I did wind up with 1,700 edges in this example, as opposed to 4 edges in the alpha-image version.

We are working on a separate project to create dual complexity components - with the full component in one layer and a simple alpha-image, (or even a simple rectangle) in another layer. This idea should work well for your plants, because the user could use the alpha-image while modeling, and switch to the HI-POLY layer (for shadows) only when ready for the final rendering in SketchUp.

Our tests have shown that this works well for components like plants with multiple instances. SketchUp remains very fast while the HI-POLY layer is turned off.

The SMustard script did the trick. It took about 5 minutes to process an image.

Here is a grouping of 5 plants as placed in SketchUp. (Note the shadows. The advantage of these polygon - non alpha plants is that they can casr shadows in SketchUp.)

Here is the grouping rendered with IRender Plus. Note the 3D shadows, and the 3d, fractal effect of the rendering. Each plant is rendered differently. This is especially noticeable at the top of the plants on the left.

Conclusion, the SMustard component works pretty well, and is better in SketchUp because of the shadows. The only other disadvantage it that the SMustard component is about twice as large (because of the complex polygon created to cast the shadows.)

@sorgesu said:

1. Edit the component.
2. Select all of the geometry of the component and "copy" which will copy it all to the clipboard.

I;m not sure how to emulate "Edit Component" from a ruby script.

Ditto, for copy and paste.

I'll take a look, since I'll only have to run it once per image to create the library.

On the other hand, if it only traces alpha around the outside, and doesn't pick up transparent holes, it won't work for many of the plants. Such as this one with lots of internal holes:

Note to SUC moderators: Is there a way to set the forum to show alpha parts of images in white rather than black?

@remus said:

Do they render well? (if at all, i imagine its quite dependent on the engine )

They probably render well in any renderer. Especially those which cast shadows properly from alpha images.

As it turns out, they automatically render as realistic, fractal, 3D plants using IRender Plus. (But they are of use for anyone, and should be quite useful as they are)

Here is an example:

We wanted to take all the exiting geometry in a component definition, and put it in a group within the same component definition, so that we could put the new group in a layer, which made it easy to turn off a layer to hide all of the existing geometry. (This was part of our project to convert hi-poly components so that they could be displayed as lo-poly proxies just by turning layers on and off)

We came up this this logic:
(I thought I would share this here in case anyone ever need to perform a similar task)

1. Make a new dummy component instance with the same definition as the original component

2. Make the dummy component instance unique. (It now has a cloned copy of the geometry in the original component definition)

3. Erase the geometry from the definition of the original component.

4. Add a group to the original component definition and give it the layer desired.

5. Place an instance of the dummy component into the group in the original component.

6. Explode the instance, in the group in the original component definition.

• (You can also add additional information to the original component - such as the low-poly geometry.)

7. Erase the dummy component instance and the dummy component definition.
   (I was not able to figure out how to remove a component instance - I used purge_unused instead)

Here is a sample ruby script which does this:

lo_poly.rb

This seems to work well, but is rather complicated.

Is there some easy way to do this that I am missing?

@andeciuala said:

hi all, i'd like to cooperate, with you, and share some models that i found in 3dwarehouse, most for architecture, toys, cars, furniture, interior objects, etc...

One good way to do this is to create a collection in the 3D Warehouse and add all the existing 3D Warehouse components which you want to share to the collection. Then you could post a link to the collection.

We have created a library of about 350 plants. These are created as "always face camera" alpha-transparent images. They work pretty well in SketchUp, however, unfortunately, alpha-transparent images do not cast shadows. (There was someone working on geometry to convert alpha-transparent images to complex faces, so you could place the image on the face and have it cast shadows properly. I'm not sure if that was ever completed.)

The alpha-image plants have a very low poly-count (1 image per plant), so they do not clog up your SketchUp drawing.

Here is a sample: (also available as a .SKP model below)

(The black background is the transparent part of the plant. It will be transparent in SketchUp)

Sample Plant is SKP file

Here is a SketchUp drawing with a couple of the plant components:

I am attempting to load this library to the 3D Warehouse, but I am waiting for help from SketchUp on how to perform a bulk upload. A few samples are uploaded already in this collection:

Error 404 (Not Found)!!1

(sketchup.google.com)

I have also placed the entire library in a zip file:

ftp;//ftp.renderplus.com/_download/IRender/IRender%20Plants.zip

(This is a 40 meg download)

You can extract this zip file into your SketchUp Components folder and use the plants.
Perhaps someone can try one of these and let me know what you think.

We're working on a new tool to place a material on a face in a component (without having to open the component for edit first.)

The tool locates the face under the cursor, backs up until it finds where the current material is defined for the face (if it is in the default material), and fixes everything so that the face appear with the proper material - all in one step.

Is this a useful idea and/or has someone else already done it?

We are doing it directly from an explorer like material selection tool. (The Windows Explorer approach is nice because you can have multiple branches of the search tree open and move back and forth quickly - for instance if you want to place materials from more than one branch onto to various surfaces quickly.)

But the real value is quickly placing materials into components, and nested component, group, combinations.

Thanks,

These previous three comments are quite helpful.

What we are working on is a way to automate the process.

