Asset Browser - why still so WIP?

We have Asset Browser available since 3.0, I have now tested 3.1 beta and 3.2 alpha, and to my surprise it is still so limited, that I find it barely usable for any real usage, for example:

  • still no options for actual placing an asset (I would expect some snapping options, ignore rotation / random rotation, etc)
  • no way to reload linked assets (reload button just refreshes the library listing)
  • no support for objects with children (only after some Googling we can find out this can be achieved through instanced collections, which is kind of workaround - for example it doesn’t show actual hierarchy, there are Empties created after Making Instances Real etc.)
  • still no thumbnails generating for instanced collections (need to use a clusmy 3rd party addon for that)
  • when placing linked, no option to link only data - that’s also why linked assets ignore cursor position when placed (always at 0,0,0) and can’t be transformed until Localize → Object
  • weird deleting an asset through right-clicking a category → Clear Asset (instead of right-clicking the actual asset)
  • 3.0.x creates extra copy when placing instanced collection asset (bug?)
  • and many more

Are my expectations just too high when I just want Asset Browser to offer me a library of objects (including those with children), show their thumbnails automatically and keep them linked once placed (on a cursor position)?

1 Like

What bugs me is complete absence of news regarding the support for other data-types

For example, poses came before fully animated actions, even though Blender doesn’t have any kind of way to reliably store and manage multiple animations for a character. Instead all of your animations belong in a single drop down menu, where you will lose them if you don’t fake user them one by one

Brush management is a direct and highly requested use case, and we know that the main blocker for it was the asset design and lack of UI dev resources. For some reason now that we have a functional asset browser, this is still not even discussed

It seems like the main focus for asset browser was to give online platforms a way to integrate assets natively, which is fine I guess, but resulted in a much less user-centric design

I think the key word here is patience. It may look simple from the outside, but the asset browser is a very complex thing to get working/right. So far there has been a lot of work done on getting the core implementation working, and then adding in features incrementally.
We need to remember that the asset browser is still very young, and has already improved enormously, with a fully working system for catalogues, naming, importing etc. none of which is an easy task, plus support for adding objects, materials and poses.

Other data types and quality of life features will come, we just need to have a little patience.

6 Likes

But I am talking about very basic functionality, like creating an asset consisting of an object & children.
Nobody will ever find out how to do that without searching on internet (and even then it doesn’t work properly as it behaves differently in different Blender versions, no thumbs etc).
I find this quite an issue, considering the Asset Browser is included in stable 3.0.0 - one would expect such release to have basic features polished (i.e. deleting an asset by clicking on category - is it really that complex thing to fix?).
I am noticing similar problem with many other new features in recent Blender versions - there is often some tool / feature added in very early / WIP state, and then instead of polishing it’s basics functionality (general UX & fixing main issues), it rather keeps getting new niché features.
I am not sure why is that.

1 Like

Interesting, like what?

TBH, I think that features like this should probably have a WIP tag or something given to them until they are fully complete so that people aren’t unhappy when something hasn’t been implemented yet.

I would prefer not including WIP features into stable releases at all. They could be enabled optionally in User Preferences for example.
I wasted all today by trying to figure out how to use Asset Browser for objects with children (game asset consisting of the model & it’s collision mesh as children - very basic thing in my opinion), and be able to drag&drop them into the scene to a cursor position with links prevented … without success, as you can either Append and then drag&drop to the cursor position, or Link and then it is always placed to 0,0,0 (then you need to Localize Object to transform it).
Yesterday was wasted as well, by figuring out why assets made from instanced collections create another copy once placed into the scene … just to find out 3.1 fixes that (so probably bug).

What I am trying to say is that if Asset Browser is marked as WIP / alpha, I would keep using Libraries and wouldn’t waste those 2 days in believe such simple features must be supported in stable release, so I should keep trying …

Generally I like using Blender and there have been definitely many useful features added in last releases - don’t get me wrong. But sadly, starting from 2.8x I always find a bug or serious limitation in any new feature just after few minutes of fliddling with it, even in stable releases.
It seems like pushing new releases to hype the community has bigger priority than offering as polished sw as possible.

