Speedbird193

So true. The amount of Vibe/AI coded stuff people push out these days is crazy. Use at your own risk.

This is actually quite entertaining to follow. What’s next - top security clearance and background checks before I’m allowed to make a purchase? They have well and truly lost the plot (and presumably a substantial chunk of potential sales…).

Hard pass from me. I'm not running DRM for sceneries - there's already enough background stuff running in the sim that actually has a purpose. I have no problem with DRM on aircraft like the Fenix or for other things that I use on every flight. Not for a scenery...and where does it end? If I have 50 addon airports installed from 25 different developers - do they want me to run 25 different DRMs in the background?? Their qualifications are also laughable. I hope the community reject this to the point that it sets a precedent for others to steer clear of. I don't know what they're even trying to achieve. People who weren't going to buy the scenery won't be buying it just because they put a DRM in. The only thing I can see this achieving is upsetting actual paying customers who were going to buy this but now won't because they don't want some AI coded DRM running for a scenery they legitimately bought. Really makes you scratch your head as to who comes up with this stuff...

Looks far better and more in depth than Qualitywings. They were good but also quite basic in what was simulated.

Great addition, though I usually fly in the evening and pick a real flight to simulate. That means I might fly at 8pm real time and 10am sim time. So unfortunately it won’t be much use for me. If they could somehow include a 24-hour history from FR24 that would change things completely and be amazing!

I just got my new PC with the 9950X3D. Core parking wasn’t working when I first ran the sim but it’s been flawless since then and no issues at all. Correctly parks the cores as intended each time - must be something their end.

Tomshardware posted a benchmark of the 9950X3D2 in MSFS24. It puts it on par with the 9850X3D. Personally, I don’t trust their benchmarks as the 9950X3D performs 20% below the 9800X3D and they cite core assignment issues as the main reason for the poor performance of the 9950X3D in their benchmarks: “Although the Ryzen 9 9950X3D2 doesn’t claim any lead over the Ryzen 9 9950X3D in most games, it leads by 23.8% in Flight Simulator 2024, keeping up with the Ryzen 7 models. The base Ryzen 9 9950X3D experienced core parking issues in this game, even after we triple-checked that the correct packages were installed. It was utilizing the non-X3D CCD, even on a fresh Windows image, which helps explain the big uplift we’re seeing here.” I’ve spent a lot of time reading through user reviews in the last couple of weeks and just ordered a new system with the 9950X3D as a result (coming from the 13900KF). Virtually nobody else seems to have those core assignment issues with it and general consensus seems to be that the 9950X3D should achieve very similar FPS to the 9800X3D/9850X3D and offer a slight edge in smoothness in demanding situations. So in reality, I would expect the 9950X3D to sit right alongside the Top 3, meaning the 9950X3D2 should perform the same, thus not offering any tangible upside over the X3D in MSFS24. For a minute there I was worried I should have gone with the X3D2 instead, but reading why they show such a big gap in performance put my mind at ease somewhat. Will be interesting to see some hands on experiences with the X3D2 in MSFS24 over the coming weeks though.

Fully agree! It’s a shame a few irresponsible loudmouths ruin it for the rest of us. I don’t want the hype we get from other devs and actually quite like the Fenix way of just dropping quality without all the noise, but their last development update was almost half a year ago. Some kind of news on where things stand would be very welcome. Either way - can’t wait for the VNAV update and the NEO when they do drop.

I hope this pans out. I was really disappointed to see Fenix almost made a 350 but stopped when ini announced theirs. It just doesn’t live up to the hype they created around it. If this new one is really high quality I’m interested. Oh, and please, please give us the option to use sim rate above 4x! I don’t get why PMDG have been able offer 16x working perfectly for decades and yet ini claim “it’s not stable above 4x” when I can use the default sim shortcut to fly at 8x and 16x no issue, other than fuel burn being off. I fly real routes but just don’t have the time to spend 3 hours on a long haul in the evening when crossing mostly ocean.

I’ve never had any WASM crashes apart from the early days of the ini A350. Never even seen those orange screens since they were introduced a while ago - and now had 2 WASM crashes in the last week. One was the PMDG 738 when loading a specific flight plan from simbrief (3 times in a row) and yesterday my ini A300F had the orange screens as I was changing the arrival runway into EDDP. Not sure if that’s just bad luck or something changed…

Can confirm, I also had a repeatable WASM crash the other day loading a Simbrief plan on the 738 via FMC. Only happened with one route/flt plan and I tried 3 times and had a WASM crash 3 times at the same moment. Then decided to fly another route and it all loaded correctly again.

