Noob to Blender with lots of questions and needing general advice

Hello all,

As the title stated, I am new to Blender and have just started my first project- Designing an abandoned insane asylum. It’s a bit expansive possibly, but I figure if I tackle something intricate, it will give me the best exposure to a wide variety of the tools, tricks, and techniques involved with Blender, and I’ll be working with everything from environments and lighting to object modelling as the set comes together, and so far I would say I’m doing a pretty good job.

At any rate, after going through the forums for a couple of weeks, I have so far been unable to find a good guide as to the whys/whens/hows as far as project structuring is concerned. So, I thought I would basically outline what I’ve been doing, what I hope to do, and hopefully this is of use to other newbies and possible some of the veterans can comment on ways that I should approach the project differently and give me a heads up on some of the problems I will come up against and how to avoid them.

My building has multiple wings and floors, and of course numerous cells, offices, open areas etc. So I started by creating the basic floor plan in GIMP and then pulling it into Blender as a background image and building my walls over the blueprints. However, I immediately noticed problems of scaling- I thought I had set the image up to be at the right resolution, but when I pulled it into Blender, it drew the lines thicker than I had, and no longer matched the orientation of the world grid. So, I scrapped that and have been building it basically free hand, extending cubes and planes into a configuration that roughly corresponded to my drawings. But, then I ran into an issue with a lot of vertices and sides that overlapped just slightly, and so when I would make an adjustment, the walls would act in ways that I didn’t anticipate.

I tried using the “delete duplicates” and was able to clean up a number of them, but still had to go through manually and check every corner, because no matter what the unit was set at for the tool, numerous vertices still wouldn’t snap properly.

For the first part of the project, I was joining all the pieces as they became finished, hoping that this would help prevent surprises, but as the building expanded, I am increasingly concerned that while it might prevent problems, it could also impede later editing. So I started breaking them up into smaller chunks using layers. Also, to help prevent duplicate vertices, I started applying grids to each piece. Since most of the geometry as far as the building itself is concerned is pretty regular, building on square meshes and using snap to vertex and automerge editing to make sure that everything lines up squarely and it seems to be going all right so far.

At this point, I believe I have most of the building in place and will only have minor tweaks going forward.

So, the first questions are, what advantage is there to joining versus parenting, is there any advantage to having the building as one piece, as opposed to many? Should I be working to eliminate unnecessary polygons as I go? Like the grid meshes that I am using to line up the floors- Should I be merging those into a single piece, or leave it in squares and subdivide later if I need to make changes to an area.

The next step in my project is (I think) putting holes into walls, floors and ceilings, putting in doors, windows and the some of the larger elements. Any tips, tricks, or tutorials that you would recommend for this?

After the building is exactly how I want it, I will be going through and adding textures and lighting, as well as things like standing water and other elements.

Then, once those pieces are done, I will be going through and adding furnishings and other objects to make it feel used and occupied over decades. I have a number of example objects that I have found on other blender sites, but will end up building my own using them as rough guides, and building my own stuff as well.

Then, I will be adding debris and other finishing touches.

At that point, I think it will be done (though I realize that I am probably looking at another couple weeks of work at my current rate) and it will be time to actually try rendering the thing. I fully expect it to take a long time since I am going for a photo-realistic, or near photo-realistic look, but if there are things that I can do to reduce render time, I would appreciate any tips.

I am at a stage in the project where, if there’s a better way to do what I want, I am not committed to my current methods and can take the extra time to make sure that it’s done according to best practices (as much as there are any), and can rebuild areas relatively easily if it will make the whole thing go smoothly.

Additionally, here are some of the questions that I have had as I’ve been going along:

  1. Is there a way to go through the environment in 3D? I want to make sure that I’ve got things lined up in ways that work well for the camera, and trying to manipulate the view is kind of finicky at times. I don’t care if it’s low-res, but I want to make sure that the sense of scale is correct- Trying to find that balance between things feeling like it’s an actual hospital, while maintaining the claustrophobic sense I am going for is difficult with the standard views and trying to pan/zoom.

