WM plugin a must

Started by edlo, January 13, 2007, 09:18:26 AM

Previous topic - Next topic

Will

@JimB Is it tiled? I know they look the same while moving and such but I think that they are diffrent when you acauly select them. I could be wrong though, What I was saying is that you could create an entire world with erosion and what not in around five minutes I would, of course, depend on how much that 512x512 area would repersent on your planet if its only like the size of the default heightfeild then yes it would take a long time, but if (and this is what I was thinking) it took up maybe an sixth of a planet then it would not take long at all.

Regards,

Will
The world is round... so you have to use spherical projection.

edlo

Quote from: JimB on January 16, 2007, 05:22:03 AM
Quote from: Will on January 15, 2007, 06:27:54 PM
less then 5 minutes. depends on your computer, World Machine already mades a world (exploration mode)
On a 512x512 tiled part of the entire planet maybe, but not on the whole procedural terrain, no way. Are you saying you want repeated tiles, or you want an entire procedural landscape from WM?

Quote from: edlo on January 15, 2007, 06:35:54 PM
You can apply much better erosion models right from WM
Which is what I was talking about - within WM.

Which is the whole point of these post; compatibility of WM files on TG2, not terrains...Files, and yes WM applies weathering and erosion to the complete procedural model.

JimB

Okay, I think I see what you're driving at. But, if, for example, you created a nice Perlin procedural terrain in WM and added an erosion node, etc, and then imported that into TG2, wouldn't it still have to calculate the erosion on the perlin terrain anyway, somewhere down the line? I don't know how powerful your PC's are, but doing the final build on a 2048x2048 tile in WM with extra quality settings on the erosion can take up to half an hour on my machine - and it ain't the slowest in the world. I don't see why the same thing wouldn't have to happen in TG2.

No matter what, the procedure has to be calculated on the heightfield to get the final terrain.
Some bits and bobs
The Galileo Fallacy, 'Argumentum ad Galileus':
"They laughed at Galileo. They're laughing at me. Therefore I am the next Galileo."

Nope. Galileo was right for the simpler reason that he was right.

edlo

Quote from: JimB on January 16, 2007, 11:40:27 AM
Okay, I think I see what you're driving at. But, if, for example, you created a nice Perlin procedural terrain in WM and added an erosion node, etc, and then imported that into TG2, wouldn't it still have to calculate the erosion on the perlin terrain anyway, somewhere down the line? I don't know how powerful your PC's are, but doing the final build on a 2048x2048 tile in WM with extra quality settings on the erosion can take up to half an hour on my machine - and it ain't the slowest in the world. I don't see why the same thing wouldn't have to happen in TG2.

No matter what, the procedure has to be calculated on the heightfield to get the final terrain.

What WM does right now when you do the 2048x2048 terrain its rendering the information onto a height field that you selected, either a .ter file a .bmp or any other, and then you load that onto TG 9.43 OR TG .
So it takes a long time because it is rendering an image; if you could import the node file onto TG2 the result would be the same as when you add a power fractal, which does not take years to apply and covers the whole planet.
With this compatibility you would be able to build up your own power fractal or alpine fractal terrains to be used on TG2.
Cheers.

Will

^What I was trying to get arcoss, sorry.

Regards,

Will
The world is round... so you have to use spherical projection.

Oshyan

Quote from: edlo on January 16, 2007, 12:57:25 PM
Quote from: JimB on January 16, 2007, 11:40:27 AM
Okay, I think I see what you're driving at. But, if, for example, you created a nice Perlin procedural terrain in WM and added an erosion node, etc, and then imported that into TG2, wouldn't it still have to calculate the erosion on the perlin terrain anyway, somewhere down the line? I don't know how powerful your PC's are, but doing the final build on a 2048x2048 tile in WM with extra quality settings on the erosion can take up to half an hour on my machine - and it ain't the slowest in the world. I don't see why the same thing wouldn't have to happen in TG2.

No matter what, the procedure has to be calculated on the heightfield to get the final terrain.

What WM does right now when you do the 2048x2048 terrain its rendering the information onto a height field that you selected, either a .ter file a .bmp or any other, and then you load that onto TG 9.43 OR TG .
So it takes a long time because it is rendering an image; if you could import the node file onto TG2 the result would be the same as when you add a power fractal, which does not take years to apply and covers the whole planet.
With this compatibility you would be able to build up your own power fractal or alpine fractal terrains to be used on TG2.
Cheers.

That's not really true - the "rendering an image" part of it has very little impact on calculation time. It's just that many of WM's functions are very processor intensive, similar to TG2's own erosion and Heightfield Generate algorithms, both of which take a bit of time. Loading a WM file in TG2 could be great and I'd love to see that supported, but it's not going to make generation of even a finite area of WM terrain any faster, let alone an entire planet. Erosion in particular is a highly demanding, iterative process. To do it at any decent resolution on a whole planet would take ages - days at least, if not weeks.

It depends of course on the resolution of each tile you were working with. The Explorer mode in WM does seem fairly fast but there are several things working to your advantage in that case. 1 is that the tiles are generated progressively (they start out at low resolution and then are updated with higher resolution versions as they are generated), 2 they are less detailed the further away from the camera they are, and 3 they have a detail cap of 256 pixels square maximum (usually set to 128 or lower for speed reasons). If those resolution limits were acceptable to you then total planet generation might be done in an acceptable time of a day or less. But this doesn't take into account the memory and hard drive use of all these tiles either.