The user will:

1. right click on an existing component instance and select "Create Lo-res Proxy"

2. The user will select options from a dialog box:

2a. Alter existing component definition or create new component definition with "_proxy" added to the name.

• There are good reasons for both choices - for instance if the components are being counted by component name, then you would want to retain the current name.

3. The user will choose one of several low-res options

3a. simple box using component extrema

3b. Always face camera image of component thumbnail (will not loof perfect, but will at lest use an alpha transparent image of the thumbnail so you can tell which component is which)

3c. Cris-cross images of thumbnail

3d. Thumbnail places on all sides of box

4. The new component will have old geometry in a group in layer "RPS HI-RES", and new, proxy geometry in a layer "RPS LOW-RES", so it will be easy to flip flop between the display.

5. External renderers can automatically extract the HI-RES geometry without having to turn on the layer within SketchUp.

In the next day or so we should have a "timing test" for SketchUp models and some complex components. (we are going to time a single operation - like "Zoom Extents" to make sure the approach works, and to see what the differences between options 3a-3d are.

@burkhard said:

Nice one but it doesn't help Sorry

I got Remus set up properly.

Did you want to join SuWiki as well. If so, send me an email.

@gksl4 said:

Sketchup exporter to Indigo from Whaat does a great job with instancing.

This thread isn't really about instancing in external renderers, but about applying a similar concept - (perhaps I should have called it "Dummy Instancing") - for SketchUp - to use low poly instances while modeling and the converting them to high-poly instances when ready for final rendering within SketchUp.

If the idea makes sense (to make SketchUp faster), then external renderers could take advantage of the layering concept by turning on the High-Poly layer and turning off or ignoring the "Low Poly" layer when extracting geometry from the SketchUp model.

We are going to try this today.

We will put two layers in the component:

"Low Poly" will contain the image of the component, or a "block shape" as suggested by sorgesu, (probably a user option) and

"High Poly" which will contain the original geometry.

We will either: a.) create a new component with a new name so the high-poly layer can contain an an instance of the original component. or b.) keep the original component name, but place the existing geometry in a group so we can give it a layer. - Both of these have advantages, so we may make it a user option.

This should just work. (It will be easy to test it by creating the same two layers by hand)

The current plan is to add this to RpTools and IRender. People who don't want to purchase either product would be able to use it for free in the IRender Trial version (even after the trial period expired) and perhaps would want to try the "Instancing" in the IRender engine to see how well it works. (In IRender you can still test renderings after the trial version - but you get a watermark which gets worse and worse as time goes on.)

We had to shut down manual sign-ups because we were getting too much spam.

I will set you up as a new user, send me your email address and desired password to

al DOT hart AT renderplus DOT com

I have been thinking about a "dummy component" tool for SketchUp to replace high-poly component instances with low-poly versions while modelling, and then restore the high-poly versions for final rendering.

Basically, the idea is this:

1. You would place one or more high-poly instances of a component into your model.

2. The "dummy component" tool would create a low-poly, dummy version of the component by creating a new component with a alpha transparency, raster image of the original component in it instead of all the geometry and replace all but the first instance with the dummy component, and then let you place additional copies of the dummy component into the model.

3. The "dummy components" would probably look dumb, but would lower the number of faces which had to be handled by the OpenGL viewer. This should make it much faster to work with the complex model.

4. When you were ready for a high-quality rendering, the tool would replace the dummy component with a copy of the high-poly version (by including the high-poly version as a sub-component) (This could also be easily done with layers - "low poly" for the image, and "high-poly" for the sub-component)

5. Rendering engines could be "smart enough" to read the high-poly layer even though it was turned off, or you could turn on the high-poly layer just long enough to create the rendering.

Is this a good idea?

Does Didier's "component spray tool" already do this. (If not, it could still be used to spray paint with the low-poly version of the component)

Is there another script out there which does this already?

Thanks TBD.

However, there is a method to upload multiple .SKP files all at once rather than one at a time.

You create a Spreadsheet of the files, with other information, Description, Tags, etc. and then upload the whole mess at one time. Hopefully creating collections and sub-collections as you go.

This is supposed to be part of the SketchUp SDK. But I am having a hard time tracing it down.

I want to upload about 500 trees and plants to the 3D Warehouse. Thesse will be always-face-camera alpha image components.

I believe there are some tools in the SketchUp SDK to upload multiple models from a Spreadsheet. I have looked for them before and will look harder.

Cab anyone guide me to these tools, or offer any other advice on how to upload multiple models to the 3D Warehouse?

RPS 3D PDF uses an adobe 3D features - U3D.

We are able to place 3D faces into U3D, but have not been able to figure out how to place text and dimensioning.

Be sure to include solid materials (such as wood below - which automatically molds to the curved surfaces), and fractal plants - which create random, complex, detail without clogging the SketchUp model in the criteria for a good renderer.

(Just kidding of course - but it is an example of how a renderer comparison could be slanted to favor a particular rendering engine)

One of the nXt support guys is working with this model with the latest version. He is trying to get more subtle lighting under the awning, using the sun, sky and cloudiness intensity settings.

See the nXt forum: http://nxt.accurender.com/forums/t/817.aspx