kestrel7e7

Realistic artillery would be absolutely mind blowing in a well scripted single player campaign in ArmA scale maps! ... Here's to waiting for BIS game 2.

Once again... the only limitation to the number of AI, in any game engine, is your CPU and the quality of the AI. Sure, you can have thousands of dumb drones that bounce off walls aimlessly but if you want them to have any intelligence, you need to sacrifice quantity in favour of quality. Crysis has HDR that affects the gameplay realistically, clumps of trees have dark shadows that are hard to see into and looking out from a dark area into sunlight is blinding. ArmA really needed this: HDR that actually works.

Only CPU power limits AI capabilities in CryEngine 2, not forgetting the trade-off between quantity and quality. How many physics entities can ArmA simulate? :P Considering BIS had minimal financial backing compared with the likes of Crytek, the ArmA engine is quite an achievement.

CryEngine 2 is in a league of its own.

Excellent idea. Not sure if the die-hards would buy it though...

What about head movements? You can already nod your head up-down for "affirmative" and left-right for "negative". You could give these head movements other meanings if you wanted to and create more complex combinations. eg one left-right and one up-down means "I've got TrackIR and I'm stretching my neck."

Sometimes I wonder if I'm playing a war game or Seagull Simulator 2007.

I'm an electrical engineer, so I wasn't expecting much using a webcam as a head tracking solution, but I was pleasantly surprised. With the right settings, Cam2pan + Newview can give you excellent 2dof head tracking which satisfies all the criteria of a good head tracking solution: *Smooth: Newview smoothing fills in the gaps of the 30fps camera. *Stable: 4 percent floating deadzone, removes unwanted noise and lowers sensitivity to small head movements to provide a stable view at all times. Otherwise constant view movement makes it hard to see clearly. *Responsive: Â 30fps gives you only a 33ms response lag, unless you want whiplash by turning your head in under 33ms, this isn't a problem. This is a 2dof TrackIR beater in my books. People are making 3rd party 6dof TrackIR alternatives as I speak. My settings: 30fps webcam (eg Logitech Quickcam Express) Cam2pan: http://www.mousevision.com/game/index.htm Flashlight mode using wide angle (>30degrees half angle) head mounted IR led IR filter over camera lens (developed, fully exposed, camera film) or no background light. Video Format: 160x120 resolution with I420 pixel depth "FB/Lock On" Panning mode 20% X & Y sensitivity 0% smoothing (smoothing increases lag) 0% sensitivity Newview 37E: http://translate.google.com/transla....&u=http Get data from: External interface (Cam2Pan etc.) Smoothing level: 5 (minimal smoothing, just to overcome camera's 30fps limit) Threshold: 6000 X-axis Speed Factor: 2.5 Y-axis Speed Factor: 2.5 K1 10 K2 2 K3 1 Deadzone: Floating: 4% for X and Y (for stable view) Hard response curve with a dip: first values: 24, 44, 83, 97, rest 100. This gives a central low sensitivity area and the rest is a nice realistic linear response instead of an annoying floaty exponential response. Edit: ArmA head movement covers a smaller range than most TrackIR flight sims like IL2 so the results will be smoother. Slow head movements can have noticable steps due to the low resolution of the Video Format setting in cam2pan. This doesn't bother me because I prefer fast and stable head movements, similar to using a POV hat on a joystick. If your CPU is good enough, you can increase the resolution of the Video Format in cam2pan to 352x288 (TrackIR4 camera has a resolution of 355x290) with a significant increase in CPU usage, but head tracking will be as smooth as silk. If you have dual core, the higher CPU load won't be a problem. Try and use as much of the camera's field of view as possible when moving your head by positioning the LED away from the central axis of your head.

Click on the 'Open Profile' button in Newview and make sure the ArmA ini file is selected. Each game ini file is created with default settings when you play a new TrackIR enabled game, so you need to change the settings for each new ini file, either manually in Newview or by renaming another ini file. Make sure everything is set correctly in the helmet tab, especially 'Enabled at startup' and 'Get data from: External Interface'.

