TIG, true, white fog can work in many cases. For my elevations I want to keep a blue sky, and if I use a white fog my sky disappears. If I use a blue fog to keep my sky, I get a blue color cast on my elevations.

Another technique I use is to create a rendering of my building as a png with sky and ground off so I can get a mask of the building itself. Then I can drop a sky in with Photoshop using the profile mask made from the png; but the reason I favor rendering out a fog vs. building mask is that I can get more control over the effects of depth and can also use curves or levels in Photoshop to manipulate the distance falloff effect.

+1 on interest in learning about adding lights with IES in Shaderlight.

Couldn't attach two files, Here is the sample elevation using fog to fade building elements further away from the camera. Pretty subtle effect but still visible.

So here is a quick mockup. I set fog to black, play with fog sliders to get full grayscale range over depth of model. I also show shadows as their own exported image. I do this on my elevations so that I can set "use sun for shading" to off and get consistent materials colors from one elevation to the next, and then overlay shadows in another photoshop layer (set to multiply" The fog layer is used as a mask to mix two variations of the basic fog and shadow free rendering. You could alter saturation, contrast, and blur in based on depth, whatever you want. You can also use curves in Photoshop to make the changes less linear.

The last image is a practical example of this used for an old elevation of mine. The effect is more subtle than the mockup, but a big time saver when quickly making changes and not wanting to re-tweak Photshop files.

One method I have used it to create a duplicate copy of a scene with a style set to hidden line so faces are all white, and lines are near white (turning them off created issues). I then turn fog on and set it to black.

This scene when exported serves as a depth mask for your original scene. You can use it is an image mask in Photoshop to blend two source renderings. This can be used to alter contrast, brightness or other characteristics of the model as you get further away from the camera.

Could be in 90's or there could be a foot of snow on the ground, that is what makes spring and fall so much fun here. Best guess is mid 80's in the day, mid 50's at night, but either of those could be 20 degrees off either way. (all temps Fahrenheit). It won't be humid.

David

We are having some problems getting the transformation correct when adding a child group to a parent group when that parent group is active. If the parent group isn't active then it works fine.

Has anyone else had this problem. Details and technique to duplicate are below.

Our workflow is that the user adds child groups to a parent group while the parent is open. The script below shows my problem, you have to enter it line by line in the ruby console because there are steps (Steps A and B) where you have to click around to change the current active group. I have tried calling make unique on both groups to no avail and can’t even do a work around of setting the child’s transformation to the parent’s inverse transformation because when I try to get the parent’s transformation it comes back as identity. I would really appreciate any help in finding a work around for this, I don’t want to have to go to components instead of groups unless I really have to.

Script is below:

# clear everything out
Sketchup.active_model.entities.clear!

# add the parent group
group1 = Sketchup.active_model.entities.add_group

# didn’t work
# group1.make_unique

# set parent’s transformation
group1.transformation = Geom;;Transformation.translation(Geom;;Point3d.new(10.m, 0, 0))

# add some points to the parent group
group1.entities.add_cpoint(Geom;;Point3d.new(0.m, 0.m, 0.m)) 
group1.entities.add_cpoint(Geom;;Point3d.new(5.m, 5.m, 3.m))

# the parent group’s transformation shows the correct translation
puts group1.transformation.to_a.join(', ')

# step A) double click group1 to open it

# parent’s transformation now shows as identity?
puts group1.transformation.to_a.join(', ')

# add the child group
group2 = group1.entities.add_group

# didn’t work
# group2.make_unique

# we want the child group’s axes to be the same as the parent group’s axes, unfortunately this seems to put it at the absolute origin
group2.transformation = Geom;;Transformation.new 

# add some points to the child group
group2.entities.add_cpoint(Geom;;Point3d.new(0.m, 0.m, 1.m)) 
group2.entities.add_cpoint(Geom;;Point3d.new(5.m, 5.m, 4.m))

# step B) exit out of all groups to top level model

# the parent group’s transformation is correct
puts group1.transformation.to_a.join(', ')

# the child group has a negative translation to bring it back to the origin?  Why?
puts group2.transformation.to_a.join(', ')

# this will put group2 where it is supposed to be
group2.transformation = Geom;;Transformation.new 

# group transformations are now correct
puts group1.transformation.to_a.join(', ')
puts group2.transformation.to_a.join(', ')

The simple answer to only use MatchPhoto on images you think are not cropped.

In your situation with a cropped photo I see three options.

1. Have someone who physically can stand in front of that bed take a photo for you
2. Have someone with access to that bed or spec sheet for you measure the bed.
3. Just guess on some measurements and model from scratch (how accurate it needs to be depends upon what you want to sue it for. But if you make the assumption that it uses a standard twin mattress then you should be pretty close.

Chuck,

Looks like a nice app. A really interesting feature would be to use the geotagging of the 3DWH models along with the geo-location and the compass of the phone to automatically identify 3DWH models that are close to you. In theory using the compass you can focus on models close in a specific direction, vs. all around.

If everything was very accurate you could even mimic the actual viewpoint, but I expect that data and geolocation isn't there yet.

