Just shooting in the dark here.

One thing that is clear to me is that the subject of aerodynamics has been much better addressed both in modeling and documentation in flight/helo sims than aquadynamics has been for subsims. Most sub models have been pretty much flat and perhaps on par in their physics to the flight sims of the very early 90s. There have been only a few minor concessions such as momentum and surface state affecting surface cruise speed.

So, provided that Dr. Sid can get unclassified access to information he is quite likely to render a model of sub physics which is a few quantum leaps beyond anything seen in any game ever available on the commercial market.

Of course, one might ask in terms of game play how important is this. If you are not trying to dock the sub or conduct espionage at PD is this level of physics detail important to game play? I suppose that is hard to answer until such time that full range of behaviors are revealed. And again it goes back to what other aspects will be simulated. Recently, in another thread we discussed the cold war and the very close tracking of one sub by another. I would imagine that that of perhaps all possible areas might be the most sensitive to physics and fine control.

Well, sit back and enjoy the ride, since only Dr. Sid knows where this is all taking us! :)

I think getting the physics right is a fundamental building block for any simulation. Everything else in some way or another depends on how good that foundation is.

The most exciting aspect about this approach to me personally is the possibility of casualty and emergency procedure modeling that is far beyond anything we've seen in the modern subsim genre--actual accounting for flooding and changes in buoyancy, attitude, performance, etc. That sort of thing wouldn't be possible if too much of the physics is abstracted the way it is in the SCS series of sims. I'd rather see and experience the sub losing depth control than just be told by a voiceover that such control has been lost.

BINGO!

If you start off with a stong foundation, then you will get back in spades what you invest.

Couple this with the work he has done in ray path analysis, spice it up with what some of us can guide others on, and you will end up with something that will be a SIM and not an arcade game.

Since someone mentioned the comparison with dynamics in Flight sims in comparison to sub sims, it might be of interest to take a leaf out of X-Plane's approach to flight simulation. In contrast to pretty much every other flight simulator, X-Plane calculates the flight dynamics from the actual shape of the aircraft model in the sim, as opposed to most other flight sims where the 3D model and it's flight modelling are completely separate entities, and up until fairly recently, X-Plane was the work of just one guy too, which just shows you what can be done by someone working on their own.

From the website: 'X-Plane reads in the geometric shape of any aircraft and then figures out how that aircraft will fly. It does this by an engineering process called "blade element theory", which involves breaking the aircraft down into many small elements and then finding the forces on each little element many times per second. These forces are then converted into accelerations which are then integrated to velocities and positions... of course, all of this technical theory is completely transparent to you... you just fly! It's fun!'


More on X-Plane here:

http://www.x-plane.com/

:D Chock

BubbleheadNuke: I also derived some formulas for full auto TMA. ;)

Choke: For me, Orbiter is demonstration what one man can do. Challenge too.
As for dynamics based on 3d model, that would be very possible for subs too. But parametric model is easier to tune.

On the other hand I plan physics model based on parts. It is not so at the moment, this is something like first prove of concept.

Ship will be divided into parts. Part will be anything which can fall of, which can malfunction independently, and in same cases large parts will be divided into smaller parts so they can break apart. Sail or planes is the best example of such parts.
Dynamics will be computed for each part alone, then their connection will be used to affect the whole. Then you just loose the part, and everything works as it should.
The ship body is pretty large. For correct hydrostatic I should compute hydrostatic lift for all volume. If I divide the body into lets say 5 parts, each part will be easy to approximate by sphere for this.
3D model should reflect this too. This can thus be used for damage modeling too. You all have seen those pictures of frigate broken apart by mk48, right ?. Well it would not be 'exact' damage modeling, but this will allow pretty good damage model.
Also all sensors and crucial machinery would be parts, with 3d position and collision box. Missile hits frigate and blows sail part, but oops ! Antennas was mounted on it ! So no antennas any more. You manage to do lucky hit of submarine sail with 70mm shell. Now did it hit periscope tube hidden inside ? Did that maverick which hit frigate damage anything important ? Engine ? Tanks ? Missile stock ?

