blingthinger

Just me? No it's very clearly stated in the interview. I assumed everyone sifting through this thread would have also read the article in the OP link. My bad. "Moving into payware is obviously a big step. Was that a difficult decision for you? ... we are not exactly “selling” the product. We are simply asking for a small download fee to cover our costs and to help keep the other downloads free. It is more or less a form of donationware...."

In the reference I quoted you were very clear that it caused a slight issue for YOU. There was no reference to other potential buyers. So yes, that is exactly what you said. You need not fear for other customers' reactions, though I'm sure he's grateful for your good intentions and concerns. He's not offering this as a product with the intent to earn revenue. He's merely covering server costs with the fees. He very literally does not care if people buy it or not. Hahaha yeah that word doesn't mean what you clearly think it means, but hey, some folks proactively chose to be insulted regardless of the interaction....

"Issue" meaning that you absolutely will not be buying the Pro version? As in... your eyes are generally unable to de-focus from the blurry ortho that almost nobody actually ever sees (unless they run on 10 yr old hardware or laptops) and re-focus on the autogen buildings and their textures (the actual product in question)? You sat there staring, screaming a little bit inside your own mind? I too noticed the blurry textures. I thought, "at that altitude the XP default textures probably look better", but I can't say that it triggered any "issue"-like response. Certainly not to the level of balking at the actual product.

There's no intent to actively sell or market this. He's not even proactively looking for kindness tips. Just looking to cover server costs. He doesn't care one bit what the terrain textures look like in any "promo" video or if anyone doesn't download it. A true hobbyist.

You must have another layer injecting it somewhere. Deleting 1400 and 1401 removed all the airport vehicles on my install.

If sigma is the primary knob it's turning to tweak Dryden's equations, it's going to remain pretty close to what IRL data sometimes looks like. Spectrum plots in Fig 7 look pretty darn similar to the polynomial equations that Dryden came up with. https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022JD037491

And what happens after that? If that statement was all there is to it, there would be no need for iterating on feel for each acft. Pending clearer descriptions, you're saying that 8.3% of apples are oranges. Not a matter of "good enough". It's simply not using Dryden. Maybe starting there to some degree but the model itself appears to get smeared away based on what you're describing. It's a good starting point to "feel" better, but "feel" isn't "apply" as you've been claiming. LR wouldn't need to implement it internally for manual weather mode. I mean...that's exactly why the MIL document is bringing it up: to set a standard for flight simulators. The algorithm would have to calculate more accurate forces using the turbulent velocities and some basic geometry definition (fuselage areas, force coefficients, much like XP already calculates wing forces) but it's a complete model as is.

Very nice. Silence on the other question though...there's some scaling or morphing of the spectra curves happening somewhere. Question is if it's a linear shift or a scale factor and if it's uniform to the full spectrum. The reason I asked about L and sigma is because that's what's setting the spectrum shape at a given altitude. Depending on how it's being scaled or shifted, you might still be able to claim that it's using the Dryden model. Your silence is worrying there. It might be starting with Dryden sure, but if it's then morphing the model beyond recognition simply because you ran an iteration on a plane and say 'this doesn't feel right', and it wrings the curves out of shape, it's suddenly no longer Dryden. If you're maintaining the spectrum slopes and therefore energy dissipation into the higher wavenumbers/frequencies, your claim holds water. Otherwise, I currently don't see how you can claim that it's using Dryden.

So everything except turbulence eddy velocity delta, which isn't an option because it isn't provided by the sim. That's comparing apples to oranges. The amplitudes will be all manner of mismatched and I couldn't consider that by itself an implementation of Dryden's model. So where is Dryden's model still entering the picture? What values of length scale and intensity are used? Is it using a single value for each? If yes, that would certainly be a pity given that the report offers full up atmospheric distributions for both values. This must be where it's using a transfer function somehow. Had to do my own LLM search on this bit as I've never needed to actively alter a real-time signal IRL.

Behavior of which variables? Which physical datarefs are being compared to Dryden's spectrum curve fits? Per MIL-STD-1797, the model's outputs are the amplitudes of velocity oscillation of a passing turbulent eddy in one of the 3 Cartesian directions: vx (or y or z), at a given frequency in the modeled spectrum for that component. I'm surprised to see you using terms like "filter" or "transfer function" or iterative training of an acft model, given that they are neither found nor implied in the report. Appears that you ran into the same brick wall I did on this one. I couldn't find a way to turn it off either. Same question as above. How what responds? Which datarefs is it monitoring?

