How much area is too much for low resolution scenery - good flyover - recognize the lakes, roads, small towns, etc.

  • I am thinking I would like to give FSET a try for fairly low resolution scenery so I can cover a fairly large area. Say 100 nm by 60 nm. I am thinking maybe level 9 or level 11 max. Is this too much to bite off in the beginning? Any ideas or suggestions?

    This is the middle of Mississippi and it almost all woods, steams, and small lakes. A few small towns here and there.

    Regards,

    Ray

  • Just to give you a reference here, I tested a Pennsylvania FSET scenery that is 140nm x 162nm at -1 (0.25m) FSET resolution and I processed at levels 10 - 14. The geoconvert processing time was around 2.5 hours and it completed successfully with no errors. The resolution is also amazing in VR.

    I've also run similar tests at various levels and sizes but all at the same -1 FSET resolution. What takes time is dependent on how many levels you add to the TMC.

    IPACS Development Team Member

    I'm just a cook, I don't own the restaurant.
    On behalf of Torsten, Marc, and the rest of the IPACS team, we would all like to thank you for your continued support.

    Regards,

    Jeff

  • Well, it must be way to large an area. Maybe I will try about 1/4 of that as a starter.

    My upper left corner is -90.38 32.60

    and my right lower corner is -89.73 32.23

    What is the next logical step to see if it is even workable?


    Regards,

    Ray

  • Just to give you a reference here, I tested a Pennsylvania FSET scenery that is 140nm x 162nm at -1 (0.25m) FSET resolution and I processed at levels 10 - 14. The geoconvert processing time was around 2.5 hours and it completed successfully with no errors. The resolution is also amazing in VR.

    I've also run similar tests at various levels and sizes but all at the same -1 FSET resolution. What takes time is dependent on how many levels you add to the TMC.

    Wow. Now that is what I call a large area. Do you remember about what the competed size of that PA project is in GBs?

    Regards,

    Ray

  • Hi Jeff, what's your computer specification?

    My computer will crash when converting an area larger than 15nmx15nm, 0.5m resolution (0).

    My computer is 16GB memory, CPU 4790k@4.2.

    It will also take more than 6 hours to compile an area smaller than 15nmx15nm.

    Edited once, last by frui (July 19, 2017 at 7:56 AM).

  • I did the whole state of Washington at 4 m per pixel for good clarity at 2000 feet or above. FSET generated the whole area as 5 x 7 1 GB chunks and because I only have 12 GB of RAM I had to do the Geoconvert as a 7 block strip at a time or else it would crash. I used level 12 as the max.

  • Time to turn in for the night. I have 15 huge bmps to wash, spin, rinse, and dry tomorrow. My area according to the FSET downloader in less than 30 x 50 nm. The bmps look very bright and vivid. My hopes are high that I am entering the fray pretty high on the learning curve.

    Regards,

    Ray

  • Dave,

    Once you do an entire state like you have for Washington, can you chop it up into more bite size pieces? For instance could the Puget Sound area be carved out as a stand alone scenery area? Or maybe something close to the size of the original PNW from Orbx?

    Regards,

    Ray

  • Hi friends,

    you are all here convened.

    I did central europe in seamless strips like this shown below.

    The area coverage is 300x180nm, but with FSET4 and only levels 9,11.

    Note that you can enter 'straight' coordinates into the window.

    Rodeo

  • There's no clear limitation, but before launching a huge zone in high levels, I'd advise always first launching it in level 9 only (by commenting higher levels in your TMC file), to ensure everything works fine. Compilation in level 9 is pretty fast and the result coverage can be checked in the "location" window by zooming in.

    Once everything is ok, just uncomment the higher levels in your TMC and launch.

    It may save you hours of compilation with nothing in the end because there was a mistake somewhere or the file naming bug skipped some imagery, or whatever...

    My 2 cents

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

  • Good idea, Antoine. I made the first pass with levels 9 and 11. Smooth as glass. I am now running the full load.

    Any reason for levels 9 - 14? Would 10 - 14 save anything? Could we tell the difference?

    Regards,

    Ray

  • I think it's a question of memory usage related to when the levels are being loaded in the memory and how long they remain.

    If there's no level 10 underlayer, level 9 of your scenery keeps displaying until level 11 gets loaded.

    Level 9 will use much less memory than level 10.

    Jumping from level 9 to level 14 is a too big step. I assume levels 11, 12, 13 and 14 provide a very smooth texture transition at all altitudes...

    The first pass in level 9 only is fast and allows checking everything, including the AIC files. Adding level 11 for that first pass is just a waste of time IMO...

    Cheers

    Antoine

    Config : i7 6900K - 20MB currently set at 3.20GHz, Cooling Noctua NH-U14S, Motherboard ASUS Rampage V Extreme U3.1, RAM HyperX Savage Black Edition 16GB DDR4 3000 MHz, Graphic Card Gigabyte GeForce GTX 1080 8GB, Power supply Corsair RM Series 850W, Windows 10 64 bit.

  • Hi Jeff, what's your computer specification?

    My computer will crash when converting an area larger than 15nmx15nm, 0.5m resolution (0).

    My computer is 16GB memory, CPU 4790k@4.2.

    It will also take more than 6 hours to compile an area smaller than 15nmx15nm.

    I'm running this on one of my testing machines; i7, 32GB memory, GTX 1080Ti, W10 64, 25TB NAS fiber connected, 1GB internet connection. And yes, before I hear it :) I've tested on a PC with 16GB, i5, GTX 980 as well with no problems either.

    IPACS Development Team Member

    I'm just a cook, I don't own the restaurant.
    On behalf of Torsten, Marc, and the rest of the IPACS team, we would all like to thank you for your continued support.

    Regards,

    Jeff

  • What are some opinions on the source images? I used the default setting of Virtual Earth, but see choices for Google, etc.

    I am downloading at Level 0 for my next test area.

    Regards,

    Ray

    I'm using Virtual Earth for everything. FSET level 0 will also give you good imagery that looks good in VR. Anything less doesn't look too good in VR (my subjective opinion here)

    IPACS Development Team Member

    I'm just a cook, I don't own the restaurant.
    On behalf of Torsten, Marc, and the rest of the IPACS team, we would all like to thank you for your continued support.

    Regards,

    Jeff

  • Just for a contrasting opinion, I did Washington State (350 x 250nm) at FSET 3 (4meters per pixel from VE - levels 9,10,11) and it looks pretty good at or above 2000 feet AGL.

    It takes 3.7gb of disk in its FS2-ready final state. I then add high res areas ( 1 m/pixel or maybe 0.5m/pixel ) for airports or cities.

    Dave

  • Excellent Dave.

    Yep, we are going to be looking for something to add buildings and hangars very soon now. You wouldn't happen to know a source for a few throw down runways, would you?

    Regards,

    Ray

  • What if I created let's say five runways – maybe 2500 foot length, 3000, 4000, 5000, 6000 with reasonable width and BMP textures running north and south so that each individual could edit the BMP with gimp or equivalent to get the runway number correct if they care and then they can set the orientation. How does that sound ? That might be OK for a beginning. You would have to then run the content converter. If you don't need to change the numbers for now then all you would do Is edit the TSC file for orientation and possibly height