Jump to content
Sign in to follow this  
turner112

How much does increased resolution mesh affect FPS?

Recommended Posts

Hi Andrew,generally speaking, the higher the mesh resolution (i.e., LOD), the more polygons the display engine needs to draw (also depending on your Terrain_Max_Vertex_Level setting, which should be at 20 to see the full LOD10 detail but can remain at the default 19 if desired).However, FS displays mesh in a series of "rings" around the user's position, with higher res mesh drawn to much smaller distances. Steve Greenwood estimated mesh display radius as:19.2m mesh (LOD11) - 12 miles (apparently not fully utilized in FS9)38.2m mesh (LOD10) - 27 miles76.4m mesh (LOD09) - 45 milesThus, the number of mesh polygons doesn't really increase that much and the impact on frames is relatively minor. I'd say you might lose about 1-2 fps when stepping up from LOD9 to LOD10 mesh (with TMVL=20), i.e., much less of an impact than most other types of add-on or realistic weather/clouds.Cheers, Holger

Share this post


Link to post
Share on other sites
Guest allcott

Rather than start a new thread, perhaps I could ask a question here, related to this one?What causes the greates fps hit - Very high detail mesh, but with Max Vertex Level set lower, or mid-detail mesh, with the MVL set higher?It seems logical that in either event serious arithmetic calculations are involved by the sim engine, but is there any insight into the relative frame rate hits of, say, using 76.4m mesh at MVL 20 versus using 38.2m at 19. And is there much of a difference visually?I'm contemplating upgrading mesh for my FS installation, but already have a freeware mesh at 76.4m for most areas I want covered, but I'd like to know if gongto 38.2m would be a beneficial step for the British Isles?Allcott

Share this post


Link to post
Share on other sites
Guest allcott

Rather than `bump` I thought I'd share my experience from a few hours if testing last night. These are not benchmarks, they are just comparative tests on my own setup.I downloaded free 76.4m and 38.2m mesh and set up a flight in the default Mooney from Shoreham, on the South Coast. This is an area I am familiar with in the real world. I set up a flightplan that took me NNW out of Shoreham on the runway heading climbing to 2,500 feet at 1,500 ft/min, followed by a 270 degree right (not 90 left) turn at 2,500 ft to track the South Downs. With 114.75 set on VOR1 and the 240 radial set I tracked along the Downs until I collected the radial then flew down toward Goodwood, before making a 180 turn over the VOR and tracking straight to Ockham VOR (115.30) climbing to 3,500ft. When 12 miles out of Ockham I set up a speedbraked descent to 500 feet, captured the 348 ADF and flew a circuit approach into Fairoaks. This, with a couple of exceptions, is a pretty typical flight for me and I am `comfortable` with the topography without maps.What I am looking for is a similar degree of `comfort` with the FS scenery. At default mesh levels, the South Downs and Shoreham environs look nothing like the real world. Enough said.With the 76.4m mesh at TMVL= 20 the change is noticeable on 2 counts1/ The ground scenery `maps` the mesh more effectively. The nuances of the texture tiles become much more pronounced (I use Ruud Fabers FSCENE 2004 textures)2/ The South Downs are properly represented with their prominent rise from a generally undulating landscape. They are not nearly as prominent, and the landscape not nearly as undulating, with the default meshFPS remained constant at a locked 30 fps throughout (no weather set, max vis set at 40 miles to introduce the haze that is an almost constant feature of vis in SE England)In addition the area around Fairoaks is no longer flat, and the airfield (Gary Summons) merges much better into the surrounding scenery. The hill to the North now looks like a hill, not a result of a farmer digging. No texture `popping` was seen. In particular, the approach to Fairoaks RW24 is sort-of between two glades of trees, properly represented in the Summons scenery. In the default sim, this approach has to be made higher than in the real world as the glades of trees are actually in a small depression below RW alt and they don't really impact on procedures, unless you're tight and late. With the 78m mesh the depression is properly depicted, and the Summons scenery mapped to it, making for a much more realistic view on approach. Next I stuck in the 38.2m mesh. First time out I left the TMVL at 20 and reflew the flightplan identically. No real difference in image quality or ground texturing noted, BUT the refresh rate of the textures was slowed (perhaps by the conversion calculations required to drop 38m res to 76 at TMVL=20). Overall the sim seemed less smooth for no gain in image quality.I then flew the same FP again, with the 38.2m mesh but this time with the TMVL=21. This was surprising as I've never really looked at high detail mesh before but I expected nearly as big a jump bewteen 76 and 38 as between default and 76m. But I didn't like what I saw. FPS varied more widely (to be expected I suppose) but the sim did not seem as smooth, regardless of the numbers. The texture `popping` had returned, indicating approaching limits on system resources and overall, unless I picked out specific ground features I couldn't see a difference. The approach into Fairoaks seemed identical. On the face of it the terrain elevations might have slightly more detail, but you have to really look to see it. But visually, it seemed to me the mesh overrules the textures. Not a glitch or an error, simply the fact that the presentation of the detail in the landscape is actually MORE detailed than what a real pilot sees. In real life, minor undulations in the ground terrain are practically invisible above 2-300 ft AGL and it's the `sticky up` things like forests, masts and tall buildings that you notice primarily. This was the case with 76.4m mesh but is most cerainly NOT that way with the detailed mesh. Strange when I could see little or no difference in the mesh itself.As a last experiment I tried both meshes with the mesh detail slider slid right down to 30 percent (this would seem a fair comparison of how both meshes scale with sim settings, although I appreciate 30 percent of 38.2m is not the same scaling as 30 percent of 76.2m).Again the 76 metre mesh surprised by being almost as good at those settings (TMVL back at 20) but the 38m introduced too many `sharp edges` at that setting (TMVL 21). I don't know why this should be as scaling from higher detail should give better results than scaling from lower.So, unless I've done something wrong, I see no point in spending money on detailed 38.2m mesh unless you're a topography student. Regardless of what FS is technically capable of producing, the detail in ground textures, the IQ, the fps and the virtual representation of what a real pilot sees is best served by 76.2m mesh.To check that I wasn't being confused by mesh claiming to be 38.2m but actually not being, I also downloaded another 38.2m mesh (freeware) and saw exactly the same results. FS 2004 is really optimal at 76.2mDid I miss something? A sim setting change maybe? I can see how the LOD10 might benefit in really mountainous areas but for the average simmer?Allcott

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
×
×
  • Create New...