Proposal: Remove Skeleton Sketching #52809

Closed
opened 2017-09-18 07:33:08 +02:00 by Campbell Barton · 15 comments

Blender 2.7x has a Skeleton Sketching feature which I suspect is not used much.
(mainly because we've had bugs in this area that weren't reported for years ... or at all, I found errors just looking over the code. eg: 28f31bdfd8 a8487fc7fe 86a2e9d0e4).

My impression is that riggers normally want to control where their bones are placed with greater accuracy then you'd get from a sketch.

I'd be interested in feedback: are riggers using this and finding it saves them time?


Note that I'm not against the ability to sketch armatures at all, it just seems the current code to do this, while quite advanced - is not being used much and could probably be implemented in a simpler and more accessible way.

Current skeleton mode with it's own drawing, operators, key shortcuts... etc. doesn't integrate into existing workflows.

Blender 2.7x has a Skeleton Sketching feature which I suspect is not used much. *(mainly because we've had bugs in this area that weren't reported for years ... or at all, I found errors just looking over the code. eg: 28f31bdfd8 a8487fc7fe 86a2e9d0e4)*. My impression is that riggers normally want to control where their bones are placed with greater accuracy then you'd get from a sketch. I'd be interested in feedback: are riggers using this and finding it saves them time? ---- Note that I'm not against the ability to sketch armatures at all, it just seems the current code to do this, while quite advanced - is not being used much and could probably be implemented in a simpler and more accessible way. Current skeleton *mode* with it's own drawing, operators, key shortcuts... etc. doesn't integrate into existing workflows.
Author
Owner

Changed status to: 'Open'

Changed status to: 'Open'
Author
Owner

Added subscriber: @ideasman42

Added subscriber: @ideasman42
Member

Added subscriber: @JulienDuroure

Added subscriber: @JulienDuroure
Member

Hi all,
Never used it.
Mainly for the reason you suggest: I want to control exactly where my bones are. Using sketch need to go over each bone to adjust position afterwards.

Julien

Hi all, Never used it. Mainly for the reason you suggest: I want to control exactly where my bones are. Using sketch need to go over each bone to adjust position afterwards. Julien

Added subscriber: @mano-wii

Added subscriber: @mano-wii

I'm in favor of removal. This feature gets in the way when you try to make changes to some other parts of the Blender code. For example, the snap code.

I'm in favor of removal. This feature gets in the way when you try to make changes to some other parts of the Blender code. For example, the snap code.
Member

Added subscriber: @JoshuaLeung

Added subscriber: @JoshuaLeung
Member

I'm leaning towards removal of this too.

While it is undoubtably quite advanced and also rather cool that we have such functionality built into Blender, it also seems that hardly anyone actually uses it in practice. Last time I checked on an issue about this and made some notes about it, a lot of users were surprised that we had this functionality. But by and large, it appears that most riggers these days either use the autorigger scripts (Rigify, Blenrig) or hand-place bones, instead of using this to generate rigs for them.

Given that any drawing code it has will likely have to be rewritten for 2.8, it's worth asking whether those parts of the functionality would be better served by just using GPencil strokes, and maybe having the advanced functionality as an addon. Perhaps if it helped riggers set up all the Bendy Bone rest pose info (to maintain curvature), it might be worth keeping for certain use cases.

I'm leaning towards removal of this too. While it is undoubtably quite advanced and also rather cool that we have such functionality built into Blender, it also seems that hardly anyone actually uses it in practice. Last time I checked on an issue about this and made some notes about it, a lot of users were surprised that we had this functionality. But by and large, it appears that most riggers these days either use the autorigger scripts (Rigify, Blenrig) or hand-place bones, instead of using this to generate rigs for them. Given that any drawing code it has will likely have to be rewritten for 2.8, it's worth asking whether those parts of the functionality would be better served by just using GPencil strokes, and maybe having the advanced functionality as an addon. Perhaps if it helped riggers set up all the Bendy Bone rest pose info (to maintain curvature), it might be worth keeping for certain use cases.
Author
Owner

@JoshuaLeung, was also thinking this could be grease-pencil + addon.

