EAW's greatest short coming

[VonBeerhofen 63]

ALL eaw 3D models are flawed, except, for a few simple shapes and those which came from other Microprose games and remained unaltered.

Now this statement is NOT an attack on anyone's endeavors in 3D modelling, it's just a simple fact. All modified planes are stuck together with spit and string, as Moggy so aptly described in another forum. In 18 years of flying EAW I haven't encountered one single 3D model which wasn't stuck together like that.

As a result of lack of knowledge on the subject of calculating and understanding the rendering sequence, nearly all models show bits and pieces randomly switching on/off at certain viewing angles or distances or 3D parts being visible when none should be visible. Perhaps I'm spoiled but this fact has always annoyed me, simmilar to the viewdistance switch which happens between lores and hires model when aproaching an object. I find it distracting and it takes away much of the immersion EAW provides.

For this reason I've visited many websites dealing with the subject and read whatever I could find on the topic. It's taken me 10+ years to understand, plus I spend a lot of time dabbling with other commercial 3D packages used for other flightsims but most of these are not comparable to how EAW deals with a 3D world. I had to revert to a more hands on approach using the tools made available by some more knowledgable people then me.

I learned that there are no shortcuts and very stringent rules for creating 3D models and that not a single 3D model will allow the same approach in order to get it to work. However, more important was that after very slow progress I verified that Gurney's R/S calculator is working 100%, as long as you stick to the rules of 3DZ modelling. This again means that there is no need for flawed 3D models, they can all be fixed provided that someone will spend enough time on figuring out why some polygons cause the calculator to screw up.

It is a painstaking process which requires a lot of ingame testing, but there's always a solution even though the solution to the problem is different for each model. Often there are multiple solutions, which is where decisions have to be taken as to which one is the best. Now the sad part is that a solution may remain illusive for months or even years and it's useless to stare yourself blind on it, it's like finding the right pieces of a puzzle or like Rubic's cube, where a single polygon change can cause a cascade of errors in the model which will always cause Gurney's R/S calculator to get stuck.

I want to reitterate, that the calculator is fine and when it locks up there's something not right in the model. When it's not right it can cause the game to crash, which is obviously another indicator that it's not right. I can't teach people how to solve such issues, it would take too much of my time, only trial and error will bring this understanding. I also hope you will understand that fixing every flawed 3D model from the past, that's nearly all of them, is a herculean task which I'm not going to take on.

It's more fun to create your own models, as you're in charge of what you're doing when you build it, then it is trying to spot where others have gone wrong in their endeavours and you can only use the current tools available for doing so. On a personal level there are some simple good working models which are flawless because of their simple design, but everything  flawed is total failure in my eyes and proof of not understanding of how it works. I know it's a harsh judgement but in doing things right there simply is no compromise.

Even I do have models I have been unable to fix up till now, which started 10 years ago. They just didn't make it because I can't see what's wrong, and untill I find the solution it stays under the heade of total failure and will not see release untill it's been fixed. I know there is always a solution so I come back to it to see if I can solve it that time. Any other approach is useless, wether you can see the model in game or not.

I hope some people will rise to the challenge and have the determination required to truely become a 3DZ modeller, it's about not giving up or trying to take shortcuts, no matter how much time it takes. Start small and work your way up, study the Microprose models as some of the answers are in there.  Its the only advice I can give.

VonBeerhofen

[rotton50 19]

Sorry, but not true.

SOME models have R/S element issues, hardly noticeable to downright awful. However, it is incorrect to say that nearly all models have errors.

I've got plenty of perfect models flying throughout the EAW world. I have some that have minor R/S issues but they look fine too. Haven't had any complaints in 15 years.

I mean, how many times while flying do you jump out and check to see if the middle axle on your bi-plane landing gear is rendering properly? Do that too often in you're going to get shot down.

I mention the bi-planes because all of the originals were dreadful but I patched then up pretty well except for looking up directly under the model.

