Traditionally, calling up these views have snapped the view between different modes. With the introduction of the high-quality virtual 3D cockpit with free zoom and pan, these views are nothing more than preset parameters storing line-of-sight (LOS) direction and zoom level.

Why not implement the traditional views as programmable viewing presets? Reserve a set of viewing keys, say the num-pad, that can be programmed. By holding down Control (say), you store the current viewing parameters (LOS and zoom) into the key. Pressing the key alone will recall the setting, and the view will transition swiftly to the preset view. This is just a temporary glance (ie. the view returns to original parameters when the key is released). Combining the key with Shift (say) will change the view to the recalled preset permanently.

Of course, the sim should provide usable defaults for the presets, but these can be overridden by the user easily as described above, even in flight.

All changes in viewing parameters should be done by smooth transitions to enhance the spatial feeling (awareness), and to provide a consistent simulation of the virtual cockpit (ie. the rotation of the camera/pilot head). Of course, the speed of the transitions need to be fast --- ideally it should be user configurable.

Manual panning should be smooth and configurable as well (speed). Modification keys (eg. Shift) should allow two speeds for more control (ideally we all should have had trackballs on our thumbs, but until then... :-).

Now, add a general view padlock mechanism that locks the LOS vector to any visible object (in air or on ground) close(st) to the current LOS. If no object is found and the LOS vector is pointing towards the ground, the coordinates on the ground is locked. If pointing towards the sky, the compass direction is locked (point at infinity). All viewing keys, manual pan and presets, break the padlock mode, but for glancing the padlock mode is automatically reissued when the view returns to the orginal direction (eg. it does a new object search). The padlock mode should be visualised by eg. drawing a targeting cross (and putting "Padlocked" in the corner of the screen?).

So what are my main rationales for these proposal? First of all, promote a single integrated and consistent in-cockpit viewing solution based on the virtual 3D cockpit paradigm (ie. the way things would look and behave if relayed by a camera in a cockpit). Secondly, simplicity and flexibility.

----

How would all this work? Here's how I envisage it in action:

I am manually panning the sky for bandits, and find one over my right shoulder. I know he's over my right shoulder because I see part of my aircraft's right wing, as well as canopy reflections/markings. I press KP* (padlock key), the sim checks if any object is close to my LOS, finds the bandit, and locks my LOS vector on him. This is confirmed by the target cross. I zoom in and do a visual identification. Luckily, it's an old MIG-23.

As I maneuver to deny him a firing solution, I press KP5 now and again to glance forward (preset) to check horizon/attitude. The camera rotates smoothly and gives me a good spatial sense of where the camera is pointing. When the camera returns to the over-the-shoulder position it repeats the padlock search. If the bandit is still visible it re-locks, if not I have to manually search and padlock again.

Now a missile launch warning sounds. I press and hold KP3 and the view rotates forward and zooms in on the Threat Warning Display (preset). It shows that the direction of the tracking radar is at 9 o'clock. I press Shift+KP4, the padlock breaks and the camera rotates to the left. I start panning manually to locate the missile trail, when I find it I padlock again. The padlock locks on to the smoke, as the missile is to small to be seen. This forces me to correct by manual panning, and to use judgment of where the missile is heading. Now I press KP8 to glance in the mirror, and to my relief see the bandit go up in flames courtesy of my wingman.

Like I said, it's not exactly what you proposed but it's probably the closest of all sims today.

Of course, there's always room for improvement.

Blaze

Kewl Idea but some things to keep in mind:

1. Target Step.
If there's more than one target you have to 'step' the Padlock. This applies to A/G especially where increasingly 'real' (iconic as opposed to texture) targeting gives you multiples of buildings as well as vehicles and even (_the worst threat_ of the modern battlefield, IMO, JF-18 having none so far): MANPADS.

