X-Plane: The State of Visual Accuracy
January 26, 2018
In many ways, being detailed is an essential component to the modern life. We've succumbed to the pressure of raising our standards at work, producing material and deliver better results; the core skill you need for that is to be detailed. As you grow, you'll learn to notice that it's the little things that matter, and if you happen to overlook those little things- it's enough to significantly affect the result of your work. When you learn how to drive for instance - you'll develop your skills to be mindfully aware and prepared in case something bad could happen while you drive; you have to look our for pedestrians, small animals or objects on the road and maintain your speed to comply with the limits. You'll develop your sensory skills as well: in particular, your vision, and you'll learn to adapt to the traffic conditions with those skills, being visually detailed is a must.
In the Flight Simulation add-on development community, the process of getting your work to your customers is a rewarding yet painfully mundane proces, which involves several components, aircraft for instance requires you to collect relevant information and data to construct your visual model, flight model and of course the systems, you then do the aforementioned components like the systems, each with its own frustrating setbacks. In regards to 3d modelling, being visually detailed is key- particularly in a time where customers are increasingly aware of the aircraft's shape and form.
I'm a visuals person, possibly a perfectionist even, where my unrealistic expectations prevent me from sharing my work to others for their enjoyment; because it's not good enough I thought. Fortunately this hobby of mine involves recreating airports casually, and I do not take it as a full time permanent job oppurtunity. If something is wrong, that's enough of a reason to go back to the drawing board. This is a recurring problem for me and for my teammates, and it has gotten bad enough that one decided to start their own brand and make their own incredible scenery. But besides me needing to cope with the reality that you can't replicate every crook and cranny, I felt that some members of the community and the developers couldn't be two polar opposites- which is definitely fine, but it gets to the point where you see how lax their standards are when it comes to visual accuracy.
If you do it, do it right.
When an aircraft is released to everyone's enjoyment, everyone would eagerly purchase and fly it to their heart's desire, thats the typical consumer behaviour, indifferent to the details that characterises the aircraft. For me however, I'm never really interested in seeing how the aircraft's brain will prevent me from slamming onto the ground, I'd install it and rip that exterior model to pieces to see whether it respects each and every curve of the real world counterpart. To me, a perfect add-on takes into consideration of both design and visuals, while acknowledging the intricate or sophisticated systems on the aircraft.
Visual accuracy of aircraft in X-Plane has significantly increased thankfully, and developers do have an increased awareness on how the aircraft looks in real life. But still, it's 2018 and we as a collective of developers are yet to meet or surpass the likes of other simulators like P3D, DCS, and FSW- visual accuracy wise. Of course, we have the amazing artists from Rotate who combines their magical inking on the textures and modelling with good system simulation, or the IXEG 737, which is by far, both personally and constructively- the best balance between visuals and systems. There's not enough of those developers however, and we need to realise that its time to champion visual accuracy.
With X-Plane 11 and its amazing PBR capabilities, theres no excuse as to why you can't put effort into getting the form correct (note, I'm speaking about form, not detail.) Sure, if you're a newbie developer- exceptions can be made, but when the developers are practical veteran developers who has years of experience in X-Plane development, it's silly to see some of the most blatant features ignored or modelled incorrectly. One thing that still struck me as odd is the rampant oppression of constructive feedback, which in some cases, even the developer or a representative would also overlook these feedback and believe that it really is " just 0.005 mm off". No Sharon, the door is like 3 meters off, stop drinking.
As a result of these continued dismissal of genuine feedback, the products stays flawed and unfixed, even though user-made modifications shows how easy it is to fix. A prime example of how an aircraft with so much potential, gets neglected is the Flight Factor 777. the team's adamant position to not improve the 777's visuals since its release in 2012 caused the development of the PAWDA 777.
Chronicles of Flight Factor
I've experienced and witnessed plentiful amount of oppression from constructive criticism, and most of the drama stems from FlightFactor. Seeing how Flight Factor (at the time...maybe now still) was very uncooperative and unreceptive to feedback, they weren't very instrumental either. The 777 was definitely groundbreaking in many areas such as the systems. But the modelling was a major, significant letdown in terms of accuracy, and I'm not the only person who noticed the deformities of the aircraft. The team member's previous works however, ironically looked better; previous aircraft developed by the main visual artist (Roman Berezin) excelled in accuracy in contrast to the 777.
Obscenely huge landing gears, colossal passenger windows, weirdly modelled wings with spazzing flex, and of course...those iconic engines that does not resemble a GE-90-110B at all. Perhaps they added a touch of artistic freedom; adding 3 too many blades for the engine fan. A lengthy discussion on the .org full of customers calling out for the wing flex to be fixed is still unacknowledged to this date.
Since nobody is going to fix it, I thought... why not put some of my knowledge to use in fixing the model myself?
A year and a half ago, I began to document and post my findings and fixes on a facebook group called X-Plane Dedicated which was at the time one of the most active groups and is the locale of your daily community drama.Today, it's definitely more calmer now, and the admins seems more reasonable.
The 3d model of the landing gear was imported and was resized to a more appropriate size. The result is *better*, it's still pretty big but it's definitely not fit for off-roading now:
Then, the next step involves moving back to the engines, first by removing three fan blades in the engines and readjusting the position of the remaining, and then touching up the nozzles so the model replicates the form of the real engine more closely. as you can see here:
The series of posts on Facebook garnered plenty of attention, mostly positive because the masses wanted a fix for awhile and the posts finally come to show the ease of doing quick touchups to the 3d model to make it that much better. Some members understandably didn't take it lightly however; thinking that by criticising the aircraft in the group, I'm not showing my edits and feedback to them directly and would tarnish their reputation. if anything, the posts does serves as an indication that there is a sizeable amount of disappointed customers who's still waiting for an official fix for the 777.
