Flight model:

A lot has been written about the absolute accuracy of various flight models but in truth the threshold value for this category is quite low compared to some of the other aspects of a good sim.

For example, consider Falcon3's flight "complex" and hi-fi models. Both have some severe defects when you look at realistic F-16 performance; start with sustained 9G turns in complex and the odd model change between 250-300kts in hi-fi for example. And yet, in their day Falcon3's flight models were not only state of the art, but also more to the point were touted by many fans as being the ones to beat.

Su 27 1.5

Now the state of the art has moved on considerably and expectations have raised the bar right along with the art. Nevertheless the threshold value is well within reach with current offerings. Su-27 offers probably the best high fidelity flight model in a fast jet sim today. However, EF2000 and Hornet3.0 both demonstrate that lesser fidelity can result in a flight model good enough to provide the all important suspension of disbelief.

Make no mistake, however, a sim that doesn't attempt to flight model angle of attack performance in such a way as to require pilots to think about and profit from efficient use of energy management falls well below the threshold.

Heading into Q4'98, Jane's F-15 must surely be the current champ in this department. Played on a machine with sufficient horsepower to give a decent frame rate at the higher graphics settings, the feel of the flight model just seems right. Stay in the "normal" flight regime (after all this is not a test pilot sim its an attack jet sim) and Jane's F-15 is about as good as it gets without a security clearance.

If we are to raise the bar in this category then it must be in slow speed flight handling characteristics. It is only here that Jane's F-15 is vaguely suspect. This comes into play in approach to recovery or "low and slow" turning fights for most game players. It would be interesting to see spin modeling or the (in)famous Viper deep stall factored in but these are really the province of test pilots not game players.

Frankly, I think we're pretty well off in this category already!

Graphics:

Su-27's detailed object model and sparse terrain detail has many fans. Indeed this approach brings considerable benefits in frame rate, especially on lower speed systems. In fact, for the perfect sim, frame rate is probably the most important criteria in the area of graphics.

Once you've flown a sim, with a flight model (one that meets the threshold described above of course) that delivers 25+ frames per second you will never want to fly anything else again. The improved sense of fluidity and control over more choppy presentations is extraordinary.

When this was originally written, I was under the mistaken impression that 25FPS, no more no less, was all that you ever need. I have since learned a bit more about "motion blur" which is how movies at 24FPS never leaving you wondering about frame rates. Nevertheless, watching various sims this year played at 25FPS, this does seem to be the break even point for fluidity of movement making control intuitive.

For the prefect sim then, high frame rate is a must. That is not to say though that a little eye candy is not going to be in the mix. Recent work on dedicated hardware accelerators shows that graphics like those in Flight Unlimited or EF2000 v2.0 are a realistic expectation. As a result, the frame rate criteria and graphic detail levels exemplified by EF2000 v2.0 must now be considered the threshold for graphics.

The other aspect that still eludes us at present is the true sense of speed that accompanies fast jet flight at low altitude across country. Graphics are key to this sensation since it is visual perception that must provide the clues for sim pilots. Sound can help but there's no substitute for sufficient ground detail to allow you to resolve details that you can relate to scale as you fly by.

Current generation sims mostly offer what amounts to a lunar landscape outside of modeled areas such as citys or airports; there's just nothing to calibrate your sense of scale to out there so how can you "feel" the speed?? This would rank high on the desired list of graphic trompe l'oeil that would grace the perfect sim.

Cockpits, Padlocks and Multiple Displays

Of course the other dimension of graphics and display work is their presentation in the simulation. Here there are a couple of key items that define the threshold: cockpit views and padlocks.

For cockpit views, the Warbirds-style fixed viewing system seems to have more fans than "virtual cockpit" implementations where you can pan around the view field. I suspect that if we had VR glasses that had high resolution and very low latency (and current generation products fall well short in both of these) then tying into a scrolling virtual cockpit makes sense. Until that time, consider the Warbirds-style viewing system the threshold.

