PDA

View Full Version : G4 Ground behaviour...


Troll
11-13-2007, 01:37 PM
I just took the plunge and bought G4 after enjoying G3.5 for some time now. I was quite happy with G3.5, but sea flying and the new controller was too tempting!

Overall I'm happy with G4! The new controller is great! Sea flying is a hoot!

However, if I could wish for one thing, it would be the way aircraft behave on ground... They move as if on a glass surface. Ok, there is a difference in resistance between grass and tarmac, but the aircraft should be bounce and shake a bit when rolling on grass. Even on tarmac, a little.

Think about it. When you taxi, takeoff or land your RC aircraft in real life, the wings will wag, the tail will bounce and hop as the ground is ever so slightly uneven.

The way float aircraft behave on water is much more convincing than how the wheeled aircraft behave on ground.

Am I way off here, or does anybody else agree?

Cheers!

Jorgen "Troll" Toll

pplace
11-13-2007, 01:54 PM
95% of us agree with you and hope that KE will address this issue. You could probably search for "ground handling" and find dozens and dozens of threads dealing with this situation.

I also was very pleased with the "water handling" and was shocked they have it as good as they do. Seems to me "ground handling" would be easier to achieve??

I know I've seen it on another sim....where you roll off the edge of the runway and the airplane rocks side to side....much more realistic.

josh0987
11-13-2007, 02:11 PM
As far as I know all that has happened when I taxi..

josh0987
11-13-2007, 02:12 PM
95% of us agree with you and hope that KE will address this issue. You could probably search for "ground handling" and find dozens and dozens of threads dealing with this situation.

I also was very pleased with the "water handling" and was shocked they have it as good as they do. Seems to me "ground handling" would be easier to achieve??

I know I've seen it on another sim....where you roll off the edge of the runway and the airplane rocks side to side....much more realistic.
Where did you get the 95% from? :confused:

phrank
11-13-2007, 02:18 PM
I disagree.... We need more realistic ground handling.

Maybe you could post a video of the ground handling you are seeing... 'cause I certainly don't see it.

Just wondering,
Do you fly an R/C plane in real life? or is all your exposure sim related?
I don't get it. :confused:

redherring
11-13-2007, 03:18 PM
I disagree.... We need more realistic ground handling.

Maybe you could post a video of the ground handling you are seeing... 'cause I certainly don't see it.

Just wondering,
Do you fly an R/C plane in real life? or is all your exposure sim related?
I don't get it. :confused:

My thoughts exactly.

Norton
11-13-2007, 03:36 PM
I agree with you guy's! The 95% is everyone except you (josh0987) saying that it is not perfect or correct ground handling.

On another note, maybe Josh is talking about say Flight School Airport where you do bounce when hitting the curbs.

Were all talking about cement to grass, pebbles, imperfections in the runway, tail wheels bouncing etc.

int2str
11-13-2007, 03:56 PM
Just putting my vote in....
Ground handling is just plain silly.

pplace
11-13-2007, 04:08 PM
Where did you get the 95% from? :confused:


I agree with you guy's! The 95% is everyone except you (josh0987)

Exactly.....I was honestly thinking more along the lines of 99% as I can't recall anyone else that thinks the ground handling needs zero improvement. I figure I'm being a little forgiving with my percentage in case I was missing someone!!

willsonman
11-13-2007, 04:49 PM
though I agree that teh ground handling could be improved I also realize what a challenge it would be. First off you would have to create THOUSANDS of additional polygons just on the tarmac. Then for grass you would probably have to add more. The calculation of all of these polygons in real time would DRASTICALLY kill performance even on the hottest video cards. Its just not practical right now. Perhaps in the future. I think that the learning curve from going from sim to the field is one that you should expect. I understand that realism is part of the experience but I try and evaluate what COULD have happened on the field as opposed to what SHOULD have happened. It makes me less cocky when I get to the field and, I feel, makes me better in the long run. Just my $.02.

pplace
11-13-2007, 05:10 PM
Maybe on my next model I'll make the collision meshes on the wheels....out of round? This would make the aircraft "wobble" and "shake" as you roll across the ground? Granted there would be no difference still from the tar to the grass, but it would be a rougher riding take off and landing???

Any opinions? Would that do anything??

Zelatio
11-13-2007, 05:16 PM
though I agree that teh ground handling could be improved I also realize what a challenge it would be. First off you would have to create THOUSANDS of additional polygons just on the tarmac. Then for grass you would probably have to add more. The calculation of all of these polygons in real time would DRASTICALLY kill performance even on the hottest video cards. Its just not practical right now.


I believe you are wrong there. The ground is made using a depth buffer, correct? Why can they just make parts where the depth buffer has bumps? There are no polygons in photofields also. They are just buffers so when an object hits something, it wrecks it.

0xdeadbeef
11-13-2007, 05:51 PM
Besides, if water handling is possible, why would ground handling be not?
That being said: of course you don't need additional polygons for better ground handling. There are dozens of ways do implement it without adding polys.
Look e.g. how Reflex XTR handles this.

phrank
11-13-2007, 05:52 PM
You're on the right track pplace, for example:

CAP232 from Rick Delten I believe has a "rough" collision mesh for the wheels, that one at low speed you can see it wobble around the field.
http://www.knifeedge.com/forums/downloads.php?do=file&id=615

Also look at Doug's variant of Maxkop's Sikorski, that one bounces around quite a bit, brobably more than usual, but could be tamed down some with the right adjustments to the springy landing gears.

