Is there any perception that can be shared about which is a “large object” and a “small object” for blender modeling. I’m talking about limits, practical ones, not in theory.
In my mind the object size has to do with resolution. But I don’t have a feeling (or objective reasons) to use on scale instead another.
Let me give a concrete example:
I’m working on a simple spaceship model expected to be 150m long. Working with centimeter resolution sounds more then good enough. In this case I can pick 1m = 1 blender unit or 1m = 0.1 blender unit (BU).
I’m using now with second option (1m=0.1 BU) because when I started with 1=1 I started to reach limits of 3D view with strange behavior. And I don’t know how to resize grid
This bring to my mind the question: which are the practical limits for a models in blender"
Note: I’m not an artist used to work with modeling tools, just an unskilled curious who love 3d modeling art and enjoy to give it a try from times to times. So this question may sound a bit rookie ! (and that’s exactly the case)
If you are getting clipping in the 3d view and/or camera view just increase the clip end distance in the view properties (N).
It really doesn’t matter what you have as a blender unit for modelling so don’t worry about it.
Actually it can matter, significantly, in some instances. AFAIK all the physics sims are based on the 1BU = 1m model scale. If you depart significantly from this (such as 1 BU = 1cm) you may run into issues with getting Cloth, Soft Body, etc., to look right.
Not sure if it’s been changed in 2.5x – haven’t had much time to play with it yet – but in 2.49b and earlier, there are hard-coded limits on some values you can set (such as the Clip distance mentioned – 5000 BU). Model scale can become an issue if you build at a scale that requires exceeding those limits – for example, 5000 meters is a long “reach” for the camera. 50 m is not. This can be an issue when building virtual sets to match live action, among other scenarios.
Decimal point limits on values can also be a factor – with 1 BU = 1m, your smallest-scale object will be ~ 1mm in size, so if you need a different min/max limit, you may have to adjust the world scale somewhat.
Yeah. Physics are based on Blender Unit is 1M. Even though you may never use physics, it is probably a good idea to just keep everything to scale in real world terms - for the most part. There will be exceptions. But if you want to know, most modelers try and do things to scale. It is the one thing that can remain universal. When you export a model to another program for instance you will also find that the exporter uses Blender unit as 1M.
They addressed very well the question.
The point about physics engines use often 1m=1bu is a good point.
I’m not going to use a physic engine but the use of 1m=1bu can easy a lot of tasks related to real placement and movement.
I’m thinking in switch to new 1m=1bu. But I’m finding some strange clipping in 3D view I can’t figure out why.
I did a simple and small example in five screenshots so I can show you what is happening.
First four (made of a simple scene with one camera, one 20x20x20 box and a lamp):
The sequence shows what happen to camera control axis according to placement. The Z axis has been clipped from view !
Note that the camera is about 300 units from center of working area. It’s not that far and much less than 5000 camera clipping limit.
The last screenshot show the cube resized to 250x250x250 and shows exactly where clipping is happening.
Can someone help me to figure out what I’m doing wrong?
A not uncommon problem – you are confusing the rendering camera (aka the Active or Current Camera) with the camera used for the 3D viewport – they are not necessarily the same. So you have to set Clip parameters for the 3D viewport camera or you’ll start to clip out parts of the 3D view. In the 3D viewport menu, open up View>View Properties…
To see in a 3D viewport what the Active/Current Camera sees, select the Camera you want to view through and use CTRL+NUM-0. This sets the 3D viewport to look through the Active/Current Camera, and its Clip parameters will then apply.
I touched another situation about ideal object size:
Ambient occlusion baking process gives different results according to mesh size.
Same mesh with 16x16x6 gives a nice baking result. But if you resize mesh to something huge like 800x800x300 the result almost useless. It appears the algorithm uses some hardcoded values to work.