To go into some more details - my impression is for rigging, placing the bones is not the bottleneck (not even close compared to setting up drivers, constraints, weight painting... etc), and riggers don't want to do this as an automated process because the bone orientation on join positions impacts the result enough - that poor decisions here add hassles later (too many joints may give extra work, too few, you have to redo weight painting... etc).

@JoshuaLeung, was also thinking this could be grease-pencil + addon. To go into some more details - my impression is for rigging, placing the bones is *not* the bottleneck (not even close compared to setting up drivers, constraints, weight painting... etc), and riggers don't want to do this as an automated process because the bone orientation on join positions impacts the result enough - that poor decisions here add hassles later *(too many joints may give extra work, too few, you have to redo weight painting... etc).*

Added subscriber: @zeauro

Added subscriber: @zeauro

If you remember the octopus tentacle demo, the tool was shown working with volume snapping for a mesh where the user want to produce a regular bone chain.
It was before the introduction of IKspline. But basically, it was to create this kind of bones chain and to retarget bones set-up at interesting scale.
The other demo video was about retargetting a finger set-up with 2 clicks at correct scale to 4 last fingers of a hand.

It is not surprising that people don't know or use it. It was a pre 2.5 feature that never really works correctly after 2.5 when all attention was on Rigify.
The cool thing in the feature is not working. And there is the problem to have to toggle Skeleton Sketching ON/OFF to be able to select one bone.
What is working is less interesting than Grease Pencil conversion.
So, nobody really cares if it is removed as is. We are not loosing anything.

But for the record, it would be great to have a rescaling/retarget tool. It does not necessary need to be based a GP stroke.
It could be an operator that replace a bone by a bone template of same length.
Or a general duplicating/rescaling operator that would work with objects, too, We have some selection, the tool register a start, a end. And then, we replicate it at a different scale by defining new start and end points.

If you remember the octopus tentacle demo, the tool was shown working with volume snapping for a mesh where the user want to produce a regular bone chain. It was before the introduction of IKspline. But basically, it was to create this kind of bones chain and to retarget bones set-up at interesting scale. The other demo video was about retargetting a finger set-up with 2 clicks at correct scale to 4 last fingers of a hand. It is not surprising that people don't know or use it. It was a pre 2.5 feature that never really works correctly after 2.5 when all attention was on Rigify. The cool thing in the feature is not working. And there is the problem to have to toggle Skeleton Sketching ON/OFF to be able to select one bone. What is working is less interesting than Grease Pencil conversion. So, nobody really cares if it is removed as is. We are not loosing anything. But for the record, it would be great to have a rescaling/retarget tool. It does not necessary need to be based a GP stroke. It could be an operator that replace a bone by a bone template of same length. Or a general duplicating/rescaling operator that would work with objects, too, We have some selection, the tool register a start, a end. And then, we replicate it at a different scale by defining new start and end points.

Added subscriber: @bunny

Added subscriber: @bunny

A tool with a lot of potential that got lost between 2.4x and 2.5. I've tried to use it a few times over the years, but came up against the issues @zeauro mentions.

The volume snapping and templating features would be extremely useful for rigging, but concur that the actual sketching should use the grease pencil stroke architecture rather than its own and that everything else could be handled by an addon.

I agree with tossing it if it's getting in the way, with a suggestion to retain or replace some kind of fast, robust mesh volume snapping that can be used for grease pencil stroke placement.

A tool with a lot of potential that got lost between 2.4x and 2.5. I've tried to use it a few times over the years, but came up against the issues @zeauro mentions. The volume snapping and templating features would be extremely useful for rigging, but concur that the actual sketching should use the grease pencil stroke architecture rather than its own and that everything else could be handled by an addon. I agree with tossing it if it's getting in the way, with a suggestion to retain or replace some kind of fast, robust mesh volume snapping that can be used for grease pencil stroke placement.
Author
Owner

Changed status from 'Open' to: 'Resolved'

Changed status from 'Open' to: 'Resolved'
Campbell Barton self-assigned this 2018-04-20 10:36:00 +02:00
Author
Owner

Committed 36773e35f6

Committed 36773e35f6
Sign in to join this conversation.
6 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: blender/blender#52809
No description provided.