My priority is simply how aircraft fly. I love the visuals, I get there may be some issues. I love the environment, again I get there may be some issues. MSFS has given us so much that we have been wanting in all areas other than that in which I specialise. The flight model is exclusively what I deal with as a developer, so ‘emotional dismissal’ is not what is intended. Yes, there is emotion involved because I see so many improvements in other areas and I really want the flight model to be improving as well.
So emotional frustration certainly exists, but I have been developing in this field across multiple platforms for 15 years or so and been simming since I stopped real-world flying in the late 90s. It’s never been perfect, but my personal opinion based on the background I have described is that the other games mentioned could look at their flight models and be technically argued to be more simulator-like.
‘Certain aspects’ of CFD needing work - I don’t think that’s actually the case. Certain aspects of the definition of the aircraft is the point I am trying to make. They have stated that accurate geometry is necessary, but neither give us the ability to define accurately enough nor the ability to adjust when the geometry we can define gives poor results. Looking at the CFD lines in the flight model, there are very few additional to what went before and it appears that ‘pretty coloured lines around the aircraft’ is all that is needed. With accurately defined geometry, I still do not find that the aircraft I am working on behave correctly - the experience of ‘total flight’ to me is not achieved unless the aircraft behaves correctly across the envelope.
I think both the CFD and thermals viz are great. It’s AR for VR/Digital Twins.
From an infovis perspective, I’d like some improvment to the information communicated, such as speeds and temp. This is important info to get more insight from the tool.
Temp could mapped to line color (red/hot to blue/cold spectrum), and speed mapped to line thickness (fat/Slow, thin/fast). It could also be useful to have a performance scaling slider, so you could dial the density of the line display up and down, as well as toggle off various information display options such as outlined above.
This is true, but they could go a long way to alleviate this by providing better SDK tools.
Automating placement of markers etc should be possible by just linking to the high level objects (eg: wheel, tail etc) and let the system iterate and place any needed marker objects to link the model to the simulation. Or at least place them in the best location to suit the “smoke and mirrors” currently under the hood.
Saves dev time, more accurate than hand placement, and results in better model definitions for input to CFD. Make it so, Asobo!
Same here. I’ve just started looking into the debug weather section of dev mode. Just did a flight in the analog King Air (not CFD-enabled AFAIK). There’s some very noticeable up- and downdrafts in dense cumulus clouds.
I found on badbadweather.com :: Thunderstorm that some of the Greek islands had thunderstorms. On approach to LGRP Diagoras, Rodos I flew into some very heavy clouds. As per the approach I was to descend from 3000ft to 2500ft. The autopilot was not able to do this. I checked in dev mode => options => debug weather and enabled “proximity airflow visualization”.
In the image below, moderate vertical wind turbulence was reported - basically updrafts of 1475fpm.
I also experienced significant downdrafts in these clouds. Although the plane didn’t get shaken around, handling the aircraft especially in the updrafts was challenging. Notice also that turbulence G variation was reported as light.
I can confirm these observations! The dynamics of the air feels quite real.
Now, if they could only tune down the shaking of the horizon. Eyes are really good at keeping the horizon steady when the head is moving, in real world. In the simulator, no matter how much the camera is moving due to turbulence or any other non-commanded reason, the view should remain focused at infinity.
Cool statistics didn’t know we can look how much turbulence it is
I checked that we can see that air density changes in the sim depending on how much temp it is. It’s so much stuff simulated in the background that we don’t know about.
Like dew point is actually simulated even that ATIS is reporting 10 degrees all the time.
After some testing I can confirm that clouds create rising air below them (proximity airflow) if they’re a certain density and size.
Here’s a test over a desert. This is a manual weather preset with density set to the max and slew mode enabled. As you can see, the air rises (green\teal lines) when you move into the shadow of the cloud and falls (purple lines) outside of the shadow. This seems to be pretty consistent. I also tested over the ocean with same results
Here’s me slewing sideways into the shadow of a cloud:
Hello. Has this type of 20 KM CFD and airflow in MSFS in SU 11 been done in other desktop flight simulators before? Which other desktop flight simulators have done this and roughly how does it compare to MSFS?
To my knowledge this is a first in flightsim ever. There are no consumer flight sims doing this - I am even not sure if there is a commercial one which does this …
I think this is really important. And I’m not meaning to dismiss all that has been mentioned about the CFD especially wind/thermal etc. Much is really impressive.
But an accurate physics representation of the aircraft is really important. Otherwise, no matter how accurate the visual model looks, underneath (and I am surmising here given i have no practical knowledge of sim aircraft modelling, but basing it on a couple of Dev posts i have seen) you could be flying any number of similar planes. Unable to allow the cfd to detect and manipulate anything other than a perfect cylinder, or for a wing - a tight restricted idea of a wing is not ideal.
That said given enough impetus, enough will on Asobo’s part I expect this will develop - but how do we signal we want this? Do we need to “push” the point or is the CFD absolutely naturally going to become more complex and allow for non-standard shapes?
I wonder if a Wishlist Poll could be created - ie Can the CFD model be developed to encompass complex shapes or swept or angled shapes? I don’t think I’m the right person, but someone with knowledge of developing aircraft on the Sim?
Oh, I see. If this is the case, then I think after SU 11 is released, MSFS’s flight model may start to surpass XP 12’s flight model for single wing planes.
Granted, XP’s flight model and MSFS’s flight model have different features, but there are so many new features in MSFS’s flight model, that I think MSFS has the more complete and comprehensive flight model (at least for normal single wing planes - my understanding is that MSFS has issues with double wings, weird shaped wings, etc).
If “more complete and comprehensive flight model” is what you got out of the earlier poster’s message, I just don’t know what to say…SMH.
It has potential, but the only “more” it is…is simply more limited and more constrained.
As I sit here thinking, there’s only a few addon aircraft that match the stated limitations.
Like maybe the DC-6, but none of the modern airliners.
Most light GA aircraft have rather boxy & irregular fuselages rather than cylindrical (C172, Kodiak, Caravan, et al)
etc., etc.
From a pilot’s perspective, I find the 172 G1000 with the CFD to be entirely unrealistic, whereas other planes without the CFD (such as the JF Warrior) are much better. However, I feel like Asobo have proven themselves at this point to be more than capable of making progress with the simulator in all regards, even if that progress is not necessarily linear. In other words, I’m willing to give them the benefit of the doubt that at some point down the line the issues we see now are going to be much improved.
didn’t know we can look how much turbulence it is 
