NZFF

Welcome to the support forum for VectorLandClass (or VLC for short)

VectorLandClass is a scenery addon for FSX which upgrades landclass, roads/rail, water, seasons, ground and autogen textures and object placement for the whole of New Zealand. You can use this forum to ask questions about VLC as well as keep up to date with the ongoing development of this project (currently at beta stage for a Q4 2010 release)

For more information, or if you haven't already done so, please head to the VectorLandClass website at www.vectorlandclass.co.nz

Tim , Congrads and Damn Well Done , VLC Now the journey begins

Good on ya mate! Now that's what i call PASSION!
Wallet at the ready.

Congrats and good luck Timmo!!

Well done on the free advertising too........ I just caught your plug on The Rock as I drove in to work......

Hi Timmo
I just wanted to pass on my delight and anticipation of the up-coming release of your VLC project . I've been keeping an eye on your work for a while now, and was very, very pleased to see how far things had come with the 2 vids you have up on your website (of Wanaka & Mt Cook). I don't get much time to fly these days, but when I do, I'm primarily interested in NZ VFR only - so the "look" of NZ is of paramount interest / importance to me. The only disappointing thing about it all is the crappy mesh we have - 20 metres: give me a break! Is there any way a bunch of us could pool together to purchase a much finer mesh? Even 1m would be good. Any thoughts regarding that?

Again, congrats on the excellent work you've done to date - I'm REALLY looking forward to the release so I can pony-up the cash and get FSX looking good for NZ!

Regards
Peter

Hi Peter
Bear in mind that the default FSX mesh for NZ is 75 metres, so 20 is a huge improvement. I think that we'll be using the 20 metre mesh for a long time, as this is the only real choice for NZ-wide coverage. My understanding of the 20m DEM is that the resolution is 22 metres horizontally and close to 10 metres vertically, so we may see a mesh marketed as 10 metres at some stage (we may already have seen this), but this would still be the exact same data, just four times the file size. Tim may know something I don't, he's the expert...

It would be nice if it could be higher, but there are some practical limitations to overcome. One is the file size -- if, for example, the 20 metre mesh is 500MB for NZ, 10 metres would be 2GB, 5 metres would be 8GB etc. A 1 metre mesh would be greater than 100GB, so it may be a possibility in five years time, but at the moment many New Zealanders have practical bandwidth limitations.

There's also the performance hit -- each LOD increase (20 10, 10 5) requires four times the processing power. Again, one day we'll get there, performance-wise.

One solution is to build higher res mesh only where it is most important. This is certainly something I'm looking at, not for VLC, but for resulting development, mainly airports of course.

Thanks for the kind words Peter- NZ deserves it!

Robin has pretty much hit the nail on the head in regards to the mesh- The problem is two fold: data and storage.

The reason there isn't a dtm at under 20m pixel size is because the data to fill in the extra pixels simply doesn't exist. Lidar is a technology that will be used more and more- It is, to put it crudely, a device for 'filling in the spaces' in the terrain that would otherwise simply be filled with interpolations of lower res data.

When we make a dtm, we are effectively recreating a continuous surface only using limited and non-continous amounts of data (Of Course, the user wouldn't like flying over a terrain with holes in it! So they must be filled with something). The trade off with this is that more data = greater file sizes to pack the data into. Given that the most detail able to be shown is directly related to its maximum spatial resolution (i.e. smallest pixel size), it actually means you have to set your spatial resolution based on the worst data you have.

The creators of FSX realised that these things are dynamic, the definition (i.e. pixel size) of the terrain is directly related to how far you are away from it. A user that mostly flies high above the terrain simply doesn't need to see 1m pixel terrain except for a brief period at the start and end of the flight. But a user that likes flying helicopters slow and low to the terrain would find it beneficial.
Although that amount of detail covering the whole globe is sometime away, there are small 'pockets' of this information- Generally around cities etc and areas of dense population.

In NZ the situation is similar- There are small pockets of high resolution data around the country, but this hasn't reached the stage where it has 'closed all the gaps' due to the high price of acquiring it. Therefore, it doesn't yet make sense to generate 5m mesh- The extra resolution is simply wasted for 90% of the country and takes up more space!

Having made a few Mesh files over the years it's not so much what LOD it is.. it's the quality of the resulting Mesh that counts most of all.

A good quality 75mt mesh is much better to have than a poor quality 10 or 20mt Mesh.

Edit: I always made a point of designing Scenery to suit 75mt Mesh mainly because designing Scenery to suit lets say 20mt Mesh only meant that I catered for maybe 10 to 25% of Flightsim users.