This in essence is what IL2 has. It's easy to do and it works great, for dynamics, damage control, it's also easy to create such model by non-programmer people (on DW we call them modders).

My goal is to make the configuration of the sim so strong, that no programming is needed. Possible, sure. But not needed. There will be parametric models of every sensor. Definitions of graphic interface, connection between the interface and sensor parts, I'm designing state-machine simulator which will be used on weapons (like you have to meet such and such conditions to do such and such action, which takes that time, moves that part, makes that sound and is visible here on the controls screen). Then there will be scripting language for extending this where needed, but only few expressions would suffice. No DLL (as in orbiter, for example), no need of compiler or anything.

I'd just like to add to what ML said. Physical fidelity is indeed necessary to realistically represent casualty situations involving flooding, loss of control, etc. Being a detail freak at times, I've always wished for a good diving station simulation that would let me operate the ballast control panel and trim & drain pumps to keep the boat in proper trim, the planes to control depth (depth keeping at PD in heavy seas is a challenge on the par with air-to-air refueling in flight sims), etc.

Remember how much bitching there was about the original physics model in Dangerous Waters? "Subs turn too fast", "It takes forever to change depth", "My boat is bobbing on the surface like a cork" ... ad infinitum. Players do care and a bad physics model can break a game as much as any bug.

:|\\

Seeing how far Dr Sid has come has made me put 3DS MAX back on my PC.

sid will I have to build my models any differently to take into account a frigate breaking apart? i.e. I'll have to break the hull into two or three sections so that it can break up in the sim?

Also if the physics model is correct then the general rules of physics should apply for any object in the sim whether surface, submerged or airborne so we don't get crazy behaviour of missiles and the like as we have in DW.

I don't know where Sid is at in terms of contributions to programming but surely I can't be the only model maker interested in doing stuff. At the moment I use .3DS files but maybe Sid can look at building the sim in such a way that GMAX the free modelling program can be used too.

When REALLY stomping on the power a 688 could roll quite a bit due to the sudden increase in torque, soon after the speed had caught up to the turns the ship righted itself and had maybe a degree or two of roll. I cannot remember it ever being a problem.

If I remember the books right, there WAS a 688 that had a 'trim tab' on a stern plane (starboard side) to dial out this slight roll. It was not worth the extra weight and complexity IIRC. Also, 637 class had counter rotating screws but they considered it too expensive and complex for general fleet rollout.

Xabba: I would not bother with hull parts at the moment. Shape and textures are the most of the work. To divide the model can be done at any time. But think about it. There will have to be some structures inside too :|\\
Those pivot points of animated parts are more important now. Some masts will come very handy soon too.

I hope there will be more modelers, but for the moment one good sub is enough. Also many poeple thinks they can model, but they can't. Look .. we don't even like some SCS's models, especially because of accuracy !

As for the 3d formats, I can implement and even reverse engineer almost any format (did it before). I have some experience with OBJ (almost any software can export to it), LWO, and now 3DS. GMAX should be no problem too, looks quite 'readable' in hexa. Any idea if it is described somewhere ?

Not sure since GMAX is the free baby 3D MAX I would have thought there was quite a bit of stuff on it.

I'll do the model pivots points tonight at work and e-mail before I leave, I'll also see if I can find some scopes I did.

As for internals I guess a few internal braces can be done and textured.

Interesting. My experience is with twin-screw boats, so torque effects were never an issue.

:arrgh!:

It's interesting how you guys are focusing on the simulation's failings with regards to control physics and how the boats behave in response to rudder orders, say. Nobody here has complained about how the simulation doesn't really do that great a job of simulating how sound behaves in the ocean either. Nor does it really model other important technical details, like signal processing, very well either. For example, nobody has to worry about how a beamformer works or what not. There is no measurable Doppler effect. The effects of ducting are modeled fairly poorly if you ask me. In general, any sonar effects originating from the wave nature of sound will be absent. There's no such thing as a Fourier waveform in DW, hence things like harmonics never show up on the narrowband display. Sometimes I think that if I was going to make a subsim, instead of simply having a list of frequencies coming from a submarine, each having similar magnitudes, each submarine ought to have a set of Fourier coefficients attached it it.