You don't have to configure ArmA for head tracking. Cam2pan must be in "FB / Lock On" Panning Mode, use nvtomouse to check that cam2pan is transmitting correctly. If it works in nvtomouse then it will work in ArmA. You might find this guide helpful: http://www.simhq.com/forum....fpart=2

IR LEDs used in remote controls are normally medium angle, wide angle is best, ie 80 degrees full or 40 degrees half angle. Be extremely careful using USB ports as a 5V power source because some motherboards can be easily destroyed if there is a USB overcurrent situation. I know that the Intel ICH5 Southbridge chips can be fried by static alone from the front USB port header in some ASUS motherboards, the rear USB ports are better earthed and less likely to do this. Overcurrent could easily happen if you are experimenting and accidently create a short circuit, killing the motherboard in a fraction of a second; using a PTC fuse rated at 0.6A cutoff can give you a safety buffer.

Yeah, let's make it like CS. "What do you mean, why am I holding a knife? I run faster with a knife. Everyone runs faster with a knife. " PP

First post edited with Russian Newview forum URL containing latest Newview download link.

TrackIR is no different, whether it be a reflective dot stuck on your forehead, Vectorclip attached to a baseball cap or a wired 'TrackClip Pro' it's still necessary to put something silly on your head. But isn't that what geeky head tracking is all about?

For aircraft, using a mouse in the left hand and a joystick in the right works well with a bit of practice. For a right-handed person, doing this can improve your left hand co-ordination in general.

Stop wasting your life being an unpaid beta tester, shelve ArmA and come back to it in two years. If you're lucky it might be a little better. The post-release development support that BIS gave OFP was only possible because of good sales figures. You can't expect the same support with ArmA if it sells poorly. Nobody works for free. And no, modders don't have the ability to properly fix ArmA, only BIS can do that.

Since when was realistic warfare meant to be fun? Â War is Hell.TM.

Just to be clear, those points were not mine, but someone else's in a different forum, I copied and pasted them for discussion (note the quotes) and highlighted the part about head tracking in FPS games. I can only assume the author had hands-on experience with TrackIR. Regardless, thanks for your comprehensive reply Barred.

I can appreciate that there is a lot going on under the hood of TrackIR, since it uses NaturalPoint's extensive optical tracking experience with OptiTrack and the cameradll.dll interface, but all that really matters is the end result. I believe, as in the university paper I posted, that good results (a practical alternative) can be achieved with common, commercial, off the shelf webcams and some relatively simple tracking software. Currently, amateur 3rd party tracking software is not up to the task, but it could be in the future, the biggest barrier to it being the proprietary TrackIR interface. For example, Freetrack shows promise, with good movement out of game but poor movement in-game because it can't use the monopolising TrackIR interface and few games support 3rd party head tracking. This leaves no other option but some dodgy backdoor route with a poor result. Back to topic, I found the following in another forum and thought it provided an interesting insight into the practicalities of gaming with TrackIR. "TrackIR has been around for some time. It's certainly neat but it has some major flaws, some of which were already pointed out here. As well as a near rabid fan base, usually flight simmers, who will whine you into oblivion if you dare speak out about it. 1. Moving your head but your display remains stationary. Meaning you need a form of unnatural acceleration curve to minimize head movement and maximize view. 2. Due to the acceleration curve needed you lose stable viewing at the extents of the units movement. For example I'm looking down my airplane wing at something, well it's not fixed, it's at the top of the acceleration curve. Add to this the difference between a snap view in a dogfight per say, and a steady view for calling out bombing targets. You can have only one profile. 3. Aiming with your cross hair anywhere but at center. Think about how you move your mouse to acquire a target with a fixed cross hair in an FPS. Now think about moving your mouse to a non-fixed cross hair. -- (This has been mentioned before in this thread and indicates there are some difficulties with using headtracking in FPS games.) 4. What I mentioned can be hacked and deadzoned to make a playable experience, but you'll need to map keys to switch acceleration profiles, return to center, enable/disable the device. Something no TrackIR user will ever admit to but all require to get a playable experience. 5. TrackIR (Naturalpoint the developer) enabled a cheat in IL2 where only TrackIR users could get a full 360 view in the cockpit. Something rather important if you're flying in public dogfight servers. The fanboys raged about "people with better video cards get better performance and that's the same", and other half witted statements. Naturalpoint refused to fix the issue and instead hid it in an easily found registry entry, wink wink, and posted the hack on the TrackIR forums. Similar to the ASUS driver hacks in most people's opinions. 6. It's saving grace, in my opinion, are racing games. All you need is 15 degrees of travel to look into a curve. In this respect it's perfect. All in all, neat piece of kit, but not ready for general gaming."

