Looking past struts.

[Charlie251]

Not sure what you have in mind for the 6DOF and such. I would hate to have to hit a button to change the mode from looking around to leaning around.

Wonder how much it would cost if you added another track unit that you mount to the side of your head. Then you have a dot or ALED on the side of your hat. That side unit keeps track of the distance to the dot. Then as you lean closer or farther from it, your body leans on the cokpit.

Of course you would need game integration for this to work as well. But it would sure beat having to move your head around until you are looking up at the strut, then hit a button and turn your head to lean around it.

[gaming247]

He Jim why don�t u try this:

The software is tracking the size of the dot.
(i think it already does for the filters, isn�t it ?)

When the size of the dot gets bigger u move forward in virtual cockpit. When it gets smaller - backward.

When the dot moves left without changing size the head leans to the left.
When the dot moves left and the dot gets smaller u turn your head in VC.

This could also been done with up and down.

When the dot moves up without changing size the head moves up in VC.
when the dot moves up and changes size u turn your head up.

This could already be done in FS2002, as it has a fully rendered cockpit and u can move left, right, up and down.
(not sure, if IL2 also has one. but i don�t think so)

Just an idea i had.
Keep up the good work.

[ January 19, 2003, 04:27 AM: Message edited by: gaming247 ]

[Jim]

Hello:

I am really enjoying our new Forums, we all getting a chance to discuss great new ways to improve the trackIR.

I have a plan for the 6DOF unit, but there are several ways to go about it, so ideas like these are very helpful in determining what a user experience should be like. I had acutally never thought of the idea of having a Hotkey to switch between translation and rotation control, could be good if done right and would allow for using a single dot.

The main problem of doing a full 6DOF system is that the games don't really support a full 3D rendered cockpit. I have mentioned my desire to do a full 6DOF system to the developers, but they told me that you would then be able to "see around" some objects that are really just 2D, like looking around a SouthPark animation, just cardboard cut-outs from the side view. In the future I think that games will be full 3D and then we can get to work. I will mention this to the developers as it would really be great to look around struts and shift around in your seat.

Some of the users like to translate for looking around instead of rotating their heads to change the view, this would no longer be an option, or I guess we could allow the user to bind them together. Wow, I can't image how to layout that interface.

I think that FS2002 is the best way to test this out, as it does have the built in functionality. It would be very fun.

I have thought about using the dot size as an indicator for some level of "z" axis control, like allowing you to zoom in. This could work but would only a single level zoom, in or out, as there isn't really enough information in the dot size to do it smoothly.

Using dot size to tell translation from rotation could work if we had a ton more resolution, but we really don't. Take a look at the Data Window and you can see the overall pixel count doesn't change that much when you move around. The real trick is using more than 1 dot or camera for more complex motion tracking.

Keep the ideas coming.

[Charlie251]

The idea of using a button to change from translation to rotation wouldnt be to bad. At least not as bad as I originally stated. It could be just another button that you can add to your joystick. In fact, once IL-2 FB comes out and everything is integrated, there shouldnt be a need for the view change to recenter the view. So I could easily remap the buttons that go from F3-F1 to recenter to the "slew view" button.

It would definitly work. Oh, my idea of having a seperate unit for the side of your head sucks. After I got to thinking on it, horrible idea. I would have to build a shelf next to my head to mount it!

I think a "Slew" button would be awesome, if the game developers would integrate it and not use the 2d bitmap for the struts.

Terry

[honda346]

I think the 3 dot triangle pattern posted by the guy on the Driver forum is the best so far. It should allow for all direction calculations.

Personally I really don't like the idea of having to modes of operation via a key/button very much. I just can't see going between both modes back and forth while you try to get just that right position you were looking for (unless it was simply the only way possible for a limited sim or something). The real show stopper is the 3d models not having the detail as you mentioned... I mentioned this in the other forum but I know Aces High let's you move your head around in all directions so it would work there already if they could just implement a straight trackIR enhanced mode into it... There are no "gaps" or "holes" from shifting your head there.

[gaming247]

hmm... my 3 dots methode should will work well and covers most movements.
We have only problems when TrackIR should track the distance between the unit and the dots.
That means if u are using trackIR like u are using it now (unit is in front of you, on the monitor) you can�t track a forward or backward movement.
I think the best compromise would be to have trackIR placed above your head looking down at you. With this method you could track every movement of the head. also forward and backward. The only thing u can�t get is the up and down movement of the head.(would need tracking the distance between dots and unit again).
For up and down u would need to configure an extra key. But how often do u need that ?
This would be a good compromise.imo.
You could do every movement beside up and down and only need one trackIR unit.

[ February 22, 2003, 04:41 AM: Message edited by: gaming247 ]

[DNME]

IMO if you want real 6 DOF headtracking you're gonna need something else than IR reflecting dots.

I'd then rather lean towards something that Nodak suggested (gyration mouse). IMO 3 trackIR units simply won't do and will be much fuss to setup. As much as i like the trackIR i think it's too limited to provide true 6 DOF headtracking. But for the job that it does now it's the best!

What would be cool maybe is to have a separate device registering the 6 DOF movements and something else tracking eyemovement.

[delta_9_tetra]

Great idea there! I would buy the technology to look around in a true 3D cockpit in a combat flight sim like a shot!

However I don't see a future for a system where you have to press buttons to change your view.

I got trackIR to make viewing completely seperate from what my hands are doing, like in real life.

[Jim]

Hello:

Very interesting thread here. I think that full 6DOF is the future for sure. We have been playing with it in the office for some time and have several different versions that accomplish the same thing.

IMHO you must use an optical system for 6DOF. You simply can't get the precision and repeatiblity with any other tracking method, at least one that we can afford. All gyro or accelerometers / inertia devices loose center over time and drift. As we all know, maintaining center position is very critical. The only other options are ultrasonic transducers and plused magnetic trackers. Both are cool but cost more money and tend to have other drawbacks.

Being a self proclaimed optical tracking guru I can tell you that we will have some great options for full optical 6DOF.

[Charvel]

I think it all comes down to resolution and computing power.

I was thinking along the lines of a micro scale GPS like system but that would probably be very difficult and expensive to implement (and no you wouldn't really need a satellite ).

Instead a system kind of similar to current TIR would be sufficient using optical recognition of a person's head and it's current position in 6DOF. There are a couple of hurdles with this method currently but nothing out of the realm of possibility by any means.

I would imagine the visual recognition system software would have to be built into the hardware to keep the CPU free similar to what is going on in current optical mice.

The most difficult aspect IMO would be resolution. Getting the necessary resolution to compare even to current TIR or an old 60hz mouse would be very difficult for an optical device running the optical recognition software! Even current TIR is just barely enough in resolution to be useful in games and it only has to deal with a highly contrasted dot moving in a 2D field!

On a totally different tack, as I'm by no means any kind of expert, I wonder if current technology could provide a simple RF beacon attached to our heads with a receiver that could distinguish position through calculating position (X,Y) and doppler (Z distance) in an accurate enough resolution to be feasible? I'm not sure how it could calculate X,Y this way but maybe even a combination optical/RF system...

Then again I'm sure you guys have already thought all of these ideas through long ago but it is fun to speculate!