Updated AO engine movie

[Amosco]

I doubt it looks much different. It looks like the new engine uses the same models/textures (no doubt for backwards compatiblity and quick(er) conversion) and the grid is *VERY* simple structurally....basically all squares/triangles with textures over them.

No water, no grass, no fog, no clouds, no sky, no sun, no moons, none of that 'ambient' stuff that the graphics engine is actually changing.

Now if they start updating the textures/models and/or the way the engine handles them (IE: dynamically modifing textures on the fly to remove the 'repetitiveness' of them), which could be done at any point down the road, then we'll see some differences.

True, but it would be cool to see the grid get a makeover after all these years. Make it look a little less Super Mario and a little more Tron/Matrix.

[Mastura]

NOT TOO MUCH WORK to upgrade the textures? Are you slightly insane?
Have your tried counting the number of textures ingame lately? :P

No... I haven't counted them actually...

But last time I checked, creating a new texture in any free or reasonably decent 2D program is completely inconsequential to the count/number of textures.

The count would impact performance, or take time (obviously) when considering a total value number of textures.

I was actually talking about the process being simple, the willingness to assist any way possible and expressing my personal view from a motivational support context with regards to the community. I did not mention time, nor was I talking about how long it will take, but as a friend once said... "We have tiiiimeeeeeee"

Out of curiousity Healyhatman, how do you count the number of textures in the game? Is there a command in-game, or by analysing/grepping the matches in the AODB (aka the file that auno/aomainframe/bots/most 3rd party tools use?)

Cheers

O_o

[Healyhatman]

I didn;t mean it as in "have you counted them, because I have!!!"

what I MEANT was, three are lots of textures and updating them all would take a long time.... but would look fantastic obviously :P

[jorricane]

they have done a least one partial texture update (iirc it was just some outdoors ground textures) and iirc it took 4ever, but he11s yes it made a difference, and with a new engine....oh wow, just thinking about it blew my mind...

[Chrisax]

40MB video is back...

We put a new link (set as main one) to the 40MB video on
http://www.athenpaladins.org/aonewengine/index.html

We'll try to maintain a list of all "alive" links as much as possible. (Top menu on the page)

It seems the new engine is creating an incredible interest among MMORPG players, as our little page, which is not official, born this morning, and not really advertised, got between 50 and 115 visitors/hour in the last 5 hours.... AO for the win! Great job, AO team!

[Mavritanic]

tommorow be official relise this movie.
I think FC place links in news and in community forum thread.

[Dagget]

MMORPG article on the new engine:

http://www.mmorpg.com/gamelist.cfm/s...1718/gameID/10

also includes a solid link to the video (not sure if it is the 20mb or 40Mb one)

[Ebag333]

NOT TOO MUCH WORK to upgrade the textures? Are you slightly insane?
Have your tried counting the number of textures ingame lately? :P

Have you tried comparing the amount of work of updating a texture, to updating, say, water?

I'm not saying that it's not a long project. My point is that it can easily be broken up (IE: patch 19.1 releases new textures for all Omni cities, patch 20.2 releases textures for clam cities, patch 22.4 releases textures for....) and that the amount of work to update a texture is trivial compared to that of a complete new engine.

[Mekh]

MMORPG article on the new engine:

http://www.mmorpg.com/gamelist.cfm/s...1718/gameID/10

also includes a solid link to the video (not sure if it is the 20mb or 40Mb one)

Nice find...

[Mekh]

Interesting find:

Progress update:
-----------------------
My focus now have been on the constant FPS degrade.
The mouse-syrup fix added are possible to implement in a smarter way. I have completed the implementation and need to do some real thorow inhouse testing on a wide range of hardware before it can be released. This will take some time.

The old implementation ran the mouse-sirup fix code on all clients all the time. The new fix will:
- make a smaller framerate drop when it is used.
- will only be used when the client detects that its needed.

This means that if you run with 1 frame pre-rendering or if you have never suffered from the problem then you will automatically not run the fix.

The memory leak situation:
I have not triggered the problem while working on the framerate and the mouse-syrup fix. During that time I have not done real gameplay either, so it look like it can be something that happens when really playing the game.
I would appriciate input on what you do when you for example log in and have an ok framerate, then play youreself through a couple of playfields so that you end up with a lower framerate, or with a much higher usage of memory.


Information about framerate in AO:
----------------------------------------------
I guess there is no way I can tell you what framerate you shall expect at different locations in AO. This is too dependent on your hardware.
I can say that you will never get more than 100 fps as we have implemented a clamp on framerate at this level. This is done due to timing issues in the gamecode.
If you run features such as FSAA or other features possible to turn on through the driver, then you must expect that:
- framerate can be degraded due to a more buzy hardware.
- some rendering effects we use can be disabled due to incompatibility.