I think there's some confusion in this thread. Photogrammetry is already using AI in MSFS (blackshark.ai). As far as I know, they are using a range of AI and machine learning tools in the background of the engine to clean up the source data they have (Bing etc.) and I was always under the impression that this extends into ground textures, where AI is used to clean up images, colour match and remove artifacts as good as possible. Here's an interview that goes into some of the detail ahead of the launch of MSFS24: How Microsoft Flight Simulator 2024 uses machine learning AI, and how much of your data it might need | Rock Paper Shotgun

The PMDG 737 is set to 110.5 by default. I’ve had this once or twice and switching to the stby ILS frequency and then back to 110.5 usually fixed it. Might be more of a PMDG bug with some ILS approaches that use 110.5?

I get horrible performance degradation with the CRJ. It becomes almost unflyable after a while. Don't have this with any other aircraft.

Can anyone share what level of “time compression” they have implemented? 2x? 4x? 8x? Thanks!

I’ve always had departing and parked at 10, I used to limit arrivals to 5 as the engine pre update couldn’t cope sequencing it and it became a mess. Now arrivals at 10 works great. I didn’t really notice much difference for departing at 10 vs pre-update, but my airports have always been busy. I flew back from RJAA to CYVR last night and the radio was alive with frequent pushbacks and taxi instructions for traffic. I didn’t have a massively long taxi to the departure runway but I had 2 aircraft ahead of me and had to hold behind an ANA 787 when I got to the runway that was waiting for an arrival to land. When I had lined up, a JAL 737 was already at the holding point behind me and there were at least half a dozen or so aircraft taxing around to the runway (it was night, so was easy to spot them with their lights on). Overall, departure and arrival volumes feel quite realistic now. Haven’t flown into EGLL yet, that’ll be the true test!

Did a flight CYVR-RJAA last night and it was great. Of course there’s tiny little things here and there - it’s a simulation at the end of the day. Arrival into Narita was a cool experience - I had all traffic sliders at 10 and it kept everyone separate. Was great hearing so many voices and accents on the radio and people being properly sequenced via speed instructions. I had to fly 190kts for the last 15 minutes as I was on the STAR behind a Hawaiian A330 and there were dozens of arrivals all merging together at the FAF in near perfect harmony. The traffic map looked crazy with aircraft sequenced properly and the spaghettis all coming together from the different STARs and departures going out. Tower was clearing aircraft for take off in between the landing ones on 16R and told them to hurry up due to arriving traffic on final. The Hawaiian landed about 3 miles ahead of me and exited promptly but it was quite an experience not getting my landing clearance until he had fully vacated and I was maybe at 500ft. Big improvement versus before, where I constantly got cleared before even turning final. The new local voice model is great - adds such immersion and the Japanese accent was authentic but actually understandable, unlike the old basic voices. Is it perfect - of course not! But it’s come a long way and the latest update is a great step forward. I hope we’ll get vectors/shortcuts on transitions and holdings soon.

No issue here with the new driver, haven’t noticed any difference.

Never had a single stutter or other performance impact since I started using this.

Have you tried the local voice model? It’s miles ahead of the basic voices and the latest Experimental version adds worldwide accents to the native English voices. And it’s free…

The way I read it, it simply means that the engine will start sequencing AI controlled traffic 100nm out, so they don’t all arrive on top of each other. It then uses speed instructions as required to keep traffic separated as everyone gets closer to the airport. Nothing changes from a user flight perspective other than potentially getting speed instructions as one gets closer to the airport. Update: Done 2 flights now, both a notable improvement in terms of vectoring and sequencing. I flew into CYVR after cranking up arrivals to 10 (which previously caused mayhem) and ATC was keeping everyone apart. I got a few speed restrictions, so did the AI. Was even told to "maintain minimum approach speed" as I was turning final in my B77L. Very nice so far!

Sounds like it’s not miles away (few weeks of stability testing). I’m actually very happy they’re trying to get the core engine into a polished state before adding more features.

Considering the length of the changelog, I felt this deserved its own topic. Key points to highlight: Completely new vectoring logic Completely new sequencing logic kicking in 100nm out, now including dynamic sequencing and speed instructions (can't wait to try arrival traffic level 10 now...) Local Voice model now features accents Better ground and taxi handling, including extended thresholds (hopefully that fixes the broken line up point at the ini EGLL on 09R) Backtracking fixes New logic to avoid a C/L/R stand having aircraft parked in all 3, so it only parks on L/R or C I had the feeling they were cooking up something good, considering how quiet it has been on the update front but this exceeds my expectations. I can't wait to try this update. Looks like the v1.0 release is getting closer and they're teasing a VFR demo video in the near future! For me they're one of the top tier developers out there, great job!

Had stutters to the point I couldn't use the sim introduced by SU3. Completely cured by SU4 and super smooth since.

Oh please, please let this be good! We need new aircraft devs and someone else who can deliver high quality Boeing aircraft. There are so many 787s flying around, it's long overdue in the sim. Will it just cover the -9 or are there plans to do the -8 and -10 as well at a later date?