The place where scrolling views come into their own with today's display technology is in padlock views. The entry level for this criterion is a padlock that can visually track a target locked by the radar. The threshold goes beyond this though and requires that padlock locks can track targets in visual range, and this should include enemy, wingmen and inbound missiles as well. The padlock must also be independent of radar locks.

Depending on your personal tastes, the padlocks in Su-27 and EF2000 represent the state of the art today. SU-27 is touted as more realistic, given limits on field of view, and considerable mechanical intervention required to switch between padlock targets). Arguably EF2000 provides more unrealistic tools that at the same time provide virtual pilots easier tools to manage situational awareness (SA) in a furball.

For the perfect sim, some combination of properties is the likely winner: the ability of EF2000's padlock to lock and switch visual targets instantly; EF2000's ability to track wingmen and missiles as well as bogeys; SU-27's more realistic fields of view; helmet-mounted display style overlay from EF2000 or SU-27. Without some padlock of this flavour, the 2D nature of computer displays will continue to limit your ability to keep SA high.

And while we're on the subject, there's another axis for graphic improvement that has yet to be much explored in sims. Back to Bagdad gave the lucky few a taste of what it's like to use a sim with more than one monitor. Broken down old mono monitors and Hercules cards were pretty scarce there for a while.

For the perfect sim, new tools are on the horizon: win95 will support more than one display in the next "Memphis" release. Besides that, many 3DFX owners are wondering what their state-of-the-art SVGA cards are doing while all that eye candy is getting drawn by the VooDoo boards. Additional hardware standards are also in the wings for this feature.

Click to continue . . .

For the perfect sim then, multiple display support would be nice to have to allow such things as additional views on hand or MFDs for quick glance access; dare one say do-it-yourself virtual cockpits??

Win98 is a reality now. Multiple monitor support should be a relatively easy matter from a programming point of view. Still we have no fast jet sims that take advantage, and apparently none in prospect. Truly a shame. As faster Pentium II processor systems become more prevalent, perhaps some brave developer will seize the day and devote a few CPU cycles to putting MFDs on a second screen.

As for looks, DiD's ADF and now TAW with the new improved clouds are the current trendsetters. But recent releases like Jane's IAF and iF18 demonstrate that the pack has a way to go yet before the playing field is level in this department.

Avionics:

Several sims have set the threshold for "hardcore" simulation of sensors, mostly meaning radar and threat displays. Falcon3's venerable hi-fi A-A radar mode started it, Su-27, Back to Bagdad and Hornet3.0 have brought it more up to date and EF2000 added a touch of the AWACS/JSTARS magic to the mix.

In short then, radar support for separate target search, track while scan and short range ACM type modes must now be a given. Ground mapping radar to track both fixed and moving targets are not unreasonable expectations. A reasonable facsimile of synthetic aperture radar should also be within reach of current sim technology.

Since AWACS and JSTARS (just) were both operational and heavily used in Desert Storm, these too are reasonable requests for the avionics threshold; combining these into JTIDS display may be a bit of reach for accuracy given that some of the requested airframes don't have this in the field yet but anyone who has spent time with EF2000 will tell you that JTIDS is a superb compensator for the natural limitations inherent in sims on perception of the virtual war around you.

As for threat warning systems, these are probably less of concern since most sims seem to do these quite well. Perhaps more than elsewhere, these present an opportunity for the developer to tune the tools presented in the cockpit to help mitigate simulation limitations. Common touches such as detecting and tracking incoming rear hemisphere IR missiles are hardly realistic but are welcome aids to most sim fans, even the hardcore crowd (although the real extremists would like to be able to toggle these "cheats" off, no doubt).

For a stretch to the perfect sim in this category, a full implementation of LANTIRN would be the new prize. Here this would imply both the navigation and targeting pods including HUD FLIR display and the ability to scan for and designate targets on FLIR in MFDs. For bonus points hands-off terrain following flight autopilot capability would be a welcome addition, especially if combined with the sense of speed enhancements previously listed. Throw in simulated digital moving map capability and we're well on the way to perfection.

DiD's modeling of EMCON levels in ADF and TAW are a departure this year; one that caused many complaints about bugs that were more failures of pilots to RTFM closely enough perhaps. But their video-game radar and threat displays had both believability and usability problems that marred the effort.

