another nail in the coffin of the sim industry

[MTFDarkEagle]

Now let's give the fella the benefit of the doubt. Be constructive or stop beating the dead horse or you'll only steer the thread into a dead end.

 

I fear it is already going into a dead end... But I'm still willing to help, but he has shown no sign to accept it properly..

[EtherealN]

Okey guys, seriously, let's chill and wait for the man. No point adding pages to this until and unless he returns with a wish to gain assistance and the information requested.

 

Until that happens, let's just let it all rest.

[sweinhart3]

Red box only shows where your controls are, it does not show where your controls should be for hover. You can't know if controls are within 20% zone just from red box, because this zone is not absolute, it is relative and depends on your aircraft attitude and forces (like wind) acting on your helo. The only way to know if controls are within that zone, is to know exactly where controls should be for hover, but if you are constantly drifting, you obviously don't know where they should be (don't worry, most of us don't know in advance). You need to trim helo so that there was no vector line on hud, and only then engage the hover mode. Hover mode is not supposed to remove the drift, it is supposed to prevent the drift (this is not word play, it is important). If you can't get rid of tiny vector line, hit trim button a few times after engaging hover mode - it should settle the helo.

 

I actually find that just hitting reset trim before engaging auto hover still results in some typical backwards drifts. I usually have to trim my attitude again a little bit forward to counteract that. Once Ive done that my drift is usually near zero.

[sweinhart3]

I see nothing in BS that requires much cpu and gpu resources except the graphics - in fact the physics belongs in the category that nvidia say "will run slower on a hardware accelerator"

 

does BS use the PhysX on my GPU ?

 

BS not only simulates flight dynamics but avionics modelling, damage modelling, ballistics moldelling, and plenty other stuff. I almost went to work for a company that makes and operates full motion professional flight simulators. If flight dynamics are simple to accurately compute, why do their simulators require 6 top of the line networked servers to run it? ED is trying to squeeze as much as they can onto 1 pc, and obviously if the cpu is a major bottleneck, then that alone suggests that the amount of computations being required to run on a single cpu is near the max it can handle.

[topdog]

Where did you get that line? DxDiag? If so, "English Final Debug"indicates that you are using DirectX for developers (debug version). It is slower than "Final Retail" and might take its part in your low FPS problem.

 

Just to clarify; systems will often have a couple of debug directx dll's alongside all the retail ones.

 

The "(problems)" are related to the warning it gives on the relevant page of the dxdiag app, which you have described adequately above, but probably isn't actually in use.

 

 

To be using it, you would have to do have done the following:

 

1. Have installed the DirectX SDK

2. Used the DirectX Control Panel app (which isn't actually IN the Windows control panel, that's just what it is called and would be found off your DXSDK start menu, aka dxcpl.exe)

3. In here, specifically selected to "Use Debug Version of Direct3D 9" on the D3D9 tab and applied it.

 

You would also be able to see this option if enabled under HKLM\SOFTWARE\Microsoft\Direct3D\LoadDebugRuntime with a value of '1'. If it has a value of '0' then you're using the retail .dll's anyway.

 

So just the sight of this .dll being reported in the dxdiag is fine, it will do that just from detecting that it exists, and isn't an indication that it is currently being used unless you verify by these other means. You shouldn't play with it or remove it if the above registry key doesn't exist or has a value of 0 since those files will be completely harmless to your performance.

[ChromeWasp]

my system specs are well above *minimum*

and all well above *recommended* except for cpu

 

I have 4gb physical 3.25gb available to winxp

gpu is GTS250 with 512

 

at 1024 it is mostly ok - but still dips to 10/12 fps regularly

I wrote black shark startup script that shuts down all non-essential services and system trays etc.

nothing is interfering

 

yes i have debug version of dx as i am dx programmer

no it doesn't make any difference unless i use dx control panel to activate it

 

using av8r-01 joystick with trackir 4 pro

i don't like heavily sprung joysticks as i am used to flying with finger pressure

av8r lets me adjust the springs (couple of cable ties do it nicely)

 

yes i understand the old and new (central position) trimming system and its problems

yes i have set adequate deadzones to avoid locking issues

yes i have tested what happens if I don't return to center

I understand how the autopilot works

I understand the difference between active trim (fd off)

and passive trim (fd on)

I know when the AP trys to maintain airspeed versus pitch

I know when the AP trys to maintain heading/track versus bank

I understand the autopilot has only 20% authority from the trimmed positions

 

I have tested all my assumptions and the AP works as expected

 

EXCEPT

 

I believe there is a bug related to auto-hover, here is the simple proof:

 

- if I activate auto-hover with fd *on* and I follow the guidance, I will return to hover point quickly and easily and with nowhere near 20% deflection

- if I activate auto-hover with fd *off* the AP *should* follow the same guidance

 

but it does not - in fact it cannot overcome even a 5 kph velocity vector

now that could be due to wind - so I tested it on a mission with 0 wind

 

further proof - when the auto-hover is holding the hover point - you can FEEL and SEE the AP working with changes in attitude