I guess it's simply a case of the old German saying - "The enemy of good is perfect".

 

 

[VonBeerhofen 63]

In another forum I checked a fair ammount of your planes randomly choosen Rotton, the results of this are in that forum, but it won't hurt to make the same statement as I made there, each one tested was flawed, some were seriously flawed, others less so but still flawed.

I understand your German saying, it's principally saying that you find em good enough, but that doesn't refute my statement. As I wrote for me personally only perfect is good enough. Your welcome to disagree with me and you're entitled to your opinion but you won't change mine.

Improvement to EAW comes with wanting to achieve perfection, if that incentive is missing then things aren't any different then before 2003. Surely you're treating your skins simmilarly, as everyone who has tried to build these models, making them as good as you possibly can. People who've build campaigns also strived for perfection, placing the right objects, figuring out frontlines, drawing maps, reticules etc.

I'm not telling you to achieve that perfection, that's up to you. I however am striving for such perfection because I want to do things right, so I learned how to do it. I bet even Moggy is striving for perfection for his carrier, wether he achieves it or not depends on the determination he has to reach that goal. When he achieves it I'm sure it will be very rewarding to him as there's no better feeling then solving a model's R/S and get a flawless model

VonBeerhofen

[rotton50 19]

We obviously have a different definition of serious flaws but that's neither here nor there. You have your goals and I have mine.

The main point is that the most important thing in the EAW world is not the perfection of 3dz plane models. The most important thing is that players keep playing this 20 year old marvel of the flight sim world.

My inventory of perfect and not so perfect models satisfies that need and keeps players interested, which keeps them coming back.

FWIW, the 1.6 group has now produced an inventory of 300-400 war planes from pre WWII bi-planes to early jets, in a dozen scenarios from the Mediterranean to the South Pacific.

There's also a huge selection of X-planes, some perfect some not perfect, that allow players to try their hand on experimental models.

in addition, players can use preset plane packages or build their own by picking from those 400 or so planes. So if a player IS a perfectionist, he can edit out any plane that doesn't rise to that level. It's a win-win.

And let's not forget, this is a combat flight sim, not a scenery sim.

So we will have to agree to disagree.

 

Edited September 19, 2018 by rotton50

[VonBeerhofen 63]

Agreed, it IS a flightsim, and the better the AI are the more realistic the fights, which is why EAWPRO keeps the polygon and nodecount as low as possible in it's models and has a drastically shortened gameloop inspite of it's many enhancemets. The rule of thumb is that the faster the gameloop and framerate the smarter the AI, and this is only possible because the models are not flawed.

A selection of 300+ planes doesn't really make much difference though when it comes to exciting fights, nor do 300+ flightmodels. People can use any plane they like in EAWPRO or any flightmodel created for v1.2 but it won't lead to more ballanced fights or more interesting ones then what the stock game provides, especially not when these planes aren't matched against each other.

It's not how many objects I can provide or how many planes you can, or how many of each have been created in the past. It's about allowing people to have fun with EAW and the freedom to select what they like. In that EAWPRO can provide a new experience which is different from anything else previously created, in the exact same way as campaigns used to entertain pilots.

EAWPRO isn't like 1.60 and doesn't pretend to be, it's much closer to the original game and is meant to be used simmilarly, It's user interface is as people remember from the 1.2 version but with many extras, but when they get bored with it, or don't likeit,  just as may happen with any game then you're free to erase it. EAW has held my interest for over 18 years and I hope EAWPRO will capture people's interest for another 18 years, I hope the same goes for v1.60.

It's getting a bit off topic now, I'm just saying that the models can work flawlessly if sufficient time and energy is spend on them and it could be the next step in keeping EAW alive, well that's what I strive for.

VonBeerhofen

 

[rotton50 19]

With the state of chip processor speed being what it is today, vis-a-vis a 20 year old flight sim, I dare say there's no worry about having enough processing capacity to handle, well, anything EAW can bring to the table.

