Before I get too much deeper into my current project, I’d like some advice on the best practice here:
When building a larger scene (no animation, just for rendering/art), is it fine to build my assets in one layer and use them in another, or would it be better to build them in their own file and append them in for use in the main project?
My graphics card isn’t the best, so I need to keep the file as light as possible, but I also dislike appending objects. Right now, my project is small enough that I COULD spend an hour splitting it up if necessary, but it has been very convenient having all my assets immediately at hand (and editable, when needed)…
I’m just wondering which is better practice, in your opinion, before I end up with a huge bloated mess I “can’t” come back from!
One negative experience I am having with building everything in one scene is slow exports. Since Blender export addons have to be written in Python, they adopt its crawling slow data throughput rate, and meshes are a lot of data which typically needs a lot of processing on export. Exports can easily take tens of seconds for large scenes and even minutes if you are building a level for a video game.
hello, most of the time the simpler the better !
especially if you don’t do animation ( multiples shots, with recurring assets) and if on top of that you work alone ( != collaborative workflow) then IMO it’s best to do everything in the same .blend. The only thing you can do is to make sure you uses instances when you have several time the same object or collection. This will save some memory on heavy scenes.
splitting everything in multiple assets ( = .blend) tends to make things more difficult to manage as you need to open each asset whenever you need to make a change. But this is a necessary evil when working within a team and especially when working in animation, when you want to avoid duplication of assets for each shots and also allow to change something and propagate that to all the shots.
Anyway, the simpler the better is a rule that generally works, don’t introduce a more advanced workflow unless you really see an obvious benefit, sometimes it just make thinks more complicated for nothing…
lastly, there might be one case where modeling objects in separate scene might be intersting that is if you need to reuse stuff. Maybe it’s simpler to have a chair.blend as an asset that you can reuse across projects. This is only half true though as it’s always possible to copy/paste stuff from one project to another without having assets splits anyway…
this may depend on your workflow. Some people do want to move a big mountain in the background seconds before the hit the render button. So putting “the background” completey in a “background scene/file” might not work for you.
Saying this: blender is able to move it even if it is a linked asset, but doing more manipulation does need to edit the linked file itself (there an addon to switch easily between them; but you also can load it into another instance).
Back to workflow: when people modularise even their assets they plan ahead and this is only possibe when go wild in creativity in any sketch / blockout so they know what they need (mostly in games where resources have to be "controlled). But as i said there are those who needs the powerfulliest machine because they do not push polygons but biilions of texels…
Thank for the input, everyone!
These are all about what I expected, just wanted to double check I wasn’t a complete dummy. : )
I don’t mind slow renders or exports, and my team is VERY small (3 people, working in completely separate files). The only assets we share are more complex objects, which are built individually and added in outside of Blends anyway. MY files are just for backgrounds, staging, etc. Static environments, with different views rendered as needed.
Linking is definitely better for shared assets! I use append to add in 99% complete pieces, because I’m very inefficient when modeling sometimes. Once they’re added to my scene, I like to be able to tweak them further within the project. Sometimes you just don’t spot that wonky vertex until it’s in place, you know?
See my earlier post with the extension links.
It’s very easy to jump back to your linked object, tweak it, save and jump back to your working scene.
Remember, the moment you append, all changes only are done in your scene. Not the asset you’ve built. It could make working together more time consuming when you have to update a asset as well.
Linked files definitely has been my way to build sets for different shots in my short film. I only append files IF I need to edit something like the shader for THAT particular shot and tweak it so I get the desired look. Things like repeated sets? Linked files. Characters? Linked files but Library Override.
Lighting? Depends on the situation, if there is repeated lights then yes I link those but for shots that require an extra light or four I just add them in.
But I rarely append unless it really really needs to be. I guess I got this practice from the Lightwave workflow back in the day? Very similar in theory.
I do as well. It was not the issue of really the amount of data it can handle in a scene, as more of how easily can this be edited down the line without breaking things or continuity? Back when the short film started yeah we had RAM limits but now? Not so much.
The reason more so is just keeping things in line pipeline wise.
I built the set using normal methods. I group things up in Collections so that I can quickly load things in without loading the whole thing, if it’s not seen there is no reason to link it later on. Each different shot uses a set file, which in my case is a file of linked collections.
In all the reason is just to keep things in line, shaders, textures, models everything… if I make an edit in a model, this model is linked into this and that shot. I don’t have to redo those shots from the ground up because a texture was changed.
…