Big project: I need advice

[dmoranda]

I have just landed a big project and am in need of some direction. It is a proposed residence which requires multiple visualizations. 5 Shots to be exact. I have been given the photographs, all the proposed CAD drawings in CAD ADT format, and a couple of Architects hand renderings for look and color composition. I need to match up the proposed flagging in the shots I have attached: 2 Panoramas and 3 angled shots.

 

[ATTACH]37769[/ATTACH][ATTACH]37770[/ATTACH][ATTACH]37771[/ATTACH][ATTACH]37772[/ATTACH][ATTACH]37773[/ATTACH]

 

In a nutshell, I need to build a complete model of the proposed home in 3D, Camera Match and then integrate the rendered shots into the 5 photographs. I need to really optimize my time on this project. The time frame is roughly 2 to 3 weeks. I know, its not much....

 

What is my first move on this project? I realize I need to find out exactly what areas I need to model in detail according to the shots I've been given. I don't want to detail a particular part of the model that is not even showing up in the shot.

 

Do I create a very primitive 3D block out first? to determine exactly what areas to model in detail? Obviously I've got 2 "Elevation" shots of the front and rear of the house that need to be modeled in high detail (2 panorama shots), but the 3 angled shots hide quite a bit with the trees and fences etc. (other 3 shots)

 

Camera matching:

I am a beginner in this area. These provided photographs don't give me a lot of horizon lines / x and y lines to match up. Should I go out and re-shoot the photos with some objects on the lot? I read somewhere on this forum to place a cardboard box in the shot for match lines, then take an identical shot with the box removed. Is this the best workflow?

 

Honestly, I am feeling a little overwhelmed. This is by far the biggest project I have ever landed and I need a little guidance. I have extracted so much knowledge from these forums. I know this is the best place to ask.

 

I'd appreciate any input on this project. Workflow, programs to aid in the development of the project. I use primarily 3DS Max, and sketchup. I apologize if this post is rushed. I can explain any part of the project in detail if asked.

 

Thanks in advance,

Dustin

[dmoranda]

[ATTACH]37774[/ATTACH][ATTACH]37775[/ATTACH]

 

Inserting Architects renderings.

 

Thanks again,

Dustin

[ihabkal]

2 to 3 weeks? you're so lucky, I usually have 2 to 3 days per project.

[bjornkn]

First of all, you're never gonna get any camera matching to work with those panoramas, becasue they have a lot of different vanishing points all over the place. IMO you can't use them at all unless they are better blended. Is this what you got from your client?

I use SketchUp and Lightwave BTW.

For such a project I would start with a site plan, if available.

Model very rough models of the existing buildings in the area. Footprint is most important. I typically make "skyscrapers" even when the buildings are small, because the walls are what I use for aligning.

Then either use PhotoMatch in SU, or align by hand, aligning the exisiting walls with the photo.

If you have access to the area I would definitely go there and shoot some more photos.

[1d2d3d4d]

everything bjornkn said-

 

the site plan should be your guide ...

build a basic model- model the site,

 

take more photos if you can-

*if there is a possibility to go to the site and set stakes or markers that would be a huge

help for accuracy... ( i have yet to do this but wish i did everytime I camera match-I wonder if anyone else does this)

 

but dont get too anxious about it-

it looks like there are so many trees in some shots that you will probably be photoshopping so much of the actual building....

[John Dollus]

yup. when I do site photography, I like to take along a couple camera tripods with tennis balls on them along with a laser tape for quick site survey measurements.

It sounds like you have to just work with the photos the client sent you though. Hopefully, they sent you the images for the panos unstitched

for those, just use one photo for the match and then composite in the rest of the canvas as neeed.

As for modeling, I would just model the house according to the dwgs. Trying to eliminate windows based on trees in the photo is probably not worth the time savings in the long run.

for workflow, i'd suggest modeling the house first and sending it in just a b/w AO pass from the rough angles to get the designer to sign off on it first. while they ponder that, you can work on matching it to the photos so hopefully by the time they get back to you, the cameras are already lined up. Likely, the designer will have a couple changes (they always do) so you can then send him another set of drafts after you have those modeling changes done with the model overlayed on top of the photos (but no photoshop work yet) so the designer can then signoff on both the model with the cameras setup. Ideally, you would have also lined up the sun and have some materials defined to they can signoff on that.

If you are lucky, by now the client has pretty much approved everything leading up to getting the renderings integrated into the shots so you can focus on masking/covering/hiding etc .