GSC's Hornet Korea continued an excellent heritage and the avionics systems in the E model Hornet offering from Imagic, iF18, also shone. But once more the prize of the year must go to Jane's F-15. Surprising fidelity that would surely have been impossible as recently as 5 years ago gave us operable patch maps and a very serviceable LANTIRN implementation. Much more fidelity in this department and we would all need security clearances!

Displays are also an area where sims are already exceeding the threshold required in many cases. One area of diminishing marginal return that is already starting to go perhaps too far is support for fully active, mouse clickable cockpit displays. These are both overkill and barely realistic.

It would make much more sense if high resolution touch screen capability were available perhaps so that sim pilots could reach out and plug in new waypoints on that up front controller panel but doing it with a mouse seems awkward and positively interferes with suspension of disbelief.

Apart from which, most sim fans in the hardcore camp already have highly programmable controllers that provide far better HOTAS-based control capability than is to be had driving a mouse around the screen. Mouse control for driving the "captain's bars" cursor around an MFD seems reasonble. Making the gear handle clickable on the cockpit art seems like an overkill. Developers: please save your effort for something else on the shopping list!

To each their own. The trend this year seems to be for more clickable cockpits than ever. I've tried to use this in action but thus far I always gravitate back to a well programmed HOTAS setup. On this front, another trend towards keyboard remapping tools is most welcome.

Another twist on this is to simply add more hardware. The excellent MasterPilot from Quickshot is well worth a look as an effective programmable device that will add that certain something to your virtual cockpit controls.

And speaking of controllers... It's surprising how many sims are still written with the fundamental assumption that the keyboard is the primary interface. Programmable controllers are highly prevalent among sim buyers and yet sims still come out with keyboard mappings that don't lend themselves to the controller technology.

For example: use of the so-called "grey" keys for necessary functions (chaff and flares come to mind); keys for range and antenna control that are not easily mapped to knobs provided on controllers for this very purpose; afterburner on digital keystrokes instead of analog joystick axes; the list goes on... The perfect sim could be a lot smarter about the way it supports the current crop of programmable controllers with its keyboard mappings.

Communications is another knotty problem where there's still plenty of room for innovations. Several sims have offered interesting ideas here.

Fleet Defender allowed rudimentary target sorting and stores reporting interactions with wingman that are still very rare. Sorting especially is fundamental to operational A-A doctrine so it's puzzling why this so often gets left out of sims that supposedly offer AI wingmen. [footnote: "attack my target" doesn't cut the mustard since it usually requires you to lock before you sort by handing off a target to a wingie; real life doctrine calls for sorts in search mode rather than when tracking is already established]

Falcon3 and USNF gave us wingmen worthy of Top Gun (the movie): atmospheric if not exactly regulation. But both also supplied the most common maneuver commands that you'd want a wingman to execute, brackets and drags etc.

EF2000 introduced interesting chatter from the tower and AWACS in addition to rather more regulation wingmen, albeit in sterile recordings. EF2000 also offered an interesting command menu system for talking to the rest of the world; useful for A-A but almost lost in the afterthought for A-G usage.

Hornet3.0 gave us an excellent gritty feel to radio comms transmissions together with yet more interactions with air traffic control, both of which are welcome.

Combine a little of each of the above to get the threshold for communications. This too is probably a little ahead of the state of the art for today's offerings but if you can't arrange a BVR sort with an AI wingman before anyone locks and launches, then poor communications had better be offset with good multiplayer support so you can have natural intelligence wingmen instead!

This has been a banner year for communications. ADF gave us an AWACS with a reported vocabulary of 10,000 words. Jane's F-15 raised the ante with very flexible wingman and flight commands as well as AWACS, tanker and airport interactions similar to those in DiD's most recent sims. In this area too, I think we are pretty well served now. To raise the bar we will likely need to look at actual voice comms for human wingman to use in multiplayer games. Battlefield Commander and similar efforts are third party add-ons for this capability. Making such a capabillity work seamlessly with our perfect jet's avionics fit though would be the prize.

Go to Part III