Cheers
Ian

Thanks to you guys for the comments - very interesting. My main interest in a better mesh is accuracy. In Napier, for example, there is a height difference of 4-5 metres between the "top" of the beach and sea level in many places. But because there is a 20m mesh, it means that all beaches are treated as ground-level. I note that when up in a real aircraft, tooling along at 1500 feet or so still reveals the apparent "height" difference of the beaches. The same is true of various local river beds and a smallish knoll called "Park Island" that is right in the flight path coming in from the South. At the moment, if I crank the mesh all the way up (or should that be down?) to 1m in FSX, Napier looks like the moon with these "craters" all over filled with water! But even then, features like the raised river banks / beaches & Park Island are "flat". In fact, I've often wondered why Park Island is flat - it's well over 20m in height - I'd guess maybe 60-70m. And while I call it a knoll, it's a reasonable size (rough dimensions according to Google Earth are 300m x 300m for the NW knoll, and 600m x 300m for the SE knoll). I've had it in mind for ages to fart about with the scenery for Napier, but it's such a huge job for 1 self-employed person (that is, I work for myself and that means 60-70 hour weeks), that I've have done more than initial research. But as part of that research, I did ask Aerial Mapping about scenery tiles - even as freeware, they were still going to charge me $20 + GST per tile, and I was going to need a few hundred just to do the airfield and immediate area, never mind Napier itself! And all that lead to me thinking about the mesh - if the under-girding mesh is too coarse, features like Park Island and such - which are "iconic" in terms of the look of the area, especially on approaches & takeoff - may still be "invisible" or "amorphous blobs". Hence, my desire to get a better mesh.

I hear what you all are saying about size and computing power. Indeed, whilst computers are heading ever upward into more horsepower (I work as a computer reseller / builder), we are still a ways off having truly powerful machines that can bowl over FSX in all it's maximum glory (including fine meshes). That being said, given how long it takes to develop scenery (and again, hats off to all you folk making the stuff - you are the bomb!), I think it would be useful for us all to "ignore" the current limitations and look to where we will be in 3, 4 or 5 year's time. As most will agree, FSX will be a couple of years away at least before it is meeting sufficiently powerful "day to day" hardware to run well, in the same way that FS9 is really only in the last 1-2 years running sweet on relatively "modest" hardware. As an aside - I upgraded my Quad Core 2.67 (socket 775) sim-PC to 8GB of RAM - that yielded a small but noticeable increase in frame rate / overall performance, with sliders quite a long way up. My next plan is to get a motherboard that will take a max of 32GB and try 16GB with one of the new iCore 7's - but when the cost comes down a bit!

Anyway, I'm just thinking out loud and getting a feel for the finer points - don't let me dampen your efforts: you're all doing great!

Cheers
Peter

P.S. I've attached a couple of pics to show the Park Island thing - used GE for aerial & street view GE for ground

IslandBoy77 wrote:

QUOTE (IslandBoy77 @ Jul 17 2010, 03:20 PM) <{POST_SNAPBACK}>

Anyway, I'm just thinking out loud and getting a feel for the finer points - don't let me dampen your efforts: you're all doing great!

Cheers
Peter

No worries, it's a good way to show simmers and anyone contemplating scenery design what some of the issues really are.

Some of what you seeing is not related to the mesh at all, but rather the water polygons, which tend to flatten the coastline to the level of the water, in this case sea-level. Since the water polys have to be designed to work with any mesh, they need to be a bit over-zealous, cutting into the mesh to some degree. This is an issue with any important features which border the water -- Mt Maunganui is an example, in FS2004 even those with a 20 metre mesh didn't get to see the default Mount as elevated, simply because the coast line was designed for a lower resolution. Upgrading to the 'Roads and Rivers' topo scenery fixed this, as it included a higher resolution coastline, rather than a higher resolution mesh. FSX, though, doesn't have this problem to the same extent, because the default data is a lot better.

When you talk about mesh resolution, generally you not talking about the resolution of the elevation, but the distance between the sampled points. An actual elevation can be anything, not restricted to any resolution, but will only be a computed value based on the general area.

toprob wrote:

QUOTE (toprob @ Jul 17 2010, 04:22 PM) <{POST_SNAPBACK}>

No worries, it's a good way to show simmers and anyone contemplating scenery design what some of the issues really are.

Some of what you seeing is not related to the mesh at all, but rather the water polygons, which tend to flatten the coastline to the level of the water, in this case sea-level. Since the water polys have to be designed to work with any mesh, they need to be a bit over-zealous, cutting into the mesh to some degree. This is an issue with any important features which border the water -- Mt Maunganui is an example, in FS2004 even those with a 20 metre mesh didn't get to see the default Mount as elevated, simply because the coast line was designed for a lower resolution. Upgrading to the 'Roads and Rivers' topo scenery fixed this, as it included a higher resolution coastline, rather than a higher resolution mesh. FSX, though, doesn't have this problem to the same extent, because the default data is a lot better.

When you talk about mesh resolution, generally you not talking about the resolution of the elevation, but the distance between the sampled points. An actual elevation can be anything, not restricted to any resolution, but will only be a computed value based on the general area.

That's very interesting - I wasn't aware of those finer points. So, what sort of hoops would one have to go through in order to get the likes of Park Island to show up closely to what it really looks like? And from what you're saying, the 20m mesh is not making the rivers / beaches crappy - is there a way to fix that - that doesn't require obscene amounts of money? I was thinking about Napier hill also - the 20m mesh at least gives it a sort-off "resembles the real thing" look, but there are plenty of gullies & folds that are missing - and the main cliff facing the port is off.

Thanks again for your comments and bearing with those of us still coming to grips with the sticky details of customisation... like me...