Offscreen surface technology:
---------------------------------------
This is a eye-candy feature only. Turning on this feature will most probably cause your framerate to go down. It is ment to be a feature for those that puts visual beauty before framerate. If you want as high framerate as possible then you should leave this feature off.
Currently in AO if you have this feature on you will see the following:
- Under-water camera and Adonis Abyss has a refractive effect.
- Air near lava has a refractive effect.
- Water has a fake (but realtime) reflection effect.
- Adonis Abyss has a highly refractive "liquid" plane above.
This feature is resolution dependent, if you run high resolution then you will see a higher framerate drop than if you run a low resolution. This is both due to fillrate and space used in the video-memory.


Comparing AO to FPS games such as Quake and UT:
-------------------------------------------------------------------
We have not added as many resources into making a super fast rendering engine such as developers that develop typical FPS based games have to do.
Anarchy-Online is a MMORPG and our personell focus have been spread widely out over the functionality field that is needed to create a massive game like this.
That would be the generic reply.

The more specific reply is as follows.
First fact is that AO runs rendering through the Direct-X 7 API. If or when we upgrade to use the Direct-X 9 API (or higher) then we will automatically get a performance jump due to the fact that we then can place all of our vertices in hardware. Today we have to keep all triangle-indices in system memory, and since we do not have access to vertex-shaders then we have to manipulate all vertices in system-memory before we render. This means that lots and lots of vertex and triangle-indice data is streamed down to the card every frame. In DX7 we cant avoid that. We will also get access to features that ease many of the challanges we face in DX7. Well, I can probably go on and on but I assume now that we agree that running DX9 instead of DX7 would be faster.

If you want to do an experiment and you have a card capable of running hardware accelerated T&L then go to an outdoor playfield where you have a complexity that gives you 20+ fps. Then chose to run the game without using hardware T&L. You'll now see that the framerate is lower than when you run with hardware accelerated T&L. Then look at AO. All characters and effects run without hardware acceleration on the T&L. If we add those too to hardware acceleration then you will automatically be able to see another jump in fps. Then you can think about all the triangle-indices that our scene have...

If we did an upgrade on the AO engine then we would be faster, but we wouldnt neccessarily be just as fast as Unreal Turnament or Quake. Alot of the speed gain that we can implement in the future is based on how we handle objects that shall render. How much do we manage to batch these together and how much better can we perform our culling?
Again I get back to the resource management in our development. This game has a very high degree of complexity and shall support more features than (not mentioning names) FPS games, this typpically breaks down to how one handles the human resources and which again puts a point on how much one focuses on the rendering engine and other stuff.

The history of AO also plays some role. AO first started off to be developed with software rendering. (Can you guess how old it is?) We converted to hardware rendering and at that time there were nothing such as a vertex shader. The API to direct-X was alot different from what we use today and the engine have gone through a few steps of upgrades.

Finally I would like to say that all of this is not an attemt to excuse the fact that Quake runs a higher framerate than AO, but to attemt to answer a question of why it is like that.
The future for AO will be as its past, we will optimize, we will add better techniques and we will push the performance, but we will still have to do it when we have the time. Slowly it will be lovely :-)

[Mavritanic]

first post updated

[Rewo]

If updating the first post, remove the AoC reference since it was clarified that this is not true. Lots of people are linking this thread from other places on the net, don't mislead people into thinking AO is going to look like Conan.

[Mastablasta]

http://www.digg.com/pc_games/Anarchy_Online_New_Engine

Digg it

[Tasslehoff]

Damn... I might even have to come back...

[Imnotamonkey]

That is really beautiful guys...Good work and thanks for showing some off!

[Crysantimum]

Osman is my hero!!

Private club talk much we have? *poke*

Greetings to Bekir

It looks fantastic to be honest (and one should be honest, don't you agree?)

[Berael]

It would be interesting to add some other things when they release the new engine, like maybe some new heads or body models. It might be a nice chance to show off the new engine and update some stuff that has been around for ever.

[Mekh]

It would be interesting to add some other things when they release the new engine, like maybe some new heads or body models. It might be a nice chance to show off the new engine and update some stuff that has been around for ever.

Good idea... /signed for support

PS.
Especially some new heads, so there would be a more diverse selction to choose from.... I might even consider another Plastic surgery then
(Changed from being big to small a long time ago)

[Tonberry]

WTB ingame action movie so we can se how the engine works in action on chars etc

[crymthann]

I would love to see a more diverse Character creation system at some point.It would be great if after you chose your breed instead of only choices for head, build and height;we got to choose hair color and style,eye color,scar patterns(for nonos).maybe add in some of the facial tattoos.