This paper shows how overpriced TrackIR is. "Simple, Robust and Accurate Head-Pose Tracking Using a Single Camera" University of Wollongong 2006 "In order to overcome much of the cost and deficiencies in existing head-pose tracking systems we have been developing accurate methods for pinpointing the position of infrared LEDs using an inexpensive USB camera and low-cost algorithms for estimating the 3D coordinates of the LEDs based on known geometry. Our system is comprised of a single low-cost USB camera and a pair of spectacles fitted with three battery powered LEDs concealed within the spectacle frame. Judging by our results, we believe our system to be the most accurate low-cost head-pose tracking system developed. Furthermore, our system is robust and requires no calibration. Experimental results are provided demonstrating a headpose tracking accuracy of less than 0.5 degrees when the user is within one meter distance from the camera." http://ro.uow.edu.au/cgi....opapers Edit: And yes, they are aware of TrackIR, it is listed in the references. Another Edit: "The computational cost is also extremely low, at less than one microsecond processing time per frame on an average personal computer for the entire three-dimensional calculation. The system can therefore easily keep up with whatever frame rate the video camera is able to deliver."

Thanks to all TrackIR users posting on this forum telling everyone how satisified they are with their purchase. Great to hear. I'm thinking of the future of head tracking, history shows that mass adoption of a new technology brings with it great advantages. If this were to happen with head tracking, more games would support it and the technology would be quickly refined. But this could only happen with a more reasonable price, mass production would then help to drop the price even further once it became popular, making the technology a standard. So far, NaturalPoint has managed to achieve a monopoly over the PC gaming head tracking market, which is every company's dream, giving them the power to set the price with no competition (*cough*Microsoft*cough*). They've encouraged developers to add support for head tracking but only THEIR proprietary head tracking interface, excluding potential 3rd party head tracking solutions. The developers allow this because NaturalPoint do all the interfacing and tweaking for them (and could even pay them). As the number of games that support TrackIR gradually grows it further cements their proprietary interface and blocks out competition. If Logitech decided to enter the same market with their own head tracking solution, they would have considerable difficulty contending with TrackIR. They would need to lobby developers to stop using the TrackIR interface and make a general interface from scratch that can accept any 3rd party head tracker. By buying TrackIR you're supporting NaturalPoint's monopoly and ensuring that head tracking techonolgy will remain a niche toy for uber-geeks.

There's nothing natural about rotating/moving your head while looking at a fixed monitor. NaturalPoint could have easily released TrackIR4 straight out like the freeware Freetrack equivalent. But no, instead they take their time, adding features progessively with each new release to make the previous model obsolete and encouraging people to go through the upgrade cycle. Instead of the baseball cap setup they could have released a more convenient and practical headphone setup outright, but no, they consider it an optional extra. ChaChing! $$$$$ Open your wallet and leave your brain at the door. Sucker born every minute. Â No wonder advertisers have contempt for consumers.

Overclocking a PS/2 port has merit if you move your mouse very quickly and require immediate response. In comparison, you would have to move your head fast enough to give you whiplash before a head tracking sample rate higher than 40Hz would be required. I've tracked TrackIR from its inception and the only thing that has become evident is that people are being ripped off.

Freetrack has been available for almost a year now. I fear the project may have stagnated. Unlike cam2pan in combination with newview, which is compatible with all trackIR games (but only 2dof), Freetrack requires specific dlls to be written and compiled for each game.

The PS/2 mouse interface samples at 40 samples per second and has been used for decades with Windows, CAD programs and FPS games. Cam2pan uses 30fps webcams, which is plenty, especially so for head tracking that doesn't require precise aiming. I believe NaturalPoint's 120fps claim is misleading propaganda. Higher sample rates correspond with increased CPU usage, if 40fps is enough for a precise mouse there is no need to exceed that just for head tracking. If their 120fps claim is correct, it could be because they are using a VERY low resolution camera that only picks up significant head motions. This low resolution image doesn't require much of the CPU to process and the resulting motions are filtered to give the illusion of smooth control. If this is the case then cam2pan is actually more precise in detecting small motions than TrackIR. CPU usage with cam2pan depends on your camera drivers. A multi-core processor with a multi-threaded OS can easily deal with camera CPU usage while playing a single-threaded game. Those 'parastic' movements would still exist if all 6dofs are being used and are translated into the game whether you like it or not. Using cam2pan, everything is planar, so you don't have to rotate your head to look left and right, instead you can slide you head so you can keep looking directly at your screen with the added benefit that your nose doesn't get in the way of one of your eyes. Cam2pan gives you everything you need for a fraction of the price.