Asset Browser: 3.1 (and some 3.2) Targets #94031

New Issue

Julian Eisel · 2021-12-13T18:03:30+01:00

Julian Eisel commented

2021-12-13 18:03:30 +01:00

Objective

Use 3.1 (and to some degree 3.2) to further polish the basic Asset Browser experience, before moving to the next big goals.
In a way this is a feature-fix release (like C++14 was to C++11 ;) )

Order roughly reflects importance, so that towards the bottom, items become more nice-to-have targets - that is, in terms of asset project strategy.

For each target there is a rationale explaining why this is something that should or can be addressed for the next release or in short time.

Bcon1 End Date: 29 December 2021
Bcon2 End Date: 26 January 2022

3.1 Targets

Collection Support

Why: The lack of collection support is a major limitation of the current Asset Browser. Collection support is very important, especially in more professional environments.
Status: Needs design discussions (e.g. how to handle instancing?) and implementation.
Who: @JulianEisel

Support Multiple Catalog Definition Files

#93755 (Support Multiple Catalog Definition Files)

Why: Single CDFs ask for editing conflicts when asset libraries are shared across multiple users; Simple drop-in of asset .blend files together with CDFs isn't possible.
Status: Basic design & code architecture support it, but details need to be figured out and actual support be implemented.
Who: @dr.sybren

Note that there is no big issue with delaying this to 3.2 if needed. It's of stratigical importance that this gets dealt with before too long, but it's not really urgent.

Indexing

#91406 (AssetBrowser: Indexing Asset Library)

Why: Asset library loading doesn't scale well currently. With a bunch of asset .blend files, loading time can be quite annoying; in production repositories it can take minutes.
Status: Almost done (see D12990). Basic indexing is merged but disabled, pending IDProperty support which is in late review stages.
Who: @Jeroen-Bakker

Object (& Collection?) Placement Tool

#90199 (Placement Tool for Objects and/or Collections)

Why: Current drag & drop with snapping is great, but limited. Give users quick access to important transforms and settings for object/collection placement.
Status: Needs discussion and UI design proposal/discussion.
Who:

Implementation: @ideasman42?
Design: ?

Improved Preview Support

Why: Current options for generating or loading previews are rather limited, and inconsistent between asset types. E.g. armature assets use the scene camera, while object and material assets have their own preview scene.
Status: Needs design. Multiple improvements possible, it's a bit open ended.
Who: ?

Allow Transforms After Linking

#93010 (Design discussion: Add object instancing to support transforming linked objects?)

Why: Many users seem confused or annoyed when dragging in objects with linking enabled, because they can't transform the object, and the surface snapping doesn't work.
Status: Needs design discussion. Feasibility depends a lot on that.
Who: ?

3.2 Targets

External File Dependencies

#91252 (As a user I want external files to be localized when appending an asset so they are part of the project I'm working on.)

Why: It should be possible to import assets without causing dependencies on the asset library. So dependent external files should be copied over to where the .blend file is.
Status: Needs design discussion (see #92315). Patches exist, but may need to be reworked.
Who: @Jeroen-Bakker

Responsibility of Other Modules

Grease Pencil

#95090 (GPencil: Asset Manager support)

Why: The Grease Pencil workflow could benefit a lot from the Asset Browser. Antonio is eager in working on that and already has a bunch of work done. But it needs design review.
Status Patch uploaded (see D13193), needs design and code review.
Who: #grease_pencil module

Node Groups & Geometry Nodes

#92811 (Support node groups in asset browser)

Why: Node workflows could benefit quite a bit from the Asset Browser. Not much work should be needed to get it ready and the nodes module can take do the remaining bits (low overhead for the asset team).
Status: Should be mostly there, but needs careful confirmation/testing and some nice-to-have features would be nice to have ;)
Who: #nodes_physics/#geometry_nodes