Depending on how badly you screw up the BVR and what kind of threat nation you're fighting, this may also apply to multi-aircraft fights and even vs. a single airframe, you don't look where you last saw but where your (subconscious) inertial-memory says he's /gone since/.

I hate the time-waste of 'search-dither' padlocks. I'm counting the millseconds while the view translates!

2. 3D Depth (Limit) of Visual Field Cues.
If you emulate camouflage and/or are at a shallow slant (long grazing angle across the terrain) you're '3D' view becomes complicated by having several objects which might in fact be 'BVR', or too close to resolve with even an aiming sight. You've got to have labels that ID what you're looking at.

3. Sensor Cues/Marking.
One of my biggest hangups (and there are several shortcomings in the Janes system, let me kid you not) is the lack of JHMCS being able to cue the FLIR/Radar boreline or vice versa with wide offset targeting diamonds.
In the older systems (VTAS and the Schlem) you /can/ cue the eyeball back to a target by
flashing LED, compass directionally until you flash-lit all four quadrants.

In the newer ones you can actually see a trackline with limiter radial (a large curving arc with a marker point along it) 'until X' cue. This is good for making oblique entry to a fight with FLIR track of the bogey and 'steering' via helmet point indexer angle.

AND. Especially for A/G marking, you can store these or independent (Mk.1 Eyeball) sets for reacquisition. CAS and even some hard target strikes are unbearably complex in JF-18 because they refuse to let you find, store, serve targets on an auto-recall basis.

4. HOTAS and Transition.
I /like/ the idea of an autozoom. It helps you 'see the dot and the target' (which the dot represents) in close proximity say. This would make judging ACM spacing a lot easier.
In A/G having a central tele mode in the middle of a standard view (concentric) would help you slew the center of interest /quickly/, without losing track of where you were actually looking. This as much as anything keeps you spatially aware.

BUT. View translations today are among the most-slowing of processes. And to do what you suggest requires that Viz Mark, Sensor Tag, RWR-->Outside Threat Track, Glance Forward, 4-8 angle/zoom presets (Num Keypad Map-acrosses) and Recall ALL have to be on the stick. And probably a quick-mode switch for A/A and A/G targets since the very system you describe has an air-threat situation developing into a surface-to-air one and you would never be able to sort all the labels if you displayed both simultaneously, IMO.

Or at least I would, thanks to some moron we persist in putting the keypad on the right instead of the left where it should be for about 70% of the population and I refuse to 'cross control' hands, even for something as critical as you describe.

Last but not least, your system of threat engagement is way to complex. It needs to be linked to a _HUD_ display on bearing and closure (the velocity vector is terminal) and if there is an HMD, you need to have an auto-pan to which looks at even the smallest 'dot' as a long line lead computing system like so:

\
\
\
\
-----o--|-|-

Where the \ line is the rough missile lead angle relative to your 'o' airframe direction of flight, the first | bar is the point where the predicted feet per second and steering/seeker 'advantage' puts the missile LOS rate ahead of you're direction of travel and the second |(* whatever) represents the likely terminal impact point.

This would help you beam the weapon while keeping your head _outside the cockpit_ and if, through a combination of manuever (angular lead), range/speed/burnout differential (raw fps of closure) or guidance failure (EXCM) you 'beat' the weapon, then the secondary bar would flash back /aft/ of your airframe 'o' dot and you would know you were safe.

But IMPORTANTLY, you would still judge when to 'make the break' based on visual size/speed, in-frame with the missile. JF-18's Padlock Threat system is so useless for me that I find myself using the EW panel and a mental clock most of the time and that is /scarey/.

All of this is _Very Possible_ with a Missile Approach Warning System such as the F-18E will certainly have in the CMWS.

Kurt Plummer

P.S. The only thing remotely like what you describe in the JF-18 system is the 'Padlock Package' key, wherein Shift-F4 (not on the keymap I might add) lets you look at the guy you're supposed to be escorting, or at least (magically) 'in that general direction', no matter the viewing range.