All good leads, but "rough" depth buffers would certainly give us better tools to bring up the realism on 3D and Photofield airports.


Maybe on my next model I'll make the collision meshes on the wheels....out of round? This would make the aircraft "wobble" and "shake" as you roll across the ground? Granted there would be no difference still from the tar to the grass, but it would be a rougher riding take off and landing???

Any opinions? Would that do anything??

willsonman
11-13-2007, 07:42 PM
I was unaware that Real Flight used depth buffering for the ground. Even if you were to use a "roughing" technique to make it look real for the entire map you would be limited in buffer bandwidth. The granularity would have to be quite high since you are working with scale models and not the full size thing. That high granularity would eat the bandwidth alive and the textures are already quite high in RF so bandwidth in general is already pretty taxing.
Using DX10 API there could possibly be less of a hit in performance with the unified nature of it. Just my opinion but I feel it is asking a bit much for this particular application. I think shaping wheels would be much more ideal.

pplace
11-13-2007, 08:07 PM
I just did a super quick test on the FW-190 collision meshes for the wheels...and I did a quick taxi around and the plane does "wobble" now.....but goes away at higher speed (take off) I think I can add a bit more "roughness" to the meshes without getting to "unrealistic"

If I do hit on a fairly good setup.....is there an easy way I can record and post the video here for anyone to view? I know I can just hit "R" in G4 but then only people with G4 can view it right?

There are other videos on here that you can just click and download to view?

redherring
11-13-2007, 08:09 PM
If you want to send me the model with some ideas of what you'd like to show, I'll do a quick vid for ya.

pplace
11-13-2007, 08:18 PM
I'm going to mess around / tweak with it some more to see how "active" I can get it! I'll keep the offer open

redherring
11-13-2007, 08:34 PM
Cool, it will be very interesting to see.

opjose
11-13-2007, 09:02 PM
Even if you were to use a "roughing" technique to make it look real for the entire map you would be limited in buffer bandwidth.

The granularity would have to be quite high since you are working with scale models and not the full size thing. That high granularity would eat the bandwidth alive and the textures are already quite high in RF so bandwidth in general is already pretty taxing.



He's talking about affecting the wheels only.

We have no control over the texture/bump mapping on ground surfaces.

That is "fixed" by KE.

pplace
11-13-2007, 09:03 PM
Well my wife needs to do some classes on the computer so I'm done for a while.

It's a very fine line on the collision meshes "bump" it's either way too much that the plane looks like it's driving over boulders......or I can get it that it looks very real during a SLOW taxi....but after any speed it smooths out too much!?! I can look closely and see the wheels "bounce" off the runway as it is taking off....but it really doesn't "buffet" the plane around at all.

I'll keep working on it a bit....will keep you informed if I get any solution

willsonman
11-14-2007, 08:21 AM
You could try recording with FRAPS.. free and pretty universal format.

r1derbike
11-14-2007, 09:01 AM
Yes...even more realistic ground handling, besides the touted more realistic ground handling is needed, however it may be achieved. No doubt. I vote yes!

Charles

0xdeadbeef
11-14-2007, 12:14 PM
I was unaware that Real Flight used depth buffering for the ground. Even if you were to use a "roughing" technique to make it look real for the entire map you would be limited in buffer bandwidth. The granularity would have to be quite high since you are working with scale models and not the full size thing. That high granularity would eat the bandwidth alive and the textures are already quite high in RF so bandwidth in general is already pretty taxing.
Using DX10 API there could possibly be less of a hit in performance with the unified nature of it. Just my opinion but I feel it is asking a bit much for this particular application. I think shaping wheels would be much more ideal.

Nope, shaping the wheels is just a workaround which doesn't address the real problem. Besides: a depth buffer is not quite the same as a Z-Buffer.

Anyway, the depth buffers are not really the point. The problem is that currently the ground is always flat in photofields. If it was made from polygons like in Reflex, each polygon could be assigned certains attributes, e.g. bumpiness, friction etc.
The bumpiness could either be defined by parameters for an algorithmic approach (e.g. likeliness and amplitude of bumps) or by a bump/displacement map.

Anyway, all of this doesn't really add complexity to the basic depth buffer. Indeed, using a polygon based depth buffer would even decrease the complexity of the depth buffer in most scenes, since at the moment, the depth buffer is constructed by a bulk of objects just to simulate a certain shape.

And again (for the 4th time or so): have a look at the scenery editor for Reflex XTR. This is how a depth buffer should be defined and KE could learn a few more tricks from Reflex (e.g. alpha transparency in photo fields).

redherring
11-14-2007, 12:40 PM
Excellent explanation Ox. Thanks.

redherring
11-15-2007, 11:21 PM
Video of the Tiger2 (http://www.wabashradioflyers.com/g4/g4_tiger2_springygear.zip) - normal and springy...am I correct that not all models can/will support springy gear?

For now I've given up on the Tiger2, the Tyro and a few others that had static gear. They are back to having stiff/rigid gear.

Oh well....one can only hope I guess.