What they have is a set of displays and they each have their own rules, but no over-arching model seems to unite them so that there's a underlying logic they all share. If you look at DW from the perspective of an operational model, it's okay, you still capture a lot of the "bigger picture" aspects of a certain scale of naval combat, but if you think of it like a flight simulator, then it leaves a lot to be desired.

I'm not sure how much one really wants to go down this path, though. Warships have crews precisely because there's more technical details to take care of in say, one radar set, than one person can really do by themselves.

I think Dr Sid is wanting to address those deficiencies too but this is a testing ground to get the model physics right and then progress onto the next phase in order to tie it all together later.

If "here" means subsim, then you haven't been listening. Gripes about DW's acoustic model are quite common. If "here" means this thread, then it hasn't come up because the good Dr. hasn't modeled the acoustic environment yet, so there isn't much reason to discuss how the CSS is superior in that regard. Don't worry, that issue will come up when the project reaches that point, and he already has demonstrated his ability to create a better acoustic model than SCS has so things are looking good.

Okay, after playing with it for a while I do have a comment.

You are about right in the low speed handling.

Are you using a linear curve for response time or is it dyanmically calculated based on plane angle and speed?

The reason I ask it that while the slow speed handling is about right, the high speed stuff (25+ knots or so, seems to be slow by about 20%-40%).

I base this on old memories, but I remember doing JAM DIVE drills and was standing aft AT the stern plane control station and it just seemed that the down (or up) angle would come on a lot faster than what you have modeled.

I also remember that the time from you have to initate a back emergency (as a throttleman) is a LOT less than the time it takes to go deeper than test depth in your model. You have a VERY short timespan before the boat exceeds test crush depth during a jam dive.

Basically, the boat needs to have MUCH quicker pitch movement at higher speeds.

Both rudder and planes response depends on square of speed (and since LA has about same rudder and rear planes the amount is same too). But ! There are forces which work agains planes. Dynamic stability of the body in the first place, and it's amount too raises with square of speed.
Also sim3 model does not use front planes. I already have new model with front planes. They do not pitch the boat as the primary effect, but as they move center of the boat down, dynamic stability will do the pitch anyway. With front planes the down-pitch can be done a lot faster.
I'll release sim4 with forward planes today I hope. I'll check the pitching effect and how it depends on speed exactly, it's quite possible there are some bugs in it.

Here is the new version.

http://roger.questions.cz/other/sim4.zip

Changes:

-Bank controls are gone. Screw reaction is disabled, until better simulation of torque changing with speed is made (well described in that PDF from Bill !).
-You can control forward planes (on the sail). Use Z,X,C for that. They push the sub down or pull it up without changing pitch much. The sub will pitch anyway since it tries to point where it is going, but it can be countered with rear planes (and partially with planes autopilot set to 0).
-Forward planes are animated. Rear planes will be animated soon (Xabba has to change to model a bit).
-Simulation was remade so it does not depend on FPS anymore. Now physics is simulated every tenth of the virtual second. No matter what FPS or time compression you have, all simulation runs perfectly same. Intermediate positions are only computed for graphics when needed.
- Did I say time compression ? Yes ! Same controls as DW, except slash / resets it to 1.

BTW depth autopilot is still the same old one, it does not use forward planes but angle.

FYI - I think one thing SC/DW were missing for better game play was the ability to have higher levels of time compression and to have time compression drop out triggers which is common to manual flight and sub simulation games.

Eye eye sir, will do sir :arrgh!:
Well DW seems to have troubles to do real 32x .. it is much slower than that. Silent hunter uses simplified simulation at higher compression steps, and allows it only in some conditions - like for example 'no enemy around' .. so you can drop sound simulation or collisions. Dynamics alone are quite simple. AI does not have to be computed much often. Collisions can take some time, especially with tons of bullets from CIWS. Also some models of sound propagation I'm thinking about are quite CPU hungry. I just can't say now what will be possible.

But triggers ? Simple .. for sure !

Did you get the Typhoon I sent?