nickhod

Members
  • Content Count

    81
  • Donations

    $0.00 
  • Joined

  • Last visited

Community Reputation

125 Excellent

Profile Information

  • Gender
    Male
  • Location
    Cardiff, UK

Flight Sim Profile

  • Commercial Member
    No
  • Online Flight Organization Membership
    none
  • Virtual Airlines
    No

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. nickhod

    Release date of Ms 2020

    April 1st 2020. "None of this was in game footage .. fooled ya!" Let's hope not. I would guess E3 2020 in June, also based on them starting to do marketing about a year prior to that.
  2. There's no reason that a flight simulator shouldn't achieve the performance that we see in AAA FPS titles. In many cases a flight simulator scene will be less polygon and texture heavy than a complex FPS scene. A character model in an FPS might be 50k polys, and you might have 4 or 5 on screen, all with their own physics an AI, then all the highly detailed scenery. For comparison a polygon budget of an entire aircraft (inside and out) might be 200-300k polys. Your GPU only cares about polygon count. You might have a 10 mile draw distance, but you're drawing at a low LOD past a few miles. Photogrammetry actually helps with this as they have various LOD mesh tiles. Volumetric clouds and dynamic lights are the two most GPU intensive features of a flight sim, but in a modern sim they will be almost completely handled by the GPU. The point being, in a modern sim, so much is being handled by the GPU that you get to decide what FPS you want, based on the amount of "prettiness" you want. On P3D / XP, I can't decide that I want 90fps, it's not even realistic. Obviously there's a lot of physics to be calculated per frame but: a) Increasingly some of this can be done on the GPU b) Implemented well, it can be done across many CPU cores. c) If you remove the ability to completely replace the physics engine (and I predict MSFS will), you can heavily optimise it per (a&b) I would assume and hope that MSFS has been written to achieve 90fps (for smooth VR) in heavy load scenarios. This is totally achievable with the hardware we have today. A lot of people are still in the "hey 17fps isn't that bad" mindset with older sims.
  3. Short answer: Yes, the MSFS dev team will have optimised to saturate GPU capability and minimise CPU activity as much as possible. It's what anyone developing a commercial game in 2019 would do Long answer: FSX, P3D and X-Plane have their roots in a time when GPUs weren't especially powerful and CPUs were expected to become incrementally more powerful for the foreseeable. Those sims do use multiple cores (though often weighted towards one), but do too much work on them. As a result they can't drive the GPU hard enough (or smart enough) and frame rate suffers. Many modern games will happily eat up as much GPU resource as is available at higher detail levels, yet can run great on a an i5 processor. In those cases, as much work as possible is being offloaded to the GPU, often including object physics. So, I think people will be surprised at how modest the recommended CPU spec for MSFS will be, and how your CPU will make little difference to frame rate above a certain level.
  4. nickhod

    Channelling to end users

    If you want to know whether your connection will be able to handle streaming orthophoto tiles and photogrammetry buildings, download "Google Earth Pro", navigate to a densely populated area, go to the "Tools" menu and select "Enter Flight Simulator". It's apparent from the screenshots we've seen that it'll be possible to override the Bing satellite imagery with custom imagery, so that'll open up the possibility of an Ortho4XP equivalent tool for offline download. That leaves the question of photogrammetry and autogen. This will no doubt be streamed by default, but maybe MS will allow the bulk download of specifics areas for offline use. If not, there should be ways to create autogen with 3rd party tools. I'd guess that someone using MSFS offline or with a poor connection will still be able to get a lot from the sim with some work beforehand.
  5. nickhod

    Seasons

    If the sim is orthophoto scenery only (which it seems like it might be), seasons are hard to do well. You can manipulate the orthophoto images to replace greens with whites for winter. You can maybe get smart and detect tree areas, and replace those greens with browns for autumn / fall. The end result is pretty hint and miss though. Some stuff gets recoloured that shouldn't be. Some stuff is missed. The output of existing tools that do this isn't great in my opinion. An alternative approach is have snow fall on the landscape implemented by some shaders. This works nicely close up, say, at an airport, but doesn't scale well flying at altitude with good visibility. There are no great solutions to seasons that I know of, so I'll put my money on there not being seasons in MSFS. (I think we'd have seen a screenshot showing them off by now if there were).
  6. nickhod

