Jump to content
  • 0

Static Tree Lod (Internal,External) BTO Z Axis Placement


Question

Posted
Hi sheson,

 

I am not a forums person, usually solve problems myself, but I am stumped at this one, and I am in need of expert of experts :)

 

I have SFO 2.7, SSE

 

Skyrim Normal Tree Lod Dyndolod Generated - Perfect Transition for Pines, on unrelated: SFO snow pines look bad but still good transition, Aspen have good transition too (As expected since billboards match)

 

Skyrim Static Ultra Trees - 

 

    - 3D Trees passthroughs (Hybrid) only exist for Aspen and Snow Pines except few for SFO 2.7 (I use old ones on SFO nexus that match crc) and if I use those, perfect transition between 3d passthroughs (hybrid) and full models (As expected)

 

    - I tried Internal, 2x2 and 3x2 External Billboards for Normal Pines which does not have hybrid for 2.7 SFO, Same Billboards used for Normal Tree Lod, Now there is the problem:

 

 

The Bilboard matches but with small offset on z axis placement,

 

Full Tree model appears to be %2-3 lower on the plane than the 2x2, 3x2 or Internal static tree lod, so on transition it first looks shorter, but on closer look, lod is placed higher. It is very noticable. (BTW I am not %100 sure if it is z axis placement or if lod is actually %2-3 longer, but it looked like Z axis).

 

But with same billboards and width, height txt, If I generate Normal Tree Lod, perfect transition, height and width and placement perfectly matches.

 

I tried it with Vanilla trees to verify, same problem, static lod (internal, 2x2, 3x3) has a small offset compared to normal tree lod on z axis placement while billboard textures match.

 

Do you have any insight, Am I missing something or could have done something wrong or Is there a small oversight on dyndolod static tree lod generation (internal,2x2,3x2) in terms of calculating z axis while placing on bto files?

 

Sorry I have not tested for other tree models.

 

Side notes:

 

Nothing plays with base tree records for trees.

 

Best

 

6 answers to this question

Recommended Posts

  • 0
Posted

Partially solved it by resizing nif planes in dyndolod_flat3x2_lod.nif to something like 0.95 - 0.97.

 

Still dont have any idea on what is the root cause.

 

Best

  • 0
Posted (edited)

That is by design to counter the butterfly effect.

 

A billboard only really matches a tree when you happen to look at its exact front (and maybe its back). For all other angles and especially the side views the single billboard texture typically does not match very well.

Combine that with the normal use case that billboards in object LOD are usually for LOD level 8/16 which is further away.

 

So, to counter the butterfly effect the exact height value is used for the top vertices of the plane that matches front/back and the height of the plane for the side view is randomly bit higher.

 

In case an external billboard NIF is used, the height is randomly varied for all planes if there is only the typical billboard texture available.

 

If multiple billboard textures (mentioned in Docs/trees.ultra/DynDOLOD-Trees.html) are defined for the different planes, then the exact height value is used for all planes of the external billboard NIF.

 

It should be enough to copy a billboard texture from  SomeTree01_12345678.dds to SomeTree01_12345678_1.dds and SomeTree01_12345678_2.dds to not have the height changed for external 2x2 billboards NIF.

Of course SomeTree01_12345678_2.dds should be the side view for best results. Results will be even better if this is combined with normal map textures.

 

See Docs\trees.ultra\examples\2-textures-billboard-example\textures\terrain\LODGen\skyrim.esm\ folder

Edited by sheson
  • 0
Posted

Thank you sheson,

 

But did not work :(

 

I did it correctly, since in atlas I can see 3 copies of each billboard, however still they are just a little bit shorter,

 

I think it is not that since it is not random, its too deterministic, it is the same amount for all trees.

 

My guess is, I have something, somewhere in my build messes these, maybe bad cache, I dont know, Thanks for your answer through, but I will stick with my little hack by reducing 2x2 and 3x2 scales by 3-4 percent for now.

 

I have another question though.

 

I want Dyndolod to exclude certain flora from static lod which I dont have correct or nice looking billboards. My quickest solution was to put an empty nif as Billboard4 and use it on rules for those items. I dont want esp edits.

 

Is there a easier way?

  • 0
Posted (edited)

There is really no such thing as a cache other then for the native xEdit reference building.

 

The shiftz, width, depth and height values are directly read from the billboard txt. The bottom vertex z should be reference position z + ShiftZ and the vertex top z should be then just the height value added.

If that is not the case then the height is still randomly changed for some reason and I would like to know.

 

Delete the Billboard dds and txt files that you do not want to be used from the data folder. LOD is only generated for trees for which LOD assets exists.

Edited by sheson
  • 0
Posted

Thanks sheson,

 

Well, There is no ShiftZ on affected Vanilla or SFO trees on their [LOD] billboard data.

 

:confusion:  

 

The only other think I can think of is, I dont use skinned trees for personal preference, and turned them off in the ini, skinning should only affect lush so width a little bit, shouldnt affect height or z placement.

 

Other than that 

 

Tree only SFO 2.72, with SFO 2.72 billboards, no tree scaling on base reference or tree references throughout all mods , nifs are original, mods are original reinstalled them. Nothing overrides billboards, nifs, esps ....

 

Cant think of anything else, I will let you know If I figure it out.

 

Best

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
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

By using this site, you agree to our Guidelines, Privacy Policy, and Terms of Use.