That said, depending on the quality expectation, you should be working on it instead of reading this post - that's a lot for one person to tackle

[dmoranda]

First of all, you're never gonna get any camera matching to work with those panoramas, becasue they have a lot of different vanishing points all over the place. IMO you can't use them at all unless they are better blended. Is this what you got from your client?

I use SketchUp and Lightwave BTW.

For such a project I would start with a site plan, if available.

Model very rough models of the existing buildings in the area. Footprint is most important. I typically make "skyscrapers" even when the buildings are small, because the walls are what I use for aligning.

Then either use PhotoMatch in SU, or align by hand, aligning the exisiting walls with the photo.

If you have access to the area I would definitely go there and shoot some more photos.

 

Bjorn,

Thanks very much for the input. As you said, I will start creating the site and very primitive models to represent the house and other buildings on the street. The existing house in the photographs is going to be demolished and a new construction built in place. The site plan has the existing house on it. So I guess I can use that as a matching point? After I have this very basic model of the house, should I then drop it into the photo in photoshop? Is there a way to link files in PS so that I can update the file as the model progresses? I just don't want to have to repeat any of the steps. (although I'm sure I will have to

 

The Panoramas were given to me by the architect, and yes unfortunately they are not much to work with for camera matching. The site is within a short driving distance and I have scheduled a time to go out and take some better shots. The clients do want the shots in a panorama format. Any advice on taking better Panorama shots? I'm using a Canon DSLR. And should the markers be in any specific locations? like a marker from the surveyor?

 

Another question regarding sketchup. Do you typically clean up the CAD dwg files and then import them into sketchup as a reference? I am using Autocad Architecture and I was thinking about using its 3D capabilities, and then porting it over into sketchup or max. Although I think Sketchup would be a much more clean file. Easier on texturing as well. Any input?

 

1d2d3d4d,

Thanks for the tips and the encouragement. I'm sure I am making this out to be much more work than it really is.

 

Thanks a million guys, I'm gonna get to work now.

[dmoranda]

yup. when I do site photography, I like to take along a couple camera tripods with tennis balls on them along with a laser tape for quick site survey measurements.

It sounds like you have to just work with the photos the client sent you though. Hopefully, they sent you the images for the panos unstitched

for those, just use one photo for the match and then composite in the rest of the canvas as neeed.

As for modeling, I would just model the house according to the dwgs. Trying to eliminate windows based on trees in the photo is probably not worth the time savings in the long run.

for workflow, i'd suggest modeling the house first and sending it in just a b/w AO pass from the rough angles to get the designer to sign off on it first. while they ponder that, you can work on matching it to the photos so hopefully by the time they get back to you, the cameras are already lined up. Likely, the designer will have a couple changes (they always do) so you can then send him another set of drafts after you have those modeling changes done with the model overlayed on top of the photos (but no photoshop work yet) so the designer can then signoff on both the model with the cameras setup. Ideally, you would have also lined up the sun and have some materials defined to they can signoff on that.

If you are lucky, by now the client has pretty much approved everything leading up to getting the renderings integrated into the shots so you can focus on masking/covering/hiding etc .

 

 

Wow great stuff John! Thanks so much for your post. Now I'm a little torn. Some say shoot for the site first, and you're saying jump in head first on the model and get the geometry approved. I will have a chance to go out and take better shots so I will use your techniques regarding the tennis balls on tripods. They did send me the photos from the Panorama unstitched so I should be able to mend them a bit. It will have to span that far unfortunately though. They need to see the whole block for potential view impacts. Thank you so much for laying out your approval / workflow process. Much appreciated.

 

That said, depending on the quality expectation, you should be working on it instead of reading this post - that's a lot for one person to tackle

 

Tell me about it! I'm off now I promise.

[bjornkn]

Bjorn,

Thanks very much for the input. As you said, I will start creating the site and very primitive models to represent the house and other buildings on the street. The existing house in the photographs is going to be demolished and a new construction built in place. The site plan has the existing house on it. So I guess I can use that as a matching point? After I have this very basic model of the house, should I then drop it into the photo in photoshop? Is there a way to link files in PS so that I can update the file as the model progresses? I just don't want to have to repeat any of the steps. (although I'm sure I will have to

Your project consists of two separate jobs, modeling the building and compositing it into the photos.

I would model the actual building in a separate scene file in SU, and then model the site with exisiting buildings etc on sparate layers. Then I would load the other file (building) as a component into a new Layer and place it according to the site plan and exisiting builidings. Then you can easily turn on/off the various layers. No need to see the new building while matching the camera(s).

And you can easily change the building in the buildingl scene file and reload it into the site scene.