 

but when it is drifting - you can feel NOTHING - it simply isn't trying

now if it tried and failed to overcome the trim or wind or whatever - I could accept that

 

why would you spend millions on developing an attack helicopter optimized for a single flight crew

and then create such a weak auto-hover? that is precisely the feature that allows the pilot to handle the workload

 

NEXT POINT about "realistic" simulators

 

using computer monitors and trackirs isn't the same as reality

many things are different - in reality I would glance at a switch and then my hand would find it

with trackir i have to stare at it, hold my head still, and carefully guide my mouse

 

more importantly you can not feel any forces in your body and also the resolution is never as good

 

this means that to make a computer simulation feel realistic - it has to depart from reality

to overcome the differences

 

a crass example is the size of doors in video games - they are 10 feet wide and 16 feet tall!

if they were sized like real doors you couldn't walk through them like real doors because

in the simulation your senses are DULLED

 

this applies to this case - maybe the auto-hover specification is *exactly correct*

but then that is the mistake - as we will never have the same sensory information as a real pilot

the simulation now STOPS being realistic and starts being PEDANTIC

 

in real life I should be able to come to a hard tactical stop - down to *almost* zero

and engage auto-hover and concentrate on my weapons - that is the whole point in the aircraft

 

this simulator fails to simulate the most important feature

[ObvilionLost]

Post track first or bugs did not occur.

[pelisilma]

Try to unbind your hover button and fly without it

[Glowing_Amraam]

NEXT POINT about "realistic" simulators

 

using computer monitors and trackirs isn't the same as reality

many things are different - in reality I would glance at a switch and then my hand would find it

with trackir i have to stare at it, hold my head still, and carefully guide my mouse

 

more importantly you can not feel any forces in your body and also the resolution is never as good

 

this means that to make a computer simulation feel realistic - it has to depart from reality

to overcome the differences

 

a crass example is the size of doors in video games - they are 10 feet wide and 16 feet tall!

if they were sized like real doors you couldn't walk through them like real doors because

in the simulation your senses are DULLED

 

There's a difference between what you, the human pilot, can physically do to interact with the sim, compared to how the Helicopter is simulated within the environment. This Sim *simulates* flying the KA-50, and it does so pretty darn close to the real thing, after what we've all been reading from legitimate sources.

[martinistripes]

Everytime you post I think maybe this guy is coming round, beginning to see the light then, you go ahead and drop another bomb like...

 

this simulator fails to simulate the most important feature

 

...and I :lol:

 

Funny thread.

[beers]

RAM is handy and a nice GPU is maybe helpful, BUT CPU speed is nearly the only thing that matters here. BTW, multiple cores hardly matter either.

 

(maybe i missed it, but) ...

what

is

your

CPU ?????

 

as you note that you are a programmer i don't expect we have to provide instructions on how to identify your processor type and speed, but if you are unsure, please let us know and we'll provide the extra 3 sentences of instructions. If you don't detail your processor type and/or speed, or provide the DX output, I don't think you are really taking part in this conversation between our community and you, a new and welcomed member.

[MTFDarkEagle]

(maybe i missed it, but) ...

what

is

your

CPU ?????

I second that..

[ZaltysZ]

I actually find that just hitting reset trim before engaging auto hover still results in some typical backwards drifts. I usually have to trim my attitude again a little bit forward to counteract that. Once Ive done that my drift is usually near zero.

 

"Reset trim" will always (with 0 wind) result in aircraft drifting backwards. This is because "reset trim" puts controls into default positon, which is position for compensating helo tendency to roll forward while it is standing on the ground. In other words "reset trim" trims your aircraft for parking, not for hovering.

 

Trim button not only affects trim, but also gives commands to stabilization system. While hover mode is on, trim button sets new hovering point. AP needs time and effort to bring helo to designated hovering point. With every trim button press, you basically tell AP: "Forget that hover point, which are far away already; lets hover where we are now". Carrying this multiple times will make AP slow the drift to zero. In contrast, if you have designated hover point only once, AP will need not only to kill the drift which is present at the moment of hover point designation, but it will also need to drift back to that point. With not enough small drift, this may take too long or even totally fail, so it is better to hit trim button multiple times and kill the drift this way.

 

P.S: "Reset trim" is optional game feature. Real Ka-50 does not have such thing as "reset trim".

[ChromeWasp]

a reasonable technique is to trim the helicopter for hover

then switch FD on and HOVER off for manoevring

then as you pull up into a hover turn FD off and the AP will mostly take care of the rest

you can use shkval to designate and scan then ...

once velocity vector is zero activate HOVER and confirm holding

then ATTACK!

 

that is 2 button presses to exit hover and 2 to regain

 

(the other technique is that I learn to fly better)

 

BS is the best simulator i have :)

 

(re:cpu I am ashamed to name it)

 

where can I find some good track files?

[Avilator]

Ummmmmmmmmmmmmmm...........................................what's your CPU?

The only reason you don't name it is because you're ashamed? That seems like it might be the source of your framerate issues.