What variables is it using to make this match? E.g. local_y? local_vy? Normal force? Please do elaborate on what you've interpreted these specifications to be.

amd's catching up in the raytracing. Slowly. Gradually. Most of the reputation for problems with anything proprietary in the mix is the fact that none of the most popular distros from a decade ago were rolling releases. This meant you were stuck on a specific generation of nvidia driver for a year or more until the new distro release happened. If an app (e.g. XP in my case) demanded a specific version of driver that Debian happened to not offer yet, you had to go the manual route and that sometimes caused compatibility problems. Most folks never had to deal with the real issues nvidia caused. Linus' famous middle-finger-event happened because they were being a pain in kernel dev space. User space never really had to deal with that fallout because Linus worked around it. Hence the middle finger. The reputation propagated because it's opensource world.

Yes it continues to be quite impressive to watch and see what these LLMs are capable of with appropriate prompts and interactions.

Booo. I'm not aiming negativity at you and you know it. You "signed off" of the discussion a long time ago without answering any of my algorithm questions which were never worded negatively. You were immediately bizarrely defensive against a perfectly normal query into how you wrote the code. Any reticence to suddenly embrace your reappearance and gleefully ask more questions is 100% your fault, especially since you insist on misunderstanding my points above as negative. That said, is that even necessary at this point? Is there anything still "broken" in your eyes? It sounds like you have it fixed, no? It sounded like the LLM addressed your OP point before you even asked it (which is good because it means you don't trust it 100%). If it's fixed now then all good. I'll possibly peruse your lua after you post it.

Then you misunderstand that my tone is, for the most part, not directed at you. I don't like contention either. However I do even more dislike intentional misinformation and misdirection. Why did you paste such a huge wall of text in your initial responses? Because you felt like someone was calling you out as a fraud? Because you know that this ai topic IS a hot one right now? Maybe some of both? There are entire websites and companies and organizations fully banning ai-generated content. A lotttttt of people are furious about This. Very. Situation. One that you (hopefully accidentally) think is no big deal. So my only tone to you on this topic is: it is kind of a big deal. If you're going to be posting on a public website (hobby or not) in the future, be aware that they way you treated it is going to raise eyebrows. I personally don't give a hoot if you are LLM'ing your code or not. I probed for clarification because these days there is one thing worse than not reading the report: reading it and not understanding it and then making a technical and emotional mess of the model implementation in a public forum. If I know you're LLM'ing most of it, I can phrase things in way that won't accidentally make you look like the south end of a north-bound horse. If you find an error, everyone can chuckle that the ai is a word not allowed sometimes and move on because "haha sometimes they do that glad it was found before it was released to the public". Why didn't you clarify something like this at the start!!? Don't get me wrong. I have questions/doubts about exactly which data and what the LLM is doing with the data, but seriously??....you gave off all manner of tone and reaction that trained technical professionals have... no interest in engaging with.

Their tune has been gradually changing as the LLMs that have made them wealthy are basically all being developed and then run on linux. That's aside from the fact that lately both major vendors (if amd could be considered "major" compared to nvidia) really seem to have their heads on backwards as they write (are they actually writing?) their driver code. Difference there is that the better amd drivers in linux are not written by amd (all praise Gabe), so nobody notices.

So you were protecting emotions at the potential cost of technical accuracy. Actually it's not potential. Because Bits posted his interactions with the LLM (rookie move, yo), I can see places where it's gone off the rails. I won't go further there because clearly feelings matter more than accuracy. I'm in favor of using LLMs to learn; even write code to some degree. What I wanted to clarify was the tone Bits implied at the start which was that he had: researched turbulence found a report read the report understood the report was implementing a model in the report as computer code Turns out that none of that is true aside from the LLM search telling him about the existence of a model. Why did I want to know this? Because it would change how I try to potentially help fix what was wrong, but apparently egos and emotions dominate. And to those of you mumbling right now about "who cares, F.U. it's just a 'free' plugin": zoom out to the rest of the ai discussions happening around the world right now and I suppose just sit back and let the world burn? At least your simulated bounciness will give you a warm/fuzzy while it does.

