Terrain questions

[Stephanie]

hey

I am working on a wall-model, which is more than a kilometer long. The
Terrain in Google Earth is too sloped, so the advisement is that I
should generate a sourounding matching terrain or bouild my model

[Gully Foyle]

Hey, Stephanie.

Despite the subject of your post, you didn't actually ask any
questions.

However, if the terrain you imported from GE is exaggerated along the
vertical axis, you can simply flatten it out some.

First, right-click the terrain object and select Unlock. Then, with
the thing selected, Use the Scale tool, push down the top-center
control node, flattening out the object by reducing its verical scale.
You can type in a number, like .5 (to make it half as high), or just
eyeball it.

-Gully

[Stephanie]

sorry I sent it too early - here again

hey

I am working on a wall-model, which is more than a kilometer long. The
Terrain in Google Earth is too sloped, so the advisement is that I

should generate a surrounding matching terrain or build my model in
the
sloped way.

both are not satisfiying for me, because the wall isnt sloped that
way,
building it the along the terrain it will also give a strange look
when terrain is toggeld off,
or in case somebodey changed the elevation exaggeration in the
preferences - it wouldnt
match anymore.
when I build a surrounding terrain it will look odd overall and I have
to texture it again
when the airial photo of that area changes.

so my solution is for now, that instead of building one model, I
segmented that wall
into hundreds of pieces, each is now an own model which is "clamped to
ground"
in the kml file. this is also the point where I stopped using sketchup
for creating the
model.

unfortenuately these hundreds of segments load very slowly in google
earth, so my
question is, is such a model still suitable for the building layer,
maybe loading gets faster
in the future?

or has anyone another idea for a solution?

here is the model:
http://sketchup.google.com/3dwarehouse/details?mid=108929f713a250ddf5065354ba24558d&ct=mdsa

stephanie

[VAM]

Hi Stephanie,

I am not sure wether I understood your problem. As I am not familiar
with the situation in Berlin it might help if you can show a picture
of a range of the wall that we can see how it is segmented in reality.

As far as I remember parts of the Wall were built out of prefabricated
concrete-segments with a concrete pipe on top. Are those segments
jigsawed to meet the differences in terrain-height or do they have a
smooth upper edge?

Wolfgang

> here is the model:http://sketchup.google.com/3dwarehouse/details?mid=108929f713a250ddf5...
>
> stephanie

[matthiasbasler]

> The Terrain in Google Earth is too sloped, so the advisement is that I
> should generate a surrounding matching terrain or build my model in
> the sloped way.

Your third solution doesn't sound practicable to me. Splitting the
model into hundreds of small models makes it probably impossible for
Google Earth to put them into the "Best of" layer - and it makes it
definitely impossible for people to download the whole map since they
would have to search for all the pieces and download them one at a
time. I don't think anybody would have this endurance.
Splitting the wall into 2 or 4 pieces, on the other hand, is OK. This
will keep size of each model reasonably small and will allow GE to
load those pieces first that are nearer to the camera.

If I were in your shoes I would try a mixture of
- adding surrounding matching terrain
- Have the wall be slightly sunken in (but max. 1m or so) in some
areas where there are maybe "terrain spikes"
- just live with a slightly too slopy wall. Someone not living there
will likely not notice the difference. ;-)

P.S. You say a sloped wall would look strange when terrain is toggled
off or terrain exaggeration is no 1.0.
I'd definitely not care about this. If someone turns off terrain or
exaggerates it this is his/her own fault that a lot of models will
look rather strange.

Of course this is just my opinion.

[Stephanie]

thanks for your thoughts

On 19 Jul., 10:21, VAM <wolfg...@lemke-soeder.de> wrote:

> with the situation in Berlin it might help if you can show a picture
> of a range of the wall that we can see how it is segmented in reality.

the wall at this is part looked more like being build on flat terrain:

http://www.dailysoft.com/berlinwall/xgraphics/photographs/eastsidegallery/19970503-30-0502.jpg
http://www.perspektive89.com/system/files/images/15EastSideGallery1.jpg

> As far as I remember parts of the Wall were built out of prefabricated
> concrete-segments with a concrete pipe on top. Are those segments
> jigsawed to meet the differences in terrain-height or do they have a
> smooth upper edge?

It is made out of that slabs and the top has that familar tube and is
not jigsawed. I just made a flat wall to save work and simplify the
model. Also it is not very depth: the wall is not more than 15-20cm
thick, the tube 40cm. Currently I focus on the textures and a way to
overcome the sloped terrain.

So my segmented model does not look correct because of the jigsawed
appearance when the terrain is turned on and a model which follows a
sloped path would unfortenuately never appear the correct straight
way. I also like to show the wall as a long straight line because I
think this is an important characteristics of it (at this part, there
were of course sloped parts of the wall).


On 19 Jul., 10:48, matthiasbasler <matthiasbas...@earthflight.org>
wrote:

> Your third solution doesn't sound practicable to me. Splitting the
> model into hundreds of small models makes it probably impossible for
> Google Earth to put them into the "Best of" layer - and it makes it
> definitely impossible for people to download the whole map since they
> would have to search for all the pieces and download them one at a
> time. I don't think anybody would have this endurance.