    Aug. 29th : the wing tip pic

    The wing-tip pic (full size: https://cdn.microsoftstudios.com/fsi/wp-content/uploads/2019/08/Wing-Tip-View.jpg) is definitely autogen and not photogrammetry. To the left of the wing tip itself, you can see a huge farm building. It seems like the autogen system didn't know what to place there (or the OSM data wasn't good enough) and it's put a generic grey box type building with a repeating roof texture. Interestingly the trees do look like photogrammetry trees (kind of 'bubbly' and full). However it makes total sense to build a large library of tree meshes from real trees in photogrammetry areas. If this is representative, it really is next-gen autogen. Fantastic to see that Northern Europe will look great despite the lack of photogrammetry.
  7. nickhod

    29th update ?

    Dear Microsoft 😀 Should you happen to be reading this, which I very much doubt, please post the Thursday updates at 9 - 11am your time, rather than last thing. Currently, those of us over the other side of the pond have to either stay up very late or wait until Friday morning to read them.
  8. nickhod

    Your First Next Gen Flight

    For nostalgia value I'd depart from a Chicago airport in the C172 (ideally Meigs, but I doubt that'll be included by default), and do a sightseeing flight around Chicago. I'll be thinking of when I did the same after getting a copy of FS95 as a birthday present and just how far simulator graphics have come on.
  9. Being 38, I'm overjoyed at being called young. 😀 I started my own software company at 28, which now employs people. We develop 3D industrial simulation software and have contracts in the US and Europe. (What do I know about making money from simulation software or graphics programming though, right?). Believe me, if I'd have listed to the naysayers (of which there were many), I'd never have even tried. "Stay hungry, stay foolish" as Steve Jobs said. So, we'll see.
  10. nickhod

    How should the ATC work?

    Aside from the aviation aspects, great speech synthesis and speech recognition. In VR I don't want to have to type or grab controllers to click buttons.
  11. I fear that you might be right. 😥 It's a shame as I love VR in Aerofly. All depends how well VR is done in MSFS, if at all.
  12. Yes, that's exactly what I am doing. Obviously I'm developing for AFS2 right now and following their standards of how things should be done. Porting is always a challenge, but porting is usually less work than starting from scratch. My point is that it is possible to think, even at this early point, "might this project be a good fit to potentially port to MSFS". The Cessna 208 seemed like it wouldn't be, for obvious reasons, and so I abandoned that project while I hadn't sunk too much time into it.
  13. I did say "thinking" about what could be developed, to be fair... I started developing a C208 for Aerofly (my current go-to sim), but decided to shelve that as there's one included with MSFS. So, my intention is to pick another aircraft that I suspect wont be in MSFS (pure guesswork), develop for Aerofly now and be ready to port to MSFS. I also have a European airport project near completion for Aerofly. It's not in a photogrammetry area, so I believe it's a good contender for a MSFS product. (Again, pure guesswork, who knows). If all the 3D and texture work is done in Max, I'll certainly be in better place than having done no work at all, waiting for MS to release an SDK. So yeah, I'm absolutely serious.
  14. nickhod

    Photogrammetry

    What I find amazing is that there is software that can construct the photogrammetry mesh with just machine vision. I saw a video of someone taking screenshots of a Google Earth photogrammetry building from different angles, feeding back into a photogrammetry app and it reconstructed the model pretty well. I'm sure MS and Google throw every resource at the task and also use LiDAR, as you say. Will someone find a way to decode Google's photogrammetry mesh tiles and bring them into MSFS? Long shot, but I wouldn't rule it out given enough time.
  15. Maybe so. Even then it suggests that there will be some kind of central place to buy 3rd party content. It was an interesting job ad that I thought I'd share. Open to interpretation, I agree.