 

I wouldn't do anything in PS until after rendering (in LW in my case).

In PS I usually copy the background and add a black layer mask to it.

Then I load the rendered building (with alpha) inbetween that front layer and the background.

Then I can just paint on the mask to make foreground parts visible.

New renders can easily be slipped into the sandwich.

 

The Panoramas were given to me by the architect, and yes unfortunately they are not much to work with for camera matching. The site is within a short driving distance and I have scheduled a time to go out and take some better shots. The clients do want the shots in a panorama format. Any advice on taking better Panorama shots? I'm using a Canon DSLR. And should the markers be in any specific locations? like a marker from the surveyor?

A normal, rectilinear panorama will start to look very strange, stretched and ugly once you pass around 120 degrees fov. Your panoramas are not rectilinear, and they must cover at least 160 degrees fov? You might try with cylindrical projection? Or spherical? Or even some more esoteric like Pannini? What architects usually do when drawing a street block is to draw orthographic, which isn't easy to do with a camera, and in particular not easy when you have lots of tree and other details at different distances from the camera. Take a look at some of my ortho/linear panorama photos at http://www.bknilssen.no/fasader/ . A lot of PS work was involved, and it wouldn't really be possible with your streets I think? But then you could use an orthographic camera in SU, LW or Max.

Standard procedure for shooting a normal panorama is to use a tripod with a panorama head that allows you to rotate around the lens' NoParallaxPoint (NPP). Then you could stitch tem using PTgui or some other panorama software where you can choose your projection etc.

 

Another question regarding sketchup. Do you typically clean up the CAD dwg files and then import them into sketchup as a reference? I am using Autocad Architecture and I was thinking about using its 3D capabilities, and then porting it over into sketchup or max. Although I think Sketchup would be a much more clean file. Easier on texturing as well. Any input?

It depends on the DWGs. If they are very cluttered with lots of small details that I don't need, and makes SU sluggish, I often clean them up inside SU. Then I group different parts and make some new layers where I put them. Next I build a "cage" from plans and elevation drawings and start modelling the building inside that cage. UV mapping and texturing is very easy inside SU, but there are no other projections than planar, and no other parameters than color. It's usually a very good start though

[dmoranda]

Well, I haven't really had time to post because I have been jamming on this project. Here is an update for a rough approval of positioning from the architects.

 

Thanks to everyone for the input and advice. I have used it thoroughly! Still a ways to go on these. Next step will be texturing / lighting / rendering / compositing. I just wanted to give an update. All C&C is greatly appreciated. Although its still a ways to go now.

 

Thanks,

Dustin

[bjornkn]

Well, if the background images are to be those mosaic linear panoramas at least you don't have to worry much about camera matching...

As long as you put the house at the right altitude and scale it's gonna look more "correct" than the background ;-)

 

I'd say that maybe the next step should be to go there and shoot some new and better panoramas - unless that's really how your client wants them to look?

[dmoranda]

Yeah unfortunately they want to use the photos they have shot. It does look rather amateur how they are spliced together. I did have to do 3 shots that were in perspective and required camera matching. I am not too versed in this. It ended up being several hours of sitting down with the architect and positioning them according to his wants. I am happy with the volume and perspective of the shots. I've posted them here.

 

Now that I have my cameras set up and locked I am now going to go in and texture and light the scene. I'll be using vray for sketchup. I pulled the model into max, but it looks pretty tesselated from the transfer. Any ideas on this? I am running 3dsmax 2011 and the newest version of sketchup pro. I'll be posting some of the renders as I get them, and I would really appreciate some critiques and comments for improving them. Thanks so much for the help from everyone. Cheers!

[bjornkn]

If they get tesselated you may need to change smoothing angle?

You may also need to merge points, because the 3ds format apparently denies uv mapped meshes to share vertices, and thus you end up with a lot of separated tris that needs to be joined to render smooth.

[dmoranda]

Here is a quick update. I need critiques! The architect keeps trying to have me make the textures darker and it just seems to be getting further away from realistic on the lighting side. I guess he is wanting the house to be subtle and blend with the landscape. If it gets any darker its going to dissapear! Working on the other angles as we speak. Any C&C would be much appreciated.

 

Bjorn I decided to stay using sketchup and vray as my renderer. I was having some issues importing and sketchup seems to still be doing a great job.

 

Thanks for the input, I appreciate it!

[bjornkn]

I would've used a longer focal length, and hence less perspective. I would also move the camera to the left so that the side wall is no longer visible. I think that will make it blend better into the strange ortho/linear pano background. Unless it's supposed to not be aligned to the road?