@unknownuser said:

BTW, in his guide Adam warns about LU being capable of corrupting a model. The LightUp studio file I used with the above attempt to export had something happen to it, the model just completely disappeared after that little test run, so make sure you have a backup of the file as suggested by Adam.

I have to say I personally have never had a corrupt model from using LightUp. The only issue I've had - which sounds like what you've seen - is the model looking like its disappeared but its just the Layer is disabled.

I'm certain your model disappearing is nothing more than the Layer being hidden.

Adam

@unknownuser said:

You know the demo licence only works for 30 minutes at a time then SU locks up.

At the risk of being a pedant, it doesn't "Lock up" Sketchup after 30 minutes. It times out and stops working but you can continue using SU as normal without LightUp for as long as you like.

Adam

@jorgensen said:

I have tree quite simple questions i guess

How do you walk around like in the Sponza building ?
When I try to walk I loose the lightning

If you start the Tourtool and switch on Collision and Gravity you will be able to walk around buildings etc. It isn't the Sketchup Walk tool.

@jorgensen said:

Is the light included when one export as AVI ?

LightUp completely replaces the Sketchup rendering so functions that save screenshot of Sketchup won't work. LightUp currently only supports saving high resolution screenshots that can be sampled down to a useable size in a PS, PaintShop Pro etc.

Exporting movies from LightUp sounds like a useful feature for a future version though.

Adam

@putnik said:

AdamB, looks very interesting! Does it work in SketchUp 5? I couldn't find this on your web site (perhaps wasn't looking hard enought).

Thanks.

No LightUp only supports SketchUp 6 (and higher!) because to integrate with SU nicely it uses WebDialogs and Observers.

Adam

@jim said:

Thanks for the help.

Jan, I haven't looked at your code yet, but I plan to.

Adam, I tried what you said (at least I think I did,) but ultimately, this is what worked for me: (assume a single Group is selected)

> m = Sketchup.active_model
> e = m.active_entities
> s = m.selection
> 
> trs = Geom;;Transformation.scaling( 2, 1, 1 )
> s[0].transformation *= trs
> 

Which seems counter to what you said?

It really depends on what the semantics of their * operator is. ie does it pre- or post- concatenate the matrix - but glad you're all sorted either way!

Can't ignore an opening like that!

Yes, LightUp v1.1c is on the server and has many fixes and improvements. Come and get it, I think you'll be pleased with the improvements.

Here a little video of the Sponza building processed by LightUp http://www.youtube.com/watch?v=KANx9iUqiWM which I thought was pretty neat even if I say so myself.

Adam

@jim said:

So apparently applying the following scaling transformation to a Group scales the Group along the world axes, and not the Group's axes. (Shown in wireframe to show Group axes.)

It just seems scaling along the Group's axes would have been the expected result.

> tr = Geom;;Transformation.scaling( 2, 1, 1 )
> group.transform! tr
> 

[attachment=0:r2ix4t33]<!-- ia0 -->Clipboard01.jpg<!-- ia0 -->[/attachment:r2ix4t33]

So I need to ask, how do I go about getting the Group to scale 2x along its X axis, regardless of its rotation?

In answer to my own question, the only solution I found is to create a new transformation, apply the scaling first, then apply the position and rotation transformations of the original Group. There are other ways to order the transformations, but the scaling must be done before the rotation.

Jim,

Multiplying 2 matrices together is non-commutative (unlike with regular numbers). ie A * B is not the same as B * A for matrices.

The transform method does group_matrix * your_matrix , what you want is your_matrix * group_matrix
So just all you need to do is get the group matrix out multiply it with yours and set it back on the group to ensure you get the order you expect.

Adam

@remus said:

Adam, could you not just do it with a small number of rays? If you only want a quick visual representation of what sound is going to be doing in the room i dont think youd need to many waves to get a pretty good idea of whats going on.

You could perhaps use a slider to control the density of waves being emitted from a surface and then set the wavelength.

Its not the number of rays. Its because a sound wave is diffracted by the furniture etc. I just googled and found this link http://www.acoustics.salford.ac.uk/feschools/waves/diffract.htm which might help explain.

Adam

@rdesign said:

Well- we know this is possible in other applications: <http://squ1.com/products/ecotect/features/acoustics>

Tim

I just had a look at that link and didn't read it all through but at first blush that product isn't useful for audio in my book. I mean, if I stand around the corner from you and say "hello" you hear it because of the waves 'going round corners' due largely to diffraction.

So yes, if you just want to treat sound as.. light, then yep shoot some rays. Its just you'll get an answer that doesn't match up to reality much!

Adam

I've implemented this before. I think it has lots of potential to provide properly occluded audio sources. You can go pretty far by attenuating/filtering audio sources based on how much 'stuff' there is between the listener and the source in a straight line way. eg you can highpass filter sound that is separated from the listener by a thick concrete wall etc etc. ( I simply raycasted from listener to source and from source to listener, got what material I hit at both ends, made the presumption it was solid 'stuff' between the two faces and did some filtering.)

But yer problems start when you look at transport of sound from reflections. Raytracing something with a wavelength that is tiny compared to the room features (eg lightwaves) is easy because you can largely ignore diffraction around corners.

Audio wavelengths is a different matter entirely because the wavelength of sound is much larger than light so that diffraction (around furniture etc) is integral to simulating audio transport. It all gets a bit scary on the computation side at the point...

Adam

OK, got it. [And this is all thinking out loud ]

So the idea is that if developers implement some well-known reflection methods, their plug-in will be able to work in the new whizzy plug-in management GUI that gives users a better experience, right?

What happens if they don't implement those methods? I guess the plug-in is ignored wrt the PluginManager?