### Objective Use 3.1 (and to some degree 3.2) to further polish the basic Asset Browser experience, before moving to the next big goals. In a way this is a feature-fix release (like C++14 was to C++11 ;) ) ---- Order roughly reflects importance, so that towards the bottom, items become more nice-to-have targets - that is, *in terms of asset project strategy*. For each target there is a rationale explaining why this is something that should or can be addressed for the next release or in short time. **Bcon1 End Date:** 29 December 2021 **Bcon2 End Date:** 26 January 2022 ---- ## 3.1 Targets **Collection Support** **Why:** The lack of collection support is a major limitation of the current Asset Browser. Collection support is very important, especially in more professional environments. **Status:** Needs design discussions (e.g. how to handle instancing?) and implementation. **Who:** @JulianEisel **Support Multiple Catalog Definition Files** #93755 (Support Multiple Catalog Definition Files) **Why:** Single CDFs ask for editing conflicts when asset libraries are shared across multiple users; Simple drop-in of asset .blend files together with CDFs isn't possible. **Status:** Basic design & code architecture support it, but details need to be figured out and actual support be implemented. **Who:** @dr.sybren Note that there is no big issue with delaying this to 3.2 if needed. It's of stratigical importance that this gets dealt with before too long, but it's not really urgent. **Indexing** #91406 (AssetBrowser: Indexing Asset Library) **Why:** Asset library loading doesn't scale well currently. With a bunch of asset .blend files, loading time can be quite annoying; in production repositories it can take minutes. **Status:** Almost done (see [D12990](https://archive.blender.org/developer/D12990)). Basic indexing is merged but disabled, pending IDProperty support which is in late review stages. **Who:** @Jeroen-Bakker **Object (& Collection?) Placement Tool** #90199 (Placement Tool for Objects and/or Collections) **Why:** Current drag & drop with snapping is great, but limited. Give users quick access to important transforms and settings for object/collection placement. **Status:** Needs discussion and UI design proposal/discussion. **Who:** * Implementation: @ideasman42? * Design: ? **Improved Preview Support** **Why:** Current options for generating or loading previews are rather limited, and inconsistent between asset types. E.g. armature assets use the scene camera, while object and material assets have their own preview scene. **Status:** Needs design. Multiple improvements possible, it's a bit open ended. **Who:** ? **Allow Transforms After Linking** #93010 (Design discussion: Add object instancing to support transforming linked objects?) **Why:** Many users seem confused or annoyed when dragging in objects with linking enabled, because they can't transform the object, and the surface snapping doesn't work. **Status:** Needs design discussion. Feasibility depends a lot on that. **Who:** ? ---- ## 3.2 Targets **External File Dependencies** #91252 (As a user I want external files to be localized when appending an asset so they are part of the project I'm working on.) **Why:** It should be possible to import assets without causing dependencies on the asset library. So dependent external files should be copied over to where the .blend file is. **Status:** Needs design discussion (see #92315). Patches exist, but may need to be reworked. **Who:** @Jeroen-Bakker ---- ## Responsibility of Other Modules **Grease Pencil** #95090 (GPencil: Asset Manager support) **Why:** The Grease Pencil workflow could benefit a lot from the Asset Browser. Antonio is eager in working on that and already has a bunch of work done. But it needs design review. **Status** Patch uploaded (see [D13193](https://archive.blender.org/developer/D13193)), needs design and code review. **Who:** #grease_pencil module **Node Groups & Geometry Nodes** #92811 (Support node groups in asset browser) **Why:** Node workflows could benefit quite a bit from the Asset Browser. Not much work should be needed to get it ready and the nodes module can take do the remaining bits (low overhead for the asset team). **Status:** Should be mostly there, but needs careful confirmation/testing and some nice-to-have features would be nice to have ;) **Who:** #nodes_physics/#geometry_nodes

Julian Eisel commented

2021-12-13 18:03:30 +01:00

Added subscribers: @JulianEisel, @Jeroen-Bakker, @ideasman42, @dr.sybren

Garek commented

2021-12-13 19:09:27 +01:00

Added subscriber: @Garek

Tomek Gubala commented

2021-12-13 23:48:10 +01:00

Added subscriber: @VGT

Tomek Gubala commented

2021-12-13 23:48:10 +01:00

I don't know if this is the place for a random person to post their suggestions. If not, feel free to delete my comment.