I try to maintain a constructive outlook, the feedback and findings that I posted does show that a significant amount of work was needed to bring the 777 to standard. People who viewed the posts enjoys it's objectivity and the discussion it brings to do-it-yourself modding of aircraft. But alas, the posts' existence does not last long.
From this point, I've looked back and checked the list of modelling errors that can or can't be fixed, without significant editing to the mapping too. I thought to myself: "Forget it. It's time for a new 777." Henceforth, marked the beginning of the PAWDA 777's development. I started anew, with a clean sheet, and modelled a new 777 exterior from the ground up, and after a few series of 777 development updates on the facebook group, the community reacted in excitement.
This is where we began to see how the blatant suppression of feedback develops into some petty political game. Hours after the latest series of 777 posts were published on the group, the admins began deleting it citing how the posts "offended" some members in the community and it violated FlightFactor's terms and conditions where the project has stolen "algorithms". To add fuel to fire, the admins followed up with a kick and a ban from the group, "accidentally" they explained.
The 777 isn't only the aircraft FlightFactor has released without jarring visual errors, fast forward for a few more months where we see the arrival of their 767 and 757 and you can see that they're finding alternative routes instead of just doing it properly, that's my take on this reoccurrence. Otherwise, what is the deal behind this repetitive modelling mishap? Using photos to supplement diagrams would be a helpful start instead of solely relying on documents, and using old 3d models without optimising and improving them is not the best route to take, when the marketing proclaims that its a brand new model. I've still hope for them, as there are improvement in their latest products; namely the the A320 Ultimate, which looks much better modelling wise.
The release of the 767, a solid product which received to much applause that even earned a PC Pilot award. It's also a visibly improved product- modelling wise, in particularly the wings and the flaps. But it still has its major flaws, which I instantly saw. The aft belly fairing was modelled wrongly, and looks bulky and unaerodynamic, the slats were modelled and textured strangely, where the inboard slats has a mono-metallic texture. Windows were also cut too low amongst other things.
FlightFactor's latest Boeing release, the 757 V2, looks to be a very solid product release. Until you realise that the same modelling errors from the predecessor are still present in the new version. Wrongly modelled/animated flaps and slats, windows were cut too low by a 2-3 centimetres (sounds small right? but it becomes a problem when you repaint the aircraft), side cockpit windows aren't curved like the real deal and so on. I opened up the model from v2, and v1 and compared it side by side in AC3D.
Apart from a few shuffles for how the 3d model is exported and divided, they are pretty much exactly the same. Only the texturing is redone (and looks decent) Have a look:
Recycling 3d models is perfectly fine, only when you optimise it and retouched key areas where you did wrong last time, and in the case of the 757V2-that doesn't seem to be the case. Several people complained about the flap animation, which was extending at odd angles with a huge gap when the second slotted fowler is extended, what they did to fix was akin to sweeping the problem under the carpet; simply by reducing the flap extension angle to make up for the huge gap.
I definitely don't want to single out FlightFactor for this, the team is full of incredibly talented developers that is capable of producing head turning products that can rival even the mammoths in the community such as PMDG. However, it seems that with such a high profile standing in the community, they do have a track record of poor visual accuracy, and I think that given their incredible talent, a little more effort on this goes a long way: both for them and us users. I remain hopeful that their upcoming products such as their 777v2 would be a completely reimagined product, with no traces of the old modelling and is the pivotal product that sets new standards- visually and system wise. Other developers has slipped a few modelling flaws, but they are more receptive of feedback and aren't as afraid to fix it, yes it takes quite a bit of work to make it right, but with the right mindset and passion to improve things- I think the end results would be more rewarding.
The main issue I am seeing right now is when developers create and release aircraft with really jarring flaws. It's really not hard to go the extra step to make your work look perfect. There is absolutely no excuse to attempt a perfect replica of the real world counterpart, especially when you have years of development experience and plenty of resource material at your disposal. Why turn a blind eye on visual accuracy? If you model the exterior, do it properly and with a hint of passion, you can turn that mess into the best. Other developers do overlook minor details, but it just seems to me that FlightFactor takes the cake every time when it comes to consistent modelling fails. And the worst thing yet is that they recycle old 3d models...with huge mistakes and they dont fix it but remapped and retextured it.
Even Hiroshi Igami of POSKY fame in Microsoft Flight Simulator thought about the quality of visual accuracy at the time:
It's not in everyone's minds to expect a perfectly modelled and textured aircraft exterior. Usually when such complaints about the exterior was voiced, people come to oppress it with generic arguments such as "I fly in the cockpit, so the exterior means nothing to me". I used to foolishly slap them with something along the lines of "then if the developer models a potato for the exterior, would you still buy it", and BOY did I misjudge their expectations; you guessed it.. most proclaimed their devotion to potato shaped exteriors.
Is this a rant? Possibly, but this comment is a chance for me to express my disappointment and thoughts on our expectations for developers, and the developer's reaction to constructive feedback. We shut down good and genuine suggestions but ask for the most sophisticated system simulation, leaving the visual standard to the developers to decide and control. Many of us don't want planes that look like what the real deal is. It's time to expect higher, for the developers to aim higher.
I will repeat that I'll remain hopeful FlightFactor would deliver incredible products that are solid and provides the best balance between visuals and systems, like the IXEG 737. They have the skills to do so much more, but only if they would put in the extra effort to model the details that makes what the aircraft is. Their a320 looks to be a solid release, even though some members noted that the modelling is outsourced. Visual Accuracy shouldn't be an option, but a necessity to elevate and bring our standards higher, bridging the gap between simulation and reality.
Threshold encourages informed discussion and debate - though this can only happen if all commenters remain civil when voicing their opinions.