You mentioned some ideas for what features could be offered which sound cool - might be good to have the notion of 'Favorites' to avoid searching through dozens of icons each time. Perhaps even some gesture pop-up (radial layout of icons?)

One thing I don't understand (and its probably lack of coffee!) is how this PluginManager discovers the plugins. ??

Lastly, I'd like to encourage more people to contribute to this discussion. If you've got suggestions and ideas, then post them; the worse that can happen is they're shot down in flames and. ^h^h^h.. Just kidding! No, really SketchUp users out there, lets have your ideas as to how you'd like to manage lots of little plugins.

Adam

SU has a language built-in for a similar purpose. Its called Ruby.

At the console you can execute Ruby commands to create geometry etc.

eg

Sketchup.active_model.entities.add_line([0,0,0], [3,1,4])

adds a line to your model.

Sketchup.active_model.entities.add_text "Allez Les Bleus", [5,6,7]

adds some "camera-facing" text at coordinates 5,6,7

etc etc

http://download.sketchup.com/OnlineDoc/gsu6_ruby/Docs/ruby-classes.html for every function.

Adam

Jim,

I guess you've lost me a little already! What exactly is the problem that we want to solve? I know there are few rules wrt plug-in integration, but it would be useful to articulate the precise issue that would be solved by any additional APIs.

Having a reflection system / protocol is all neat but how exactly does that help the end-user?

Secondly, IMHO if you are going to provide some infrastructure then you have to provide the developers a compelling reason to use it. It has to give them something that they would otherwise not have. If you don't, then most developers will take the view that its quicker and easier to 'do their own thing'.

So I'm a developer, what do I gain for my Tool/Plugin by implementing some standard reflection methods?

Adam

@adrw2001 said:

I've been having problems with SketchUp crashing quite frequently... It doesn't give any error messages, just plain quits..

I'm running it on a new MacBook Pro, 2.4 GHz Intel Core 2 Duo, 4GB 667 MHz DDR2 SDRAM... should be plenty of machine to support SketchUp without issues. any settings I need to change to improve stability?

On a Mac, any application that quits like that will write a log file to:

your_home_directory/Library/Logs/CrashReporter/Sketchup.crash.log

Adam

I'd be up for adding to that where I can. I think getting the right level of standardization is important - not too brittle or rigid - from pov of adoption. And not using language features for their own sake too.. But in the same breath I'd suggest that in a language as dynamic as Ruby, protocols over classes is the best bet. eg I really like the fact that the Tools are just objects that implement a known protocol and not some class derivation.

I also like TBD's idea of knowledge base for SU ruby. However, from a fair amount of experience I know its very hard to keep these things from drifting off into the long grass - its completely possible, but it requires work, regular work and perhaps more work than any of us are realistically going to be able to offer. But having somewhere to capture that - "I just wasted an afternoon rediscovering the same bug/glitch/oversight I wasted an afternoon from N years back and this is what it was" - could be a valuable resource. Catchy title too.

Adam

@zootsuit said:

Hi there.

I'm working on a large scale residential estate for an architectural project. I have a PDF of the plan site which I want to use as a base to trace over in Sketchup.

I have converted the PDF to a jpeg in Photoshop then imported it and used it as a material on a base rectangle. The problem is that it needs to be scaled up so much that the image is very blurry and hard to trace accurately. Is there a better way I can do this ? Should I model on a much smaller scale , i.e the original size of the jpeg, then scale the model to size after? If I try to scale the Jpeg up in Photoshop the file size goes through the roof!

Many Thanks

Zootsuit

If you had an EPS (PostScript) version it might be possible to write a Ruby to convert it to geometry that would scale perfectly.

Adam

@remus said:

Adam, what makes you think inferencing does take much computing power? i would have tohugh its quite computationally intensive to find the mid, end and direction of very line in an SU model.

It would, so you don't do that!

Typically one creates some spatial structure that can narrow down the potential candidates very quickly. LightUp uses kd-trees to do it, I suspect Mr.Google will have something similar. The end result is you actually find mid, end and direction for only handful of lines.

Adam

@unknownuser said:

Don't know about that...but clearly the graphics is not the most important bottleneck in current SU I believe.

I think you're broadly right. There seems to be a lot of work going on under the hood to get to the point of drawing (which incidentally is why the OpenGL/DX question is a bit of a red herring). BTW I can't believe inference is more than a few percent of frametime so I'd discount that too.

However, if you have complex models it does make a BIG difference in refresh rate to not draw shadows and not draw (thick) profile edges.

Adam

@richard said:

@jim said:

The biggest challenge is identifying the geometry for a leaf. My first thought was that if the leaves all have a surface area that is similar, it could make finding them a lot easier - you could find them by area.

Can you post a sample skp of an imported tree?

Maybe we can get you half-way there, if not all the way.

Thanks for responding jim!

I can't load up an onyx tree due to the user agreement, though the same effect would be buy grabbing any SU tree with component leaves and exploding down to base geometry.

I would imagine the script would either search for alike areas or simply alike geometry. The script would only need to locate one similar face and then select all connected and replace!

I would imagine the script wouldn't be so hard to do but could be a great time saver where ever a model contains numerous instances of the same geometry.

My 0.02 of your earth pence worth...

Area might be a good key, you might also want to group connected geometry, find its OBB (oriented bounding box) and then match on that. And of course then you'd have the necessary local transform "for free" too!

The biggest problem is that processing geometry in Ruby is spectacularly slow and if you're going to be using dense geometry you might be waiting a long time.. Oh but you do raytracing don't you, so you're used to that.

Adam

@kwistenbiebel said:

That's cool....
I didn't even know you can make physical models with Sketchup.

Would that method (Stl) work as good for architectural scale models?

In principal yes, in practice no. It would be a pretty big (read: scary) milling machine!

Adam