I would like to chip in on "Improved Preview Support" topic. From my usage so far I see that I still prefer to generate my custom nicer looking thumbnails for assets. Workbench renders aren't that nice and there are some limitations. For example one of my asset had skewed texture on it, and it turned out I didn't know that workbench is probably not supporting more complex nodes structure when it comes to texturing. Example on screenshot below. Here probably Mapping node with slight transformations is not taken into account when creating thumbnail.

So back to suggestion: When I'm working on an asset file, I have it in my opened Blender, I marked created object as asset, I placed camera in a position I like and I hit render (in eevee or cycles depending on what works better for specific asset), or view -> viewport render image. At this point I have render I would like to use as thumbnail, but in order to assign it to asset I have to save it to disk, and then load back with option "load custom preview". It would be nice to be able to assign this render directly as thumbnail for asset. It's right there in blender and it would be nice ot be able to skip saving to hard drive step.

With this solution though there is still topic of thumbnails consistency. This way each asset thumbnail could look different, you have to remember to make background transparent. I don't have an idea for that. Actually my current solution for custom thumbnail consistency is blend file used for rendering thumbnails. With this approach it won't be possible to use this render to thumbnail shortcut I just described above because when rendering in dedicated blend file prepared for this, I'm no longer in asset file and therefore can't edit it. Gain some, lose some I guess.

Maybe ability to provide custom blender file as thumbnail renderer? Sounds like an overkill though.

I don't know if this is the place for a random person to post their suggestions. If not, feel free to delete my comment. I would like to chip in on "Improved Preview Support" topic. From my usage so far I see that I still prefer to generate my custom nicer looking thumbnails for assets. Workbench renders aren't that nice and there are some limitations. For example one of my asset had skewed texture on it, and it turned out I didn't know that workbench is probably not supporting more complex nodes structure when it comes to texturing. Example on screenshot below. Here probably Mapping node with slight transformations is not taken into account when creating thumbnail. ![image.png](https://archive.blender.org/developer/F12738314/image.png) So back to suggestion: When I'm working on an asset file, I have it in my opened Blender, I marked created object as asset, I placed camera in a position I like and I hit render (in eevee or cycles depending on what works better for specific asset), or view -> viewport render image. At this point I have render I would like to use as thumbnail, but in order to assign it to asset I have to save it to disk, and then load back with option "load custom preview". It would be nice to be able to assign this render directly as thumbnail for asset. It's right there in blender and it would be nice ot be able to skip saving to hard drive step. With this solution though there is still topic of thumbnails consistency. This way each asset thumbnail could look different, you have to remember to make background transparent. I don't have an idea for that. Actually my current solution for custom thumbnail consistency is blend file used for rendering thumbnails. With this approach it won't be possible to use this render to thumbnail shortcut I just described above because when rendering in dedicated blend file prepared for this, I'm no longer in asset file and therefore can't edit it. Gain some, lose some I guess. Maybe ability to provide custom blender file as thumbnail renderer? Sounds like an overkill though.

RedMser commented

2021-12-25 15:54:57 +01:00

Added subscriber: @RedMser

Ron Walker commented

2021-12-30 02:21:27 +01:00

Added subscriber: @o6a2

Robert N. commented

2022-02-02 21:38:07 +01:00

Added subscriber: @RobertN

Pauan commented

2022-08-08 11:53:28 +02:00

Added subscriber: @pauanyu_blender

Julian Eisel commented

2024-04-17 17:19:05 +02:00

This is quite outdated so closing this. Not all tasks have been completed, but that is general development at this point, this task is not really helping with that.

Julian Eisel closed this issue

2024-04-17 17:19:09 +02:00

Blender Bot added

and removed

labels 2024-04-17 17:19:10 +02:00

Sign in to join this conversation.