The rest is pure ATF with F4 'target sensor' only working on the locked target and then only as long as it stays inside the sensor field of view. F3 frequently not translating to the threat /at all/ and then often showing me nothing but sky. And F2 being a 'center out' that doesn't seem to work with the <, /\, \/, > keys to manually slew the applicable search-start point.

Keypad numbers work for snap views but include only the one, ho hum, 'up and forward' view rather than a 'translate to here and STOP!'/snapback (second press) system like I'd like.

Regarding the need to cycle between padlock candidates in crowded environments, that's a good point. I regard padlock as an intelligent AI function doing visual analysis to help me out controlling the camera in the virtual cockpit. I know you are an advocate of even more explicit tools bordering on the 'magic' category (like your SA sphere visualizing and numbering the threats around you for easy view control). Still, if you could keep these tools working under the constraints of what the sim models (a camera in a virtual cockpit), I see nothing wrong in some innovation here... :-)

Here's an idea for the cluttering problem: Press KP* (padlock). If the padlock system finds many padlock candidates close to LOS, bring up numbers on the candidates. Then press 1-9 to select or use Tab to cycle (Flanker 2 does the latter). In my view, padlock candidates should only be searched in a small radius around LOS, hence requiring manual precision panning (here we might disagree --- I am an advocate of efficient manual panning. I think it is an essential skill adding to the experience. We need more flexibility in the sims and better input devices for the future, though).

Regarding depth-of-field, this could (should?) be a constraint on the padlock mechanism. Only targets that are deemed visible are searched. And of course, the padlock AI should prioritize targets in order of distance (combine that with the numbering scheme), and maybe threat (thus prioritizing military vehicles before buildings for ground objects).

In general, make the padlock AI as smart as needed, but don't compromise the basics, ie. a consistent viewing model (a rotating camera/head) and smooth viewing (the slider Blaze mention solves the speed problem). The rationale here is spatial awareness --- rotation intuitively tells you how the camera is pointing, hence clarifies camera attitude and consequently aircraft attitude. I believe a smooth forward glance would be very helpful here. It's about spatial awareness.

Let me illustrate. The first set of images below illustrate an instant transition. Note that you can't judge from the two images where the camera is pointing in the first image, except if you know the sim and recognise the mini-HUD symbology, then you can make a educated estimation.

The situation is made clear by a smooth transition, illustrated by the second set of images below.

Even just two transitional frames fully clarifies the spatial relationship between the views, hence the camera attitude in the first image. And since just a few transitional frames are needed, it doesn't have to be slow.

Regarding zoom, yes, that should be required for a complete virtual cockpit system. Both for internal (instruments) and external viewing. Having zoom stored with preset view keys makes it functional. Also, the zoom in/out function should have configurable speeds (as well as a speed modifier, say Shift). Add seperate preset zoom keys as well, eg. 'normal', 'wide', 'visual id', 'instrument'. Why not? :-)

Your overlapping tele-zoom in the center of the screen is innovative, but I think there would be less need for it if you could easily toggle a fast transtion between normal and high zoom. This would be consistent with the viewing model, thus my preference.

Regarding evading SAM missiles you wrote:

quote:

---

I find myself using the EW panel and a mental clock most of the time and that is /scarey/.

---

Maybe that's good! I recently downloaded a narrated training track for evading the Hawk SAM in Flanker 2 (courtesy of Rich at the Ironhand site). His skill in locating the missile smoke, judging the threat, timing and executing the evasive maneuvers is impressive. Maybe that's what a game is all about? Skill under constraints achieving success?

That said, I would love some sort of missile path symbology/prediction as training and analysis tools, especially for recorded playback. Add those even more technical suggestions for visualising the missile perspective/seeker you sent me a month ago! And the helmet sensor cues you mention for maneuvering into a fight... cool! :-)

MiG Alley is the best. It has the perfect viewing system in my opinion.