Irregular room problems


Ahhhhhhhhh! I am trying to replicate a room which is very irregular because it is an old house. Whilst each room is roughly a rectangle shape, opposite walls are different lengths, etc and corner angles are not 90%. I’m having great difficulty in trying to get the walls to autofill. Does anyone have any words of wisdom on dealing with irregular room shapes please


Try to avoid modeling anything that is slightly off-axis or non-square or out of plumb. The modeling tools are specifically designed to model objects that are square and on-axis. It’s possible to model irregular geometry, but it’s a struggle every step of the way.

Perhaps you could model the room square and on-axis and deform it afterwards.



It’s sometimes easier to go with a maximum size box and move corners in to deform it to the correct dimensions.

I would recommend assuming that the floor is flat; draw a rectangle that is bigger than you need and use one of the edges as a start point to draw out the survey. Unless you took throw sizes (corner to corner) or corner angles, getting uneven walls to tie in is almost impossible.


I don’t think the answers given have quite understood the nature of your problem. I too work on old buildings that are off square and where nothing is parallel. It does make drafting much harder.

If your lines are not Autofilling, it sounds like you have one of two problems. Either the line ends are not joining up or one or more of your lines are not on the same plane.

You may find it helpful to start with a rectangle much larger than the plan you are working on. The lines that represent the walls would then be drawn on that plane. On completion, you can delete the bounding lines of the rectangle and any unwanted Autofill regions. This approach makes it easier to stay on plane.

I have just lodged a post to do with the difficulty of selecting points that are off axis. SU is set up with inferencing that has a much greater gravitational pull than an off-axis point has. So to select such a point, I zoom in to find it. Even so, it is quite easy to fail to lock onto it and then there is a gap and Autofill won’t work.


It seems like I’ve found the limitations of the software with the first thing I wanted to achieve. After spending literally all night creating one room with two windows I know exactly how much of a struggle it is to create irregular geometric shapes. In honesty, I’m ‘disappointed’ (pun intended) so far that it can’t handle the real world. Thanks for your reply though.


Ah ha, a fellow struggler. It has been a very frustrating learning curve literally spending all night trying to achieve what would seem a simple task. On my journey I met your friend ‘selecting a point’ issue, in fact, it seems I discovered all the problems with the software simply because I’m not a square! Like you suggest, I now see how it is better to create something larger (but regular) then cut away. I just wish they would put on the box, ‘not good at odd shapes software’ or even better, make it capable of dealing with more than legoland. Thanks for your good advice


There’s really not much call for the ability to model off-kilter structures. Consider the time and expense of developing an alternative set of spatial controls to accommodate skewed structures and compare that with the value added to the program, which would be virtually negligible except for a handful of oddball (you should pardon the expression) or whimsical applications.



It can more than handle the real world: have a look in the 3DWarehouse.

Can you post exactly what you were having problems with? There are probably several different ways to achieve the end results - once you get an idea of a better way to work/think about the model and the software you may revise your opinion…
… or you may not. Let’s see what you were/are trying to do?


Even the great pyramid is slightly out of alignment. Not good software for creating war torn environments then? I feel that tools are meant to do what we need them and whilst I do appreciate my scenario is an exceptional one, the fact remains that the older a structure becomes, the more irregular it becomes. Maybe they should add a ‘filler’ tool? :slight_smile:


Come live in the southwest of France where most buildings are in old pierre stone and you’d realise that regular buildings are the exceptions.


Do you think I’m denying that there are skewed structures in the world? I’m aware that the planet is full of dilapidated structures on the verge of collapse, and one needn’t go to France to find them. I’m simply saying that there is seldom reason to model such a structure.



A question I feel dangling in the air is “what do you intend to do with a highly precise computer model of the irregular structure?” In other words, what effect will divergence between the real thing and your model have on your objectives? Could you be overworking the problem by seeking an exact model where one really isn’t needed?

If your goal is artistic, then high precision isn’t necessary. Rough it out straight and square and then nudge things around a bit as needed to achieve whatever level of realism you want. If you were drawing a sketch with pencil and paper would you worry about whether your lines are or are not exactly parallel or if the as-drawn sizes are exact? Who will ever know or care?

If your goal is to design built-in cabinetry or renovations, I think that precisely modeling the irregular structure using an app such as SketchUp is misguided. Everyone I know either starts by repairing the structure so that it is straight or else builds with scribe allowances and fits in place. Only rough measurements and a model based on them are needed in this case.

So what is your end use for your model?


As slbaumgartner says.
My day job is dealing with existing buildings and designing alterations and refurbishments using BIM tools. In most of the cases the useful thing to do is to model the existing building by reproducing the original design intent, confirmed with measurements taken in situ.

Large irregularities usually don’t cause modelling problems, small do and are best ignored, also because many of them result from measuring errors. For instance, in old brick buildings walls usually are of a constant thickness within a tolerance and don’t taper, and concrete columns are mostly rectangular. Both are often misrepresented in surveyed plans.

I can safely say that there are no rectangular rooms in the real world, but that we can quite well live as if they existed.



Also, I’ve had a flood of email from a bunch of angry Egyptians. They have asked me to remind you that the Egyptian Pyramids have been renowned through the millennia for their amazing exactitude. Khufu, the Great Pyramid, deviates from true north by only 2.2 minutes of arc.



You mean it’s not perfect ?

Even if a USA sub launched a cruise-missile to land on your exact well-known whereabouts, it could miss you by 3m/10’ and still be “on target” !
But then, it might still hurt !

So NOTHING is 100% accurate - even this statement !


Not at all, like I said, thank you for your advice. My criticism is with the software not you and I have no doubt you have shared the same frustrations as me.


In honesty, what Sketchup offers will suffice and it is just the perfectionist nature I have that burdens me. I am an artist so I create what I see rather than ‘that will do’ mentality. But you are absolutely right in this case. Regarding repairing a structure so it is straight, yes, if one intends to line walls with plasterboard/gypsum, but with old stone walls, the character is in the anomalies and in this part of the world, plasterboard is for modern builds. Thanks for your words of wisdom


Maybe my first lesson should have been it ‘doesn’t have to be that precise’. In future a rectangle or a square will do just fine :slight_smile: Thanks for you reply


Sort of a paradox, is it not? Your perfectionism compels you to build imperfect geometry.



You never need to model something any more accurately than you might be able to build and then measure it.
And because you can’t measure it to the nth degree of accuracy, then you need to apply some ‘common-sense’.

Most buildings are not Swiss-watches, race-car engines, or nuclear-reactors.
For a typical building [e.g. a house], an architect’s plans might specify the length of a wall.
However, [in most common tolerances] if it is built within +/- 10mm [3/8"] of the specified length, then it is deemed to be that length.
If there were a much greater accuracy, then nothing would get built !

So if you survey a room that’s ‘almost square’, wouldn’t you be more logical to presuppose that the designer [and even the builder] had expected it to be square, and that the innate fallibilities of the many humans involved in its construction had somehow conspired against it.

BUT if it’s obviously ‘skewed’ away from a ‘square’ form, then you can accept the skew in the form - BUT bearing in mind that if the surveyed skew is within a degree of some obvious marker angle values, then they were probably the intended angles [e.g. 22.5°, 45°, 30°, 60°, 67.5° etc]