6 Likes

Well said!
Asset browser is in it’s kids shoes…same like some little important things. (basics)

1 Like

I wouldn’t go that far… Personally, I prefer to have access to these features earlier, even if they aren’t entirely polished. The BF never puts anything that doesn’t work in, and especially if using an LTS release, I’d rather be able to use this feature for the next two years or so until the next LTS, even if it is in a limited form, than not at all.

Also, it’s not like it’s a choice between putting a feature in a release, or polishing it. Polishing and finishing features will happen at the same speed either way, it’s just whether we get access to them early. It’s also worth pointing out that there are advantages to doing it this way, as the devs get a lot more user testing to fix bugs and find out what users want.

1 Like

… I’d rather be able to use this feature for the next two years or so until the next LTS …
… It’s also worth pointing out that there are advantages to doing it this way, as the devs get a lot more user testing to fix bugs and find out what users want.

That’s the purpose of Alpha / Beta / Experimental Features, isn’t it?
In my opinion no stable release should include a feature (enabled by default) which is WIP or even has flaws in it’s basic functionality / design.

Sure, but full releases are far more likely to be bug free.

Honestly, I’m not really sure what you’re unhappy about. The alternative is that we have no asset browser at all for potentially months or years until all the miniscule details are finalized. I’d rather have access to what is there now, than have to wait for it to be completely finished.
And if you don’t like that, you can just as easily not use the asset browser for the next year and get the same effect.

The only problem I see is that the devs have not made it clear enough that not all features are available yet, and that they will come in the future…

Sure, but full releases are far more likely to be bug free.

The problem of the current Asset Browser is not it’s bugs (even though there are some), but it’s features & design which is IMHO so limited, that it can’t really serve it’s purpose (no linking support, no asset update / reload, no object hierarchy support etc).

I just think no stable release should include tools which are barely usable, because users will try to include them into their workflow (in believe stable release is a polished product with fully working features) and in the end waste their time by trying to figure out why it doesn’t do what they expected.

The design isn’t limited, it just hasn’t been fully realized yet.

It seems to me that your problems would be fixed also by just including a message saying that the browser is not fully complete yet and there is more to come. If they do that, then I see no downside to including it.

There is no concept of grouping in Blender so you idea of dragging a parent with its children won’t be even possible in future. The best workaround is using collections that were recently supported.

If making an asset consisting of an object & children (like the simple example with a game model & collision) requires workarounds, then there is something wrong with the design.

4 Likes

Asset Browser is still so WIP because, well, there is a lot of work to do and not much time to do it.

That is something you want for your work, that’s your priority. Blender developers have to take into account many possible ways of using Asset Browser by gathering feedback. So far overwhelmingly people need collection support so work is being done in that direction.

There is no “complete” absence. It’s just scattered across different (open) places.

It was stated before that other data-types is responsibility of modules and community while core Asset Browser team can provide consultation and fundamental code support. As such, there is a patch for GPencil support and work is being done for Geometry Nodes groups.
Brush management is on sculpt roadmap but Joseph have a lot on his plate already to even start discussion regarding that.
@tintwotin has worked on implementing image data-block.
Sybren, developer of Pose Library, no longer working on it, instead he is now developing Flamenco, so further work on animation data-block have to be done by someone else.

Main focus for Asset Browser in 3.0 is to roll out basic Asset Browser as polished as possible. So, you can mark objects, materials and worlds as assets, create thumbnails, have catalog, can drag-and-drop assets. Seems basic to me.
Pose Library seems like different project since it even have it’s own section on pretty release notes.

I’d argue about that. This is functionality that you often use in your work - but that does not mean it’s basic.

That is already the case, you can only enable Experimental features in alpha builds.
Blender operates on the rolling release basis, it’s not gatekeeping features when they’re ready to be included. You can always improve features later on, including based on feedback that was received upon release.