Heck, even 10 year old PC's could handle EAW requirements.

What you could do would be to try a one month experiment using only hi-res models and see if you or your players notice any difference in the AI behavior. You might be able to upgrade to better quality models if you find out there's no hit to the AI performance.

Wouldn't that be something? Upgraded models and all the bells and whistles of EAWPro.

Just a suggestion to improve the overall satisfaction of the customer base.

BTW, I forgot to add that Gurney's R/S/ checker is NOT 100% accurate.

I too thought it was until I started combining left and right side tail sections as I've upgraded old models using my "empty F" process. I do this to free up one of the add-on sections to be used for some other part of the plane.

Imagine my surprise when I assembled a new tail section, got a green light from the R/S checker and then had the horizontal stabilizers show through the vertical rudder.

And it's happened more than once.

I'm not being critical, just passing on the information.

 

[VonBeerhofen 63]

Complete planepacks including flightmodels may be interesting to some as an addon for EAWPRO and you're welcome to create them and publicise your work. Tests in the Launchpad has shown that with the current groundobjects and stock planes framerates are decent, but with 256MB memory on all our machines, except Russ Watson's, the entire addon should not exceed 50MB and that's where we are now.

The group feels they do not want to enter into a race for the latest technological improvements in computers, our rigs do the job and that's all we need. Besides that none of them wants to learn a new operating system, especially not those who're in the final stage of their lives. Besides this there's the hassle for the players who've already expressed they're totally happy with what they have and adding plane packs, no matter how simple you make it, is bound to cause severe problems with our oldest player, partially because he's 76 I believe, but also because he doesn't understand much of computers and addon, and his English is terrible (Sorry FranK).

The only thing I'm currently considering is to create one final FXEXE, which, like previously, can be added to any 1.2 addon and upgrade it to EAWPRO standards without having to make further changes. I just can't say if I can make it work, time will tell.

I realise that people may want to change the stock aircraft and they're welcome to give it a try, it's not any different then it was in v1.2 and personally I've done it many times. Still it does require some extra knowledge and that knowledge is disappearing fast. Besides that there aren't many helpers like in the past and although I'm gladly willing to help it shouldn't turn into a dayjob of creating personal addons. If people want to mess with things they should learn the ropes, just like I had to.

Whatever is thrown into EAWPRO may make it graphically better, just like what the Final Cut already is, but will not improve AI behaviour when the combined files exceed the combined size of embedded files in the CDF's. I can't prevent tampering with EAWPRO or the Final Cut, but people must understand that neither fully works as v1.2, if you want to do it right you'll need to understand the differences.

As such you may find that some of your planes will need extra work if you want to incorporate them into EAWPRO. Such changes relate to line objects and gear(well) codes, amongst other thing. Also, the code for flaps isn't the same as Col. Gibbon's code. Quite a few animation codes were added to EAWPRO which may conflict with the one's you use.

VonBeerhofen

[rotton50 19]

The only plane I know of that took full advantage of the flap animations is the P-51. It was a PITA to set up but it would only take about 15 minutes to undo. Generally speaking it's a lot easier to delete elements than add them.

Talk it over with your group. If they want to try some modified versions of any of my models that won't tax their older machines I'd be happy to oblige.

 

[VonBeerhofen 63]

I've talked it over with the Launchpad members and the only person who I think has an interest would be Russ Watson, since his machine is way more advanced then ours. He would gladly like to see a hires planepack to test out and there may be others who have an interest too. His current framerate with EAWPRO and the Windower program is between 55 and 60 in all situations, with heavy activity and low to the ground with all settings high.

VonBeerhofen

Edited September 22, 2018 by VonBeerhofen

[rwatson 26]

VB Getting 55-60 from in the clouds to 50 feet off the deck but to clarify I'm not using the windower gave up on it..with a 24 inch screen can't see giving up the real estate..Just went to mark's site and downloaded some screens and found the ones who looked best and changes to my ini i do manually because the stick setup screen is tought to read