MOC: Mini City

[alienwar9]

Phew, It's finally done (No, not the actual city )

It's the mini 1:4 scale of my city that I promised to make

Whole LXF WARNING:32,895 pieces

And for those with less powerful computers, here is the city split into sections:

Bottom Half LXF ~9,000 pieces

Middle LXF ~14,000 pieces

Top Right LXF ~8,000 pieces

And for those who want to test (or see the river together), the upper city (section 2+3) together:

Upper City LXF ~23,000 pieces

Once again, more work, more time, and more pieces than I expected

But I finally get to see the scale and complete look of the city in LDD before the real version, and you do too. I tried my best to keep the scale and proportions correct to get the overall feel right. For the most part everything worked out (a few buildings are out of scale )

Here are a few more pictures of key views I liked:

THE standard view

view from the opposite side

a sky view

one of my favorite parts, the bridge

1st person view looking up

And finally, my necessary overhead google maps view

Hope you all like

8 more pictures on my brickshelf

[JopieK]

like it, man... you must be kidding, it is just great, or really sick, or I guess... both :)

[Brickdoctor]

Only alienwar could make a city so big that even at a fourth of minifig scale, you still can't tell how big it is from the first picture.

[Abu Dun]

Even at 1/4th of your actual project, the size of that city is awe inspiring. I can't even begin to imagine how much time you needed to put this together, especially since you included every little detail of its larger relative.

Great work and I can't wait to see the final result of the actual town

[Superkalle]

Alienware, I was wondering about assembling those big files.

Here's how I do it, are we doing it the same way?

The case is that I've made a couple of submodels of about 15.000 bricks, that I need to compile into a total 50.000 brick file. Naturally LDD becomes incredibly slow even though. Then I import the various submodels (waiting a minute or two before they appear). Then it's the matter of attaching the sub-models in the correct place. I normaly select a 1x1 brick in the sub-model and move the mouse quickly to the correct stud on the master model. Then I wait a minute and hopefully they will have snapped together.

Is this how you do it too?

And also, what is the largest LDD file (number of bricks) you've managed to open and use in LDD?

[Calabar]

I normally select a 1x1 brick in the sub-model and move the mouse quickly to the correct stud on the master model. Then I wait a minute and hopefully they will have snapped together.

I usually hide all the scene except those bricks I need to connect the imported model (often a baseplate and few bricks), and, if necessary, I hide part of the imported model too.

So the scene became much more light to manage and I've not difficulties to connect it.

[Horry]

Man, that is awesome! The level of detail is incredible. You've got to have square-eyes now!

Edited October 9, 2011 by Horry

[eiker86]

Its. Magnificent... This is really good. When are u planning to buld it?

[alienwar9]

I normally select a 1x1 brick in the sub-model and move the mouse quickly to the correct stud on the master model.

And also, what is the largest LDD file (number of bricks) you've managed to open and use in LDD?

I do as Calabar does it mostly. Hiding the model part you are moving I find is pretty useless since LDD shows the whole thing when you move it anyway. But I hide EVERYTHING else except for 1 piece, a reference piece that the imported model will be attached to. That way LDD doesn't have to calculate connections for anything else (actually, this is useful whenever building a complex model, since many times LDD starts to get confused where you want to place a piece or parts. Sometimes I even place a baseplate on the bottom so LDD has only one surface to place onto). The best reference piece will always be a baseplate that is on "ground" level, since you can't place anything underneath it.

The tricky part is getting the exact location to move a model. When I separate a model, I leave a plate that stands out from the rest of the model and is attached to a baseplate in a corner. The plate goes with one part, the baseplate with the other. That way when I have to reconnect it, it will be very simple to find the connection point (since I often put huge chunks of cities together and its hard to even tell where exactly they go together )

Since LDD can take a while to load the movement, getting the position right the first time is important. So the with the imported model, I select a corner of a corner piece (or that extra plate I mentioned). This means that the cursor will barely be touching the piece (wait until you get that purple high-light) and will be a respective corner to the reference piece that you will be attaching it to (so bottom left corner to bottom left of reference piece, etc). Sometimes you will end up deselecting the whole model, but this is better than moving the model in the wrong place, so its worth it. As soon as you have "grabbed" the corner, your method of quickly moving is the next step . But you have to make sure you know where you are moving to.

This means, that the final location of your cursor will have to be perfect (at least when I've moved big models, it has to be). Even a millimeter of mouse movement throws it off. What I have found is that if you try and place the corner EXACTLY on the reference piece corner, it will actually be a little off. Usually it is about 1/5th of a stud down and away from the corner (I use stud length because I'm pretty sure it matters how closely zoomed in you are). Like I said though, this is what I have found, and it might not be true for others.

At last if it lines up, you can click and pray . I would save your model before you try and connect it, as I have found that LDD sometimes will be able to save a model when it is separated, but not together. It also helps to save and restart LDD before you move it if you are worried it might crash.