Alpha/beta build are not for features that are “not ready”. It is for features that will be in release and there needs testing to catch and fix bugs that developers was not able to notice before.

What you want is probably for features to be “production ready”, and there is no rule of thump when that’s the case and we’ll end up with features that could be used by people rot in WIP status. We would not have Geometry Nodes, Asset Browser, Cycles X available for people to this day.
Besides volunteer (and even paid) developers would lose motivation to work on features that will not see light of day until 1-2 years after they’re ready for people to use, even if not perfectly polished.

Or if we would make Experimental option back in release and put those features in it, things would be the same as before: nothing changes, anyone working with Blender just enables Experimental option first thing after downloading Blender, making that checkbox useless.

2 Likes

To me after using the Asset Browser for “map making”, ie. creating a bunch of assets and making a library out of it, then linking them in a new file.

And i find the way this works problematic and unintuitive.

First issue, i cant lock orientation prior to placing a object from the browser, this is problematic, it adds extra steps for me to fix orientation, i use the LINK option, because i want to be able to reedit an asset down the line/refine etc. without readding or changing the mesh data on every object.
On top of that LINK does not allow for any Translation/orientation/scale when the object is placed in the viewport and it gets placed at 0,0,0, which is practically useless as is.

To fix that i now have to select the said object i placed and go to Object > Relationship > Make library override to be able to Translate or Rotate the object, but this is not all, after making a library override i then have 2 copies of said object, one that’s frozen at the dawn of time which gets autoselected and under it is the “new created” overridden one, now i have to delete the original one since it’s just there wasting viewport air.
And all of this i was able to do only after reading the documentation and some forum posts, which shows that this is not very intuitive approach of the asset browser for new users or first time asset browser users. (In case of the “LINK” function)

It would be great if there was a button that ie. Quality of Life Feedback

  1. Restricts auto orientation to a axis (eg. place a object on uneven terrain always facing up),
  2. Have a button that when you drag an object in to the viewport it auto applies a library override so that you don’t have to do that for every single object by hand,
  3. It would be practical if we had an option when dragging an object in to the scene for it to be in Wireframe instead of the bounding box, this would help in placing complex objects in the scene.

I would also like to point out a small thing, because we don’t have a “Group system” (which a lot of people want including me) we cant “asset” a group of objects without joining every mesh (which in my opinion creates a nightmare if you want to edit the object later on when meshes overlap etc.)

2 Likes

It was for Movie Clip previews, however, this seems to be yet another unwanted contribution of mine and will most likely result in nothing.

What you want is probably for features to be “production ready”

Everybody who is actually working on any project which benefits from separating it into parts (including assets), always takes advantage of libraries and linking stuff together - it makes the project better organized, filesizes smaller, prevents from duplicates, keeps everything up-to-date etc.
It makes no sense to create assets for a project without keeping them linked at the same time.

That’s why an asset tool which can’t handle linking properly (among other things), is not only production unready, but just useless.

Can you link us to the code needed to implement that feature?

Can someone explain me the idea behind why deleting objects marked as asset in the “Current File” mode from the scene deletes them from the asset browser as well? That makes absolutely no sense to me.

When a user explicitly marks something as an “asset”, one would expect it actually gets treated as an asset, meaning the type of object that does not necessarily have to be present in the scene to be available.

If your “assets” still must exist somewhere in the 3D scene to be available in the asset browser, then the whole idea of having asset browser as in a library of the assets has been defeated, because we are exactly back where we were before the asset browser was introduced. It brings no real functionality to the table aside from slight UI convenience.

Yes, I am aware you can operate in the “User Library” mode, but that kind of defeats the idea of categorizing assets based on whether they are universal or scene specific.

Right now, it appears that Blender’s idea of what “asset” is very subjective one, and incompatible with the rest of the industry. The idea of having to be careful about deleting the “instances” if the given asset in the 3D scene because once you delete its last instance, you delete the asset itself is just straight up crazy.

1 Like