Currently I put all the parts together in one kmz-file to have them
all in one file/link and upload that kmz direct to the warehouse. I
can imagine that the warehouse is not intend to do so. And maybe
modeling that way, having different models in one kmz-file, is not
suitable for the "best of" layer at all?

> Splitting the wall into 2 or 4 pieces, on the other hand, is OK. This
> will keep size of each model reasonably small and will allow GE to
> load those pieces first that are nearer to the camera.

I started with that Idea, ending up then with my segmented wall by
increasing the numbers of pieces to avoid bigger steps :)
But for a new try perhaps more than 4 pieces are necessary, because it
seems not possible to import large terrain areas into sketchup
correctly. Terrain does'nt match in the farer areas (from the origin)
and the models are always skewed and/or misaligned.

> If I were in your shoes I would try a mixture of
> - adding surrounding matching terrain
> - Have the wall be slightly sunken in (but max. 1m or so) in some
> areas where there are maybe "terrain spikes"
> - just live with a slightly too slopy wall. Someone not living there
> will likely not notice the difference. ;-)

maybe I have 2 models, one more like the way you suggested and a flat
one to preserve the characteristic. Then have a kml networklink so
that a server decides (depending on current terrain on/off setting)
which model to show. that of course is only working with a server
which is parsing the informations from the network link - beside the
double work, I guess putting such a model to the warehouse won`t make
sense anymore.

> P.S. You say a sloped wall would look strange when terrain is toggled
> off or terrain exaggeration is no 1.0.
> I'd definitely not care about this. If someone turns off terrain or
> exaggerates it this is his/her own fault that a lot of models will
> look rather strange.

I can understand this, but for my model I'd love to have a solution
for at least a good view with the normal terrain and the terrain off
because here I think the terrain is displayed too sloped and a "flat"
model is more natural. But for exaggerations not 1.0 you're right it
is rather impossible to model for all settings.

stephanie

[VAM]

Hi Stephanie,

I had a quick reserch in other European capitols. In Madrid for
instants Google has no problems putting models with 2-3 MB into
the ,,Best-of-layer''. Your wall is a big important model - therefore
the data-load also needs to be more heavy. That is my opinien.

On the picture on your second link I noticed that the upper-edge of
the wall is not even everywhere. Furthermore I noticed a foundation of
about 50 cm which you could use as tolerance-zone. And randomly I
measured a height-difference of 2 m on 80 m length or 2,5%.

To give your model a realistic impression I would create a component
with 1,4 m length, 0,2 m width and whatever the height is together
with a concave base. Then take five, six, seven or how many segments
it is in reality together and give them a concrete pipe (or tube) of
about 10 m length - also as component - on top. Don't forget to
use ,,beliebig binden'' and ,,durch Schnitt öffnen'' as well as
adjusting the axis properly. These bigger segments I would let flutter
a few zentimeters. Together with the base you should be able to meet
the height differences and give the wall a natural impression as well.

If you give the faces in the components a standard-material (white-
gray) you can cover the component-instances with individual textures.
The pipe and the base can be covered with concrete-material in the
first instance as they are mosty unpainted. Using components the file-
size of the pure wall will stay low. Only the textures will increase
the file-size. But the polygon-count still will go up. Therefore be
aware to get rid of any unnesseccary edges and faces in the component.

This technic has two advantages. You can cover the backside with a
realistic texture and you won't get problems with opacity.

I hope my writing is not too confusing

Wolfgang

On Jul 21, 7:48 pm, Stephanie <s...@stylepolice.de> wrote:
> thanks for your thoughts
>
> On 19 Jul., 10:21, VAM <wolfg...@lemke-soeder.de> wrote:
>
> > with the situation in Berlin it might help if you can show a picture
> > of a range of the wall that we can see how it is segmented in reality.
>
> the wall at this is part looked more like being build on flat terrain:
>

> http://www.dailysoft.com/berlinwall/xgraphics/photographs/eastsidegal...http://www.perspektive89.com/system/files/images/15EastSideGallery1.jpg

[Stephanie]

> If you give the faces in the components a standard-material (white-
> gray) you can cover the component-instances with individual textures.

hey thats a good idea, didnt know that before.
thanks! think it will take a while before i will
come up with the new model.

stephanie

[Stephanie]

ok

here is a new aproach, its another section of the wall and I tried to
model it following the sloped terrain implementing some of your hints,
beside it isnt sloped there in reality that way, too. the tube is
still missing, but I hope you can imagine how it will look with one on
top.

http://sketchup.google.com/3dwarehouse/details?mid=b739564d2f3d03c35dfcd45c600ac8dc

I guess I will model from now two versions of each wallpiece which
appeared be too sloped, one for the actual GE terrain and one for
terrain to set off.

thanks for your ideas.

stephanie

[VAM]

Hi Stephanie,

the model looks great. It has now the real solid ,,wallish'' look. I
liked the tour best. Maybe you should explain new GoogleEarth-visitors
how to start the tour.

You're doing a great job

Wolfgang

On Aug 19, 8:08 pm, Stephanie <s...@stylepolice.de> wrote:
> ok
>
> here is a new aproach, its another section of the wall and I tried to
> model it following the sloped terrain implementing some of your hints,
> beside it isnt sloped there in reality that way, too. the tube is
> still missing, but I hope you can imagine how it will look with one on
> top.
>

> http://sketchup.google.com/3dwarehouse/details?mid=b739564d2f3d03c35d...