As a side note, when importing a model or loading a template, make sure you have enough space so that it lands free, without trying to connect to the model that is already there. Sometimes the stress of loading a model in and then right away calculating a connection can make LDD crash. It also takes longer to re-locate the model. Oh, and if you move it to the wrong connection, LDD takes longer to calculate that.

Sorry if that is a little more than you asked for. Its just I've had experiences where it has taken LDD 10 minutes to load the connection (the big trees in my Nayeli moc) and if you really are making a 50,000 piece model, I'm pretty sure you are going to run into issues.

As for your 2nd question, its a bit complicated. I've been able to get almost 55,000 pieces in LDD 3.0, but I never used rotate, and very few types of pieces. Using rotate and more variety, I've managed to get 39,000 pieces in a very large model. The worst is 14,000 pieces max when I used a lot of the flex tool. So it depends on the complexity of the model. Surprisingly for this mini city, I built it entirely in one model, and it never crashed. It's almost 33,000 pieces, uses a decent amount of rotate, and lots of different colors and pieces. I saved like crazy during the end though with 34 versions saved.

But I think I remember you saying you have a better computer now, so it might be able to pull off a little more. But at these model sizes, it becomes less a computer issue, and more a program issue.

@eiker86: I'm probably not going to build this small version since it would probably cost more than $4,000, and I'm saving up for the full size. The full size version I am still working on in LDD, but hopefully I will start making the real thing in a few months! So Excited!!!

[eiker86]

wouldnt the real size be like $40'000 ?

Im really looking forward to see that one =)

[Superkalle]

Thanks for that summary Alienwar

This whole thing got me thinking about a new feature in LDD Manager. Basically the idea is that you put small "marker" bricks (like a 2x2 corner plate) in different colors in the base model. Then you put identical markers on the sub-models. In the end, it's just a matter of programmatically merging the sub models into the base model using the marker bricks as sync points. How does that sound?

[alienwar9]

wouldnt the real size be like $40'000 ?

I hope not...

Actually, my goal is to keep it under $26,000, though I think at the rate the LDD is going, I might have to cut something. That is why I need to finish the LDD model so I can properly analyze how much needs to be taken out, and where it can be taken out.

So far, since I don't like the way the EL-line looks in the lower city, so that might be cut out first if needed. After that, the subway would go. Then if still needed the industrial district (chemical plant first). Beyond that, I just plain fear

@Superkalle: That...could possibly be the greatest feature yet! Though I wonder just how much it could accomplish. While it would make it A LOT easier to get sub-models together, the biggest problem is still having the whole thing run in LDD. The main point of having everything together is to get a proper view and feeling for the model, and if there is no possibility of opening and using the camera view with LDD, that feature would be mute.

But it might also help make opening the model more stable too, as LDD wouldn't have to worry about calculating the connection. I hope I am interpreting the feature you described correctly though

[Superkalle]

@Superkalle: That...could possibly be the greatest feature yet! Though I wonder just how much it could accomplish. While it would make it A LOT easier to get sub-models together, the biggest problem is still having the whole thing run in LDD. The main point of having everything together is to get a proper view and feeling for the model, and if there is no possibility of opening and using the camera view with LDD, that feature would be mute.

But it might also help make opening the model more stable too, as LDD wouldn't have to worry about calculating the connection. I hope I am interpreting the feature you described correctly though

Yes, I think you interpreted it just right. And yes, it would only take away the painstaking job of assembling a massive model from submodels, but personally that is the thing that is driving me completely nuts. And if it enables us to open larger files without crash - the better

Actually, there should be a way to do it even without marker bricks:

1) Open the master file (with maybe just a baseplate on it or so) and attach one submodel. Save as SubModel_1

2) Open the master file again (with just the baseplate) and attach a second submodel. Save as SubModel_2

Now, both these files (SubModel_1 and SubModel_2) share the same coordinate system with the same origo. Open the two sub-models and remove the baseplate (and everything else that is in the master file). You can now continue working on the submodels until they are finished. As long as you don't move the entire model relative to the scene floor they will always maintain their position in x-y-z space. When you are ready to assemble, all you have to do is unzip and open the master in a text editor. Then you open the sub models one by one, edit the Ref-values(adding to them so they do not "collide" with each other) and pasting into the master file. Save as LXFML file and open in LDD. LDD will re-order the Ref-values in a subsequent string. Something like that.

I'll do some test and look into adding it as a feature in LDD Manager.

[jasonsmith]

Wow was that ever fun hovering around that city. Very well done Alienwar! Was nice to test my PC as well, the whole file loaded in about 50 seconds and panning and zooming was no issue.

[Lego universe player]

Sorry i need to collect my jaw from the floor because this project is so epicly (is that even a word?) big.

[kotaco]

Wow that's really impressive!