Rule of 3rds overlay view

[Ashley]

Hi,

Is there a "Rule of 3rds" or "Grid" overlay view in the 3D navigator window?
I find I need to plot points, so far I'm holding up a piece of transparency paper with a grid drawn on it :-)

Any tool that does that in TG?

Thnx
Ash

[cyphyr]

Press 3 on the keyboard.

[Ashley]

That's great thank you,
Is there a way to change the divisions?

[Matt]

At the moment it only allows 3x3 equal divisions.

Matt

[Upon Infinity]

Just discovered this feature, and appreciate that it's here.  You may wish to modify it, however, to maybe replace it with something subtler like a dashed line, or a less aggressive colour or both.

[Ashley]

Its certainly a good feature. Would be cool to see future updates include row and column control.
Perhaps sub rows and columns could have a dashed line.

I use this to help me plot cameras for geometry export.

Cheers

[Oshyan]

I think if you really want a tool like that with greater configurability, you should probably look to a 3rd party utility that works more generally. Something like this:
http://www.phimatrix.com/

- Oshyan

[Ashley]

Actually what I am trying to do needs to be done in TG.

So here's an example

- Have a terrain that I want to export as an obj.
- Create a camera, set it to orthographic and rotate it to point down at the planet.
- Set camera position to the middle of the terrain and make sure the ortho width is square and, of course, the camera altitude is high enough to encompass the whole terrain.

Simple enough, but the issue is the resolution of the mesh is not high enough.
The solution is to export smaller patches of terrain. Since TG doesn't do this automatically, its all wysiwyg I need to position a camera at every location to extract that patch of geo.

So thankfully I can animate a camera moving across the terrain and write out an obj sequence.
The next issue is how do i accurately plot/place the camera?

Well if I have a grid overlay that would at least guide me to see where the positions of the camera need to be then I can copy paste these x,z positions to my camera and key frame it per position.

So far the only option in TG is the rule of 3rd overlay. I usually need to export more than 9 patches to achieve the resolution needed. So a grid of equal rows and columns that could be adjusted would be ideal.

I have literally be holding up a grid printed on transparent plastic sheet to my monitor 

Attached is an image of what I mean, top down screen grab of the default TG scene.

Cheers
Ash
 

[Oshyan]

Hmm. It sounds like your use-case is incredibly specific, and even more of the kind of thing that doesn't really justify a feature change in the main application. Fortunately, if I understand correctly, it should be fairly easy to solve your need. Just create an image file with the grid you want, then apply it to the terrain and set the size equal to the size of the *total area* you want to export. Then you should see the grid squares. To make them show up better in the preview you could make them solid colored boxes instead of grid lines (which, being thin, might not show up well with the preview level of detail).

You may even be able to do something natively with the Simple Shape Shader and a Shader Array. See attached for a simple example.

- Oshyan

[Kadri]

Adding to what Oshyan said have you tried to use crop rendering?
I haven't.So i don't know if it works.
You could cover the entire area you want with the camera and then only crop render the parts without changing the camera position.
I had always the opposite problem as the files were very big i got from Terragen.

[Ashley]

I could use a Shader array to plot the points, only thing is when I mouse over the actual points the co-ords default to the xyz of the shader and I need to mouse slightly next to the points to get the terrain/world co-ord.

Crop region render would work but trying to animate crop left, right, bottom, top is more frustrating than herding catz.

I think I'll stick with my transparent plastic grid for now.

Thnx for all the comments 

[Hetzen]

Have you considered exporting .ter from a generate heightfield combination, with your terrain as the input. There's an app I use to convert .ter into a displaced square mesh within max, called groundwiz. Gives clean welded results. The resolution is set by how many pixels you dedicate to the heightfield square defined in TG, by how many verts you use to those pixels.

An added advantage, is you can post process errosion into the .ter to bring back into TG.

The mapping process can be split between as many heightfield nodes as you want.

[Ashley]

Thnx Hetzen,
Sounds like we have a similar workflow. If I understand correctly you would need to generate a heightfield at 4096x4096 for approx 4mil poly. I need approx 20mil per 1km patch, actually its probly higher :-)

From my experience rendering out that resolution in pixels from TG takes a loooong time. Its actually faster to export obj at that resolution (poly count). With the added bonus that obj will capture undercuts as well.

I use heightfields and obj to move between apps and although TG and WM handle heightfields/dispalcement well the same can't be said for the apps the final will be rendered in. Its not consistent across the board so I go with raw geometry at high poly count to ensure I always get what I am expecting.

Back to the grid thingy, all I'm after is an automated way to export very large terrains. If you havn't already download a trial of World Machine and look at its tiled export features. Basically you set the extents of the world you want to export, then set per tile resolution. There is nothing in Terragen to aid in automated batch process, except what I have described above.

[Hetzen]

This may not be you preferred approach Ashley, the Obj export doesn't export undercuts if the camera is straight above either. There will be holes in the mesh where top down shadow would be. An HF will fill those gaps, not with an undercut though. But usually acceptable for camera projection on non hero parts.

You can also designate any resolution and size for the HF output too and once you set up a grid of HFs with overlap to blend you have the same type of tile output you get in WM. Admittedly it's not as easy to set up, but where it's superior to WM is the amount of ram required to process each square. WM will need to process some aspects of the whole heightfield to render a tile, so you will soon eat into a lot of ram outputting 8-16k tiles. TG will work on a tile at a time. Which you can process in WM one at a time.

Also, with this approach, you can optimise areas that will not have as much camera focus, so render out background formations at a lower rez.

I guess it's all about what the scene is, to work out the best possible route you'll need for the effect you're after. Rendering TG terrains in other apps will always fall short of how TG renders displacement due to how TG tessellates at render-time, thus cutting down on app poly count. Plus transferring texturing is limiting without camera projecting. I've found that I use exported terrain into other apps for shadow mattes and positional sense for camera and animation work, export the camera back into TG and render those plates. Although that depends again on what the scene is.

[Kadri]

... the Obj export doesn't export undercuts if the camera is straight above either. There will be holes in the mesh where top down shadow would be...

Not sure about that.
I tried it with a strange ground displacement and except some parts (probably mostly because of too much displacement) it came out without holes.And it has more then one possible problematic part even.

Edit: At the render time the camera was above the stucture. The preview you see below is to show the object as it is in Terragen.