.

It's surprising, once you try to create a tileset for this purpose, how not-far 64 tile designs goes when you have to build a connected world using those tiles and no others.

Having tried to re-engineer the existing technology many years ago for a better water-terrain solution, I don't see how it would work to use something like Voxel Farm if where you need to end up is tiled terrain suitable for the ground game.

The Speedtree stuff and related details are the last part of the developer work, to customize each tile to fit into its local world.

Share this post


Link to post
Share on other sites
jwilly
7 minutes ago, blggles said:

You want to fill out a new map quick, an intuitive mapmaking app the community could easily use would likely be necessary. The game does have the PPO tech, which is easy enough. If PPOs could be compiled into the terrain data, and you made most objects PPOs, well there ya go. 

The "compile" part is a seat-license for Creator. Expensive, and said to be non-intuitive.

Everything you move from the tile designs to PPO status has just been shifted to that last part of the developer work, where the Speedtree stuff and towns are created. Lots of slow heavy lifting, with not a lot of tools to help you do the work. OTOH you get the flexibility to have pretty much anything you want in your toolbox...other than tools, of course.

Share this post


Link to post
Share on other sites
blggles

I was just thinking of the finishing objects, the filler. Would be great if the community could help there.

Share this post


Link to post
Share on other sites
XOOM
On 5/25/2020 at 2:55 PM, dropbear said:

While looking around the latest unreal engine I fell upon this independent conversion utility.

https://www.presagis.com/en/product/openflight2unreal/

It uses datasmith as an intermediary so should in theory be able to import ALL the games modelling info into unreal, and therefore be able to tap into the thousands of hobbyist modeler's out there, especially in the Covid-19 times.

 

Thoughts?

I saw this actually last week after doing some investigation. While it does successfully transfer the model over, it breaks it up into hundreds of pieces, and loses all of the essential comments on DOF's that do not make it a 1:1 translation. CRS is actively investigating how to make that transfer process 1:1 but we'll probably have to build our own in-house converter. More to follow.

  • Like 3

Share this post


Link to post
Share on other sites
blipoop
1 hour ago, XOOM said:

I saw this actually last week after doing some investigation. While it does successfully transfer the model over, it breaks it up into hundreds of pieces, and loses all of the essential comments on DOF's that do not make it a 1:1 translation. CRS is actively investigating how to make that transfer process 1:1 but we'll probably have to build our own in-house converter. More to follow.

So this isn't a cold and hard 'No' to Unreal but a 'Pending research' it's a maybe?

Share this post


Link to post
Share on other sites
XOOM
2 minutes ago, blipoop said:

So this isn't a cold and hard 'No' to Unreal but a 'Pending research' it's a maybe?

It is my highest priority, to advance WWII Online's graphics and take us into the modern era.

  • Like 4

Share this post


Link to post
Share on other sites
CHIMM
On 5/25/2020 at 4:06 PM, jwilly said:

Scotsman for one has been telling us for years that he has access to lots of American, German, British, French and Italian stuff. The question in regard to such models is whether their availability will make a difference to the game, or whether instead the need is for stronger gameplay mechanics and better terrain.

To the last part of your quote, one of our bottlenecks is model animators. 

Share this post


Link to post
Share on other sites
SNIPER62

lol funny how this comes up as we noticed the same tool a few weeks ago, as xoom said.

  • Like 1

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...