Not a single person here has said he shouldn't use it. Not I, nor anyone else. I also don't see anyone criticizing any of any of his work. At all. In any forum or discussion thread. What is interesting is that so many of you have somehow managed to conjure those images up in your minds. Is that the conclusion you jump to when someone asks technical questions that you either don't or refuse to try to understand? If that is the case, then wow...what a community you have here. Lose lift on one side: so you're suggesting a roll in the opposite direction, forced by the turbulence?

.

And you aren't. You're the one who suddenly hyped it up with wild fonts in a wall-of-text post and deflections of being "clear". All you had to say was "yes I'm using it again". That's it. For my part, knowing that it's the driving thought process behind your work simply means that I won't bother trying to engage in a technical sense any further because I might as well just ask chatgpt and I'll end up with statistical parity with the answers. That's it.

Not at all. The LLM is doing the vast majority of the "thinking" both here and in the other thread. You're just feeding it your feelings on the topics. You're only clarifying your usage here in this thread because I brought it up. If I remember correctly, it was the same situation in the other thread (queried by someone else). You also don't mention it in your file post over on .org. Curious... I was merely probing for an intelligent conversation about the topic. Wondering if maybe I was wrong about your usage of a LLM. All good. First thing I do if I ask any of them any questions is turn off the fluffy conversational responses. I'm continually stunned that people actually flock to it and respond to it. That side of the tech has turned into a sad commentary on humanity, though I suppose the wound has to be acknowledged before it can get proper treatment...

I'm not that lucky. Maybe some day. But I am one helping make their dreams and careers a reality. Turbulence modeling is up one of my alleys, so to speak. My beef with LLMs is when there's no attribution to the source of the blood/sweat/tears. I mentioned claude because I remembered you talking about it in the haze thread and your description also seemed a bit LLMish. I wanted to ask questions but kind of knew I wouldn't really be speaking to you. I've toyed with this off and on a few times but never found a complete-enough model with atmospheric spectra (plus scaling factor distributions at altitude!) for all 3 axes. I also never found a way to disable turbulence in real weather mode or easily funnel the real weather turbulence intensity data to a plugin. The report says to use von Karman (but then table values are for Dryden, weird). This sounds like a bit like it's mixing in more low frequency content than you'd expect. If I'm understanding the model correctly, I reiterate an earlier question: how many frequency modes is the lua applying to generate the final signal? Or is it applying the continuous spectrum somehow? You're not using real weather mode to test are you?

And to clarify, that's the pdf page count. That MIL document is very interesting. It even includes a discussion of what to simulate when considering the turbulent wake of an aircraft carrier (p. 812). Also, I hadn't seen table LIV before writing: It does have the rotational moments. Very cool.

One thought about that model: it's only providing/simulating the 3 velocity components of a turbulent structure acting directly on a point in space (e.g. fsx table model acting on acft geometric center). Unless you create your own approximation (or send claude hunting for another model of) pitch, roll, and yaw moment, you'll be missing a fair amount of realism. In other words, it'll be very much "on rails" and will miss most of the dynamic sensation of turbulence. Though that'd really mostly be a problem for low flight speeds. An airliner moving along at Mach 0.8 with a lot of momentum is enough to penetrate the turbulent eddies quickly enough that the angular moments really don't have enough time to make their presence known. Most of the time.

Claude didn't give provide a page number in MIL-STD-1797? https://www.abbottaerospace.com/downloads/mil-hdbk-1797-department-of-defense-interface-standard-flying-qualities-of-piloted-aircraft/ Appendix A, p.804. Looks like a nice complete set of data to implement the model. How many frequency modes is claude using from those spectra regressions? Also, did it describe what it means by "effects and texture"? By "wind induced yaw", do you mean a cross wind? In other words, can a cross wind (be it steady or transient gust) null out the forces of a coordinated turn and hold the plane in a banked attitude with no heading change? Depends on whether or not the arriving wind mass (either front or turbulent eddy) induces an opposite yaw moment. If you have a perfectly ideal/straight wave front arriving at the nose first, you'd end up with a brief period of opposite yaw moment. If the air-mass velocity delta holds constant everywhere as it envelopes the fuselage, it would balance out and the turn would eventually continue. If it's a series of turbulent eddies, you'd get the rock-and-roll of..well...turbulence.