All that being said there is actually a way that a WM plugin could do some pretty cool stuff by working with the WM base networks and procedural functions, but visualizing them in a TG2 way - *at render time*. The reality is that TG2 doesn't pre-generate an entire planet, it visualizes only the detail it needs for each particular render, during the render itself. If it had to precompute an entire planet it would take ages, just like WM. So if a WM plugin could do the same that would be great - you could have the versatility of WM terrains but with theoretically infinite detail and planetary coverage. The only major issue I see is in making functions that are non-procedural (e.g. erosion) work in this way. I think it would be quite difficult to make erosion work properly in such a situation. But I'd love to hear otherwise from Stephen as this would be a fantastic addition to TG2 for sure. :)

- Oshyan

edlo

Quote from: Oshyan on January 19, 2007, 07:36:29 PM
Quote from: edlo on January 16, 2007, 12:57:25 PM
Quote from: JimB on January 16, 2007, 11:40:27 AM
Okay, I think I see what you're driving at. But, if, for example, you created a nice Perlin procedural terrain in WM and added an erosion node, etc, and then imported that into TG2, wouldn't it still have to calculate the erosion on the perlin terrain anyway, somewhere down the line? I don't know how powerful your PC's are, but doing the final build on a 2048x2048 tile in WM with extra quality settings on the erosion can take up to half an hour on my machine - and it ain't the slowest in the world. I don't see why the same thing wouldn't have to happen in TG2.

No matter what, the procedure has to be calculated on the heightfield to get the final terrain.

What WM does right now when you do the 2048x2048 terrain its rendering the information onto a height field that you selected, either a .ter file a .bmp or any other, and then you load that onto TG 9.43 OR TG .
So it takes a long time because it is rendering an image; if you could import the node file onto TG2 the result would be the same as when you add a power fractal, which does not take years to apply and covers the whole planet.
With this compatibility you would be able to build up your own power fractal or alpine fractal terrains to be used on TG2.
Cheers.

That's not really true - the "rendering an image" part of it has very little impact on calculation time. It's just that many of WM's functions are very processor intensive, similar to TG2's own erosion and Heightfield Generate algorithms, both of which take a bit of time. Loading a WM file in TG2 could be great and I'd love to see that supported, but it's not going to make generation of even a finite area of WM terrain any faster, let alone an entire planet. Erosion in particular is a highly demanding, iterative process. To do it at any decent resolution on a whole planet would take ages - days at least, if not weeks.

It depends of course on the resolution of each tile you were working with. The Explorer mode in WM does seem fairly fast but there are several things working to your advantage in that case. 1 is that the tiles are generated progressively (they start out at low resolution and then are updated with higher resolution versions as they are generated), 2 they are less detailed the further away from the camera they are, and 3 they have a detail cap of 256 pixels square maximum (usually set to 128 or lower for speed reasons). If those resolution limits were acceptable to you then total planet generation might be done in an acceptable time of a day or less. But this doesn't take into account the memory and hard drive use of all these tiles either.

All that being said there is actually a way that a WM plugin could do some pretty cool stuff by working with the WM base networks and procedural functions, but visualizing them in a TG2 way - *at render time*. The reality is that TG2 doesn't pre-generate an entire planet, it visualizes only the detail it needs for each particular render, during the render itself. If it had to precompute an entire planet it would take ages, just like WM. So if a WM plugin could do the same that would be great - you could have the versatility of WM terrains but with theoretically infinite detail and planetary coverage. The only major issue I see is in making functions that are non-procedural (e.g. erosion) work in this way. I think it would be quite difficult to make erosion work properly in such a situation. But I'd love to hear otherwise from Stephen as this would be a fantastic addition to TG2 for sure. :)

- Oshyan

Brilliant now we have all the facts so it would be possible and would be a great addition to TG2 but if the list of procedural and procedural affecting nodes grows beyond that of power fractal or alpine fractal we wont need WM at all. The only difference I see now between WM nodes and TG2 nodes are that WM nodes are much more terrain function oriented, instead of just function oriented which makes it easier to understand them individually and the effect that will cause to your terrain.
So its just a matter of user friendly nodes  ;D

JimB

Quote from: OshyanThat's not really true - the "rendering an image" part of it has very little impact on calculation time. It's just that many of WM's functions are very processor intensive, similar to TG2's own erosion and Heightfield Generate algorithms, both of which take a bit of time. Loading a WM file in TG2 could be great and I'd love to see that supported, but it's not going to make generation of even a finite area of WM terrain any faster, let alone an entire planet. Erosion in particular is a highly demanding, iterative process. To do it at any decent resolution on a whole planet would take ages - days at least, if not weeks.
Like I was trying to get across  ;)

Perhaps a chan file import/export to calculate and create the tiles in WM for use in TG2 would be an option? I.e., the visible areas in a TG2 animation are calculated within WM without the need to erode, etc, an entire planet, and only those necessary terrain tiles exported from WM to into TG2.

I have no plans to give up on WM, and would love to see the two apps co-operate in that way.
Some bits and bobs
The Galileo Fallacy, 'Argumentum ad Galileus':
"They laughed at Galileo. They're laughing at me. Therefore I am the next Galileo."

Nope. Galileo was right for the simpler reason that he was right.

Oshyan

World Machine and Terragen have always worked well together and I expect that to continue in the future. I too would love to see closer integration between the two, and with the more robust SDK that will be provided for TG2, it could well be possible. We'll just have to wait and see. :)

- Oshyan

monks

Is there no support then for importing tiled worlds from World Machine into TG? Have people done it manually?

monks

Oshyan

No direct support at present, but it's a good idea. I think it could be done manually, though I don't know how involved that might be. It would be made easier if there is Georeferencing support in WM.

I definitely want to see interoperability with apps like WM improve in the future.

- Oshyan

monks

Hmm, I'll have to go with a single terrain.

monks


Oshyan