  2. If corners and faces over lap, what sort of rendering issues will I have? My only exposure to 3D environments before this was in editing Doom wads back in the 90’s and I know that texture overlap caused a lot of problems, but I have no idea how well Blender does with this sort of thing.

  3. For things like doors and windows, as mentioned earlier, or wall trim and other smaller touches, is it better to actually build them, or is it better to to use textures with the fixtures on them, or is this dependent on where the camera is going to be? E.G I’m thinking at a distance, I can probably get away with using an image of a window, but as the camera approaches, there will be distortion, especially at angles. Or is Blender pretty good about making the necessary calculations that those sorts of things generally don’t become problematic? Since I don’t know exactly where I want the cameras to be, I’d like to try and do everything in a way which increases my flexibility in that regard.

  4. I am hoping, ultimately, to use this model for photos and videos using chroma-key replacement so, what I’d like to do with the finished model is break it into pieces like the sets for Nightmare Before Christmas, so that I can stage the shots I want without having to spend days or weeks on rendering parts of the environment that are never seen. Is there a good tool for this?

  5. Even though most of the shots I am visualizing are internal, I know there will be times where the outside world will be visible- Through a window, hole, etc… What is the best way to accomplish this? I know I can include background images, or possibly put the entire thing in a box with the containing walls having textures of the sky or whatever… Are there any tips as far as this goes, so that it isn’t readily apparent that it IS in a box?

  6. Do objects outside of the world grid still get rendered? For example, one of the buildings that I have, includes a basement. Since I started building on 0 Z, the basement currently exits below the world grid. Is this okay, or should I just raise the entire set and put in a plane for the ground, and have the basement start at 0?

  7. When it comes to using chroma-key to add players to the photos/videos, are there any tricks when it comes to lighting, either of the players themselves during filming, or lighting in the environment, to make it easier to make sure they blend well?

  8. Are there any suggestion, tips, tricks, or tools that Blender has, or which can be added on, which would facilitate this project in any way?

As I get further along, I expect I will have more specific questions, but hopefully this is enough to keep going for now.

Thank you everyone!

1 - Walk navigation (like FPS), shortcut Shift+F http://wiki.blender.org/index.php/User:Dfelinto/Walk_mode

2 - Overlapping faces can cause render artifacts if using the Blender Render engine

3 - If the result looks as you want it to look then that’s all that’s important

4 - Use render layers to render different scene layers. Combine them later in the compositor. In the render settings enable transparent backgrounds so easier to overlay the different scene components

5 - Either, whatever looks ok

6 - Things outside the camera view can also affects other objects in the camera view. If you don’t want to render something you can put them on a non rendered layer or in outliner make them non renderable. In the camera settings you can set start and end clipping distances

7 - In the lighting is coming from the left in the background but from the left for your dropped in player it won’t look correct. USe same lighting

8 - Watch tutorials. There are a huge number available on youtube. Also look at cgcookie/blenderguru etc for additional tutorials
Don’t try and put all your questions, especially vague very general questions in one gigantic post, it’s tedious to anyone who may want to help you. Also remember don’t be afraid to delete your work and start again. Trying to fix bad modelling can take longer than just starting again knowing what not to do and making a better result

Thanks for your response. I’ve been watching tons of tutorials over the last couple of weeks but they all tend to be on a specific element, frequently modeling single objects, and I am not yet familiar enough with working with these tools to be able to extrapolate to my project, so I was hoping that giving a larger context might yield more directly applicable results.

As far as the lighting, I understand wanting to use similarly located sources, as in your example, but how does that work in a more complex environment? For example, if I have someone partially obscured by piping or something, are there decent cheats to get it to look correct, or is that something where during filming, I will need to have those sorts of physical objects on hand to cast the shadows?

The easiest solution would be to not have complex lighting, but I am trying to determine the functional limits of the software before I work myself into the corner.

And I will keep my posts more focused in the future, but at this stage, I don’t really even know what to ask.

Looks like you have in mind a massive architectural visualization project, with some idea about using the project as an animation set later on. Your goal seems to be working on a project that will give you the greatest exposure to Blender in the shortest amount of time.