Also, you are suddenly calling this the best sim you have? That seems like a very different attitude than your earlier posts in a really short period of time. But hey, if you like it, it's that much better.

Use the search function to find track files.

[MTFDarkEagle]

BS is the best simulator i have :)

Wait... :shocking: what????

 

(re:cpu I am ashamed to name it)

Just name it.. please?

 

where can I find some good track files?

I'll add mine I just made, dont mind the dodgy landing :P Note however I never fly with FD on.

 

Oh, it was a 7 m/s sidewind in the hover.

As I engage the auto-hover, the helo pitches to the point where I initated the autohover. FD is never on in my plane.

In the F2 view, in the bottom of the screen you can see SPD, that's the true speed, if it's 0, its 0. You are not moving, note it is zero during the hover.

Hover.trk

Edited August 10, 2010 by MTFDarkEagle

[HitchHikingFlatlander]

Is it a pentium 3 overclocked to 1.0ghz? I wouldn't be ashamed but I'd know that it can't push DCS-KA50 based on the box specs.

[EtherealN]

Chromewasp, seriously.

 

You are not getting any help whatsoever with your performance issues until you name your CPU. I say again: unless the GPU is something silly like a GMA 950, 99% of FPS issues are due to the CPU not being strong enough.

 

If you don't understand why it works like that, read back in the thread and you'll find it explain three billion times, but just for clarity: there are two things that work in "frames" - the graphics and the simulation. If the CPU is too weak to run the simulation fast enough then the simulation will drop in it's framerate, at which time the graphics are not fed information about what to display often enough to maintain the framerate and thus: framerate drops.

 

Seriously, this is not rocket science. Just upload that DxDiag already so we can help. Last time I am asking, after that I'm done trying to help you with your performance issues. I don't care if you are ashamed of naming the CPU - that's the one thing we really have to know if we are to diagnose the problem.

 

Seriously man, just do it.

 

EDIT: If that really is something to be ashamed of, just go ahead and PM it to me instead of posting it here. I promise I won't tell on you for using a Celeron or whatever it is. ;)

Edited August 10, 2010 by EtherealN

[Lio]

Here is a track file of me flying around in a city going into auto hover a few times. Its not hard :D

 

Nothing to fancy. Overall the flying is properly pretty bad but i got crappy joystick so i have to take hand off throttle to use rudder etc.

HoverFlight.trk

[beers]

not 'throttle', you mean 'cyclic' right? let's not confuse this ;)

[aairon]

Since ChromeWasp mentioned the GTS-250 I happen to have one laying here EVGA Gts-250 Oc 512 Mb DDR3 ram. I have a system running 32 bit vista ultimate and 4 gig 800 mhz ram (3.454 usable) installed so I grabbed the GTS-250 stuck it in there and run BS in the the third mission (not sure the name but a lot of ground unit firing and smoke from my rockets and the things I am blowing up and my frame rates when all of that is happening rarely drop below 40 FPS using the in-game frame rate display (using fraps will kill them, don't know why a frame rate software would murder ones frame rates lol) The only thing I am doing significantly different is that in that machine I am running a E7400 duo core at 3.6 Ghz with air cooling and an older Pentium 4 heat sink (larger surface area then the one that comes with duo cores) This performance is hardly any better then my balls out machine running a 9xxx series Quad core at 3.3 Ghz with dual GTX-295's lol. You would be very hard pressed to see the diffrence. FSX acts the exact same for me it wants as much RAW speed on the CPU as it can get and anything beyond two cores doesn't seem to impress it whatsoever. and the GFX Card seems to be pretty far down the list when compared to CPU speed and Memory speed + FSB these things are like gold with Sims of this type. That GTS-250 is plenty enough that you should be getting around 40 FPS when others are flying around you in multiplayer IF YOU had a decent CPU. So I think you need to totaly get that FPS "problem" out of your head, you have made a clear decision to have poor frame rates! that is something you are solely responsible for, not the way the sim was made. That leaves your only other complaint about the sim supposedly not being accurate and I think that was covered by Real life combat chopper pilots here already. I'm thinking it's maybe a good time for you to start a clean thread asking for advice from other more experienced pilots.

That's where I go for help.;)

[HiJack]

... I promise I won't tell on you for using a Celeron or whatever it is. ;)

 

:LOL: nice try EtherealN :thumbup:

[aairon]

I would LMAO if ChromeWasp were actually some sadistic ED developer, out to find out how well the minions are informed.:lol: That's really more believable to me then the thought that someone could actually be this way.

I can visualize it:

A group of Dev's sitting in the summer heat by a pool in the country side near Moscow sipping Vodka on ice and taking turns posting on the laptop, Laughing their asses off.:megalol:

Edited August 10, 2010 by aairon

[MTFDarkEagle]

He's not.

[vortex360]

different game, different engine

 

 

change your graphic settings.... everything doesn't always have to be on "high". The engine relys more on CPU then GPU.

 

I've been playing with the water settings on high lately. ATI's 10.7 has really improved the game. Here is a track of me starting up, attacking and then bringing it home.

Attack and land.trk

Edited August 10, 2010 by vortex360