I would just make some assumptions about the dimensions and model it from scratch.

But if your interested in learning one approach to uncrop photos, here is an old link of mine. Your photo doesn't have a lot of vertical vanishing, so won't work so well for this.
http://www.sketchucation.com/forums/viewtopic.php?f=15&t=10895&hilit=+crop&sid=5ba75feb83dea26a56080dbab7c89e97

I want to see this technique on CSI show so they can figure out that the murder was cropped out of a photo, and then they can find the negative to identify him. Actually when I first search fro techniques to uncrop photos, or look for other alterations (e.g. person dropped in from elsewhere); one interesting technique looked at distortion of the shape of the eye to look for flaws.

By the way, cropping is ok if it is symmetrical, but then again if you didn't crop it how would you know. Some camera sensors (probably worse in cell phones) are not in line with the center of the lens, resulting in slightly cropped images right out of the camera. (while not truly cropped, the optical center of the scene isn't the geometric center of the image (as if it had been cropped).

@gaieus said:

I have another one that I am not allowed to wear, too.

Turn it inside out

@solo said:

David, got any good places to hang out at night?

There are number of places on Pearl Street or Walnut in downtown area (10th-15th street). Another area not far away is University Hill by CU campus. CU classes start on 8/23 so I expect they will all around by the hill. I'm not as familiar with night places as breakfast places since I my idea of hanging out at night is laying on the couch once I get the kids to sleep.

But if I had to name a few
The Sink has a lot of character by campus. (http://www.thesink.com/)
Conor Oneill's is nice place downtown (http://www.conoroneills.com/boulder/)
Rio (http://www.riograndemexican.com/)
But lots of other places around as well, these are just a few that popped into my head. Some coffee shops are restaurants are also open pretty late.

Thanks, I guess this should be on the "Duh" thread. Now I remember running across this when I first learned ruby.

What I was looking for would let me click on an edge, or point on a face and get the x,y and z distance from the origin. I thought there was a plugin like this, but didn't find it on a quick look through Smustard.

@mariocha said:

Hum.. I wonder why they are asking about our arrival time.
Group pickUp at the airport ?

This was on the main Basecamp site, maybe they are wondering when to have them.
"Outdoor-themed activities, such as a walking tour of Boulder or an early morning hike in the Rockies!"

By the way, unrelated to this, here is my list of breakfast restaurants
http://www.luciles.com/ (taste of New Orleans)
http://www.chautauqua.com/dining_hall.htm (most scenic)
http://www.boulderteahouse.com/ (best tea)
http://www.walnutcafe.com/ (basic breakfast)
The Best Western is also supposed to have a really good place right there.

Tons of others, but these are some of the highlights.

Nothing to do with milk anymore. An old dairy plant converted to an arts center. By the way, plenty of good places around town for those looking for good breakfast before Thursday's and Friday's BaseCamp.

What I thought was funny about the agenda. Is that we move from Breakfast/registration, right into lunch. Will be some full belly's when the keynote starts at 1pm.

So I know this is scheduled from 9-10am on Sept 2nd, but was wondering if anyone else thought it would be a good idea to open up a thread or threads for people interested in giving an un-session to describe what they want to talk about. I wouldn't suggest any pre-voting. Intent is just to give people an idea ahead of time of what they can expect to hear about. possible speakers can also get feedback on what subjects attendees are interested in. This could all be in one long thread in this forum, or a series of threads each with "Unconference" or some common word in the subject of their thread. Of course many attendees aren't on Sketchucation so there will still be a lot of surprises, but I thought it still might be helpful. I should disclose that I will probably attempt to give a session on OpenStudio and Energy modeling.

David

Happy to help, looking forward to meeting everyone.

My inlaws may still be in town for son's 2nd birthday ( a few days earlier). So if I don't make it for dinner will definitely come by later for drinks.

David

Updated after more testing:

So when I try this code with your plugin and SKP file the dimension is made, but often is nowhere near the edge it that it should be. My initial testing shows this working for edges starting or finishing at the origin. but many other edges produce dimensions far off in space. I'll look at the code and see what I can find out. The other thing this doesn't do that I want it to is to have the dimension associated with the edge, so that it scales with the geometry. This acts like an orphaned dimension vs. one that is linked to an entity.

@tig said:

It is not a plugin that auto-dimensions.
It is code that adds extra methods to the API, so if you have some other code you can add dimensions with it - for example...

model=Sketchup.active_model
> ss=model.selection
> edges=[]
> ss.each{|e|edges<< e if e.class==Sketchup;;Edge}
> ents=model.active_entities
> edges.each{|e|ents.add_linear_dimension(e.start.position, e.end.position, 4.0, 0.0)}

This would loop through all edges in the selection and add linear dimension to them offset 4" and at 0 degrees rotation about the edge...

I only supplied the code as an example to show what could be done
You need to work out how and which things get dimensioned and in what format...
Then code it...
There are some notes at the start of the ruby regarding its use too, that you probably read already......

Thanks! I would have searched a while for that.