If that’s the case, then building such a massive structure is overkill. Once you learn how to build and texture a single hall with an open room and several closed doors, and to fly a camera down the hall and into the room, you’ve learned all the technique your going to learn from that project. Duplicating this lesson into two wings, several floors, open areas, cells, offices, etc, etc, etc. will not teach you anything but how to survive tedium.

But you also want to use the project as a stage set for animation, and asked about a tool you could use to stage the shots you want without having to spend days or weeks on rendering parts of the environment that are never seen. There is such a tool, and it’s called a storyboard. Work out a short animation project and draw up storyboards for it, and that will tell you what you need to add to your single hallway with an open room and several closed doors. The rule is: only model what the camera will see.

Do this, and the additional work is not tedium, but gets you closer to your goal of making the animation clip. In addition, the storyboard will tell you where the camera is, so you can make decisions about the level of detail various parts of your model need.

If you would, get the Sintel DVD, with the movie files on it. Take a look at the sets they used to create the illusion of Sintel’s home city, with it’s streets, temples, ruins, marketplaces and so on. There’s actually a lot less there than you think you’re seeing on the screen.

Remember: you are studying to become a magician, and a magician always remembers that he is creating an illusion.

Welcome to BlenderArtists :smiley:

PS: there is nothing special about the grid at z=0. If the camera points at something, it will render, unless otherwise directed.

Orinoco,

Thanks for your response. I’ve actually done a bit of what you suggested as far as some basic storyboards for animation and rough sketches for specific shots that I’d like to create, and am focusing primarily on those elements. But with the way my brain works, and my general creative work-flow in other avenues, I have to feel like I have a solid grasp of the whole area- So while I may be focusing on, say, a single hallway, I NEED to know what is going on outside of that area… Call it a compulsion. However, I am not trying to create a fully realized, fully rendered set yet… though, the thought has crossed my mind as I’ve been watching a few tutorials on using the game engine.

My basic approach so far has been to block out the floor plan, and I am now putting down basic textures (with mixed results, but that’s another matter) just so I can get a solid feel for the place… So while crossing hallways or windows or whatever may not be anything shown (or show-able), I know what’s there and have a general idea of how it looks, with the emphasis being on the areas that I know I definitely will end up being part of a final render. In part, it keeps me focused on what I need, while still helping me keep a running list in the back of my mind as far as what things to be looking out for. For example, I know that I will have one camera track down a hallway of cells, and the angle that the camera will be at won’t require me to worry about glass refraction or reflection in the cell windows but, since I know that there will be glass windows, if I see an interesting thread or tutorial about those subjects, I can bookmark it for later.

But also, part of my reason for approaching it this way is that it gives me an environment in which there can be accidents. I might discover a shot with really interesting composition, or I get strange lighting or something, that I might not be able to (or even think to) set up deliberately.

So while it’s definitely slower going than a more stripped down approach, it just feels better to me.

Regarding the tool that I was looking for (and yes, storyboards help), what I was referring to was a way to get that rough walk-through ability that Richard provided the short-cut to (ctrl+f) At the time, I was on an older version of Blender that had fly-through, but no walk, and the controls were a bit… difficult. Being able to walk the environment with gravity on has helped a lot, especially in seeing where walls don’t quite meet floors/ceilings/each other or where there are texture artifacts because one texture is too close behind another from a particular angle.

However, I have seen a few tutorials with the game engine where they are able to show the entire set they had designed, complete with lighting effects/fogs/etc. that don’t show up in 3D mode with the textures turned on in the regular Blender engine (or, at least I can’t figure out how to do it yet), but I haven’t figured out how to do that just yet- But since lighting is further down on the to-do list, it’s not really a priority, but its nice to know that there is a way to do it.

My plan right now is to finish applying the basic textures and materials to keep the voices in my head happy, then go to the storyboards I’ve got and then specifically focus on those shots by hacking out all the walls that don’t get used, simplifying the side corridors and the like that are only going to be seen at a distance, and really getting into the high detail work on what’s left to keep final render times lower. But at least I will know (even if nobody else does, or even cares) that the final geometry is correct, and, if I need to create filler material for any reason, or my concepts change, I have places to start.