Skip to content
This repository has been archived by the owner on Mar 28, 2024. It is now read-only.

[BUG-10990] [Bento] A formal method of bone-translating animations is vital for the creation of proper facial expressions. #1252

Open
1 task
sl-service-account opened this issue Dec 18, 2015 · 16 comments

Comments

@sl-service-account
Copy link

sl-service-account commented Dec 18, 2015

With the addition of so many new face bones, it's unfourtinate to hear that we will only be able to rotate these bones, rather than translate them with animations. This disregardes basic facial animation pracices, where the movement of facial features (eyebrows, smiles) is necessary to properly create expressions. To give so many new bones but deny a way to animate them properly seems like a huge oversight. Proper facial expressions simply cannot be accomplished with bone rotations alone.

With their permission, I've included a conversation with a creator of fully custom mesh avatars:

[04:28:52 PM] EpicGordon Broome: I don't like getting involved in development, JIRA's or forums or whatever. But disallowing bone translation in animations is ignoring some very basic fundamentals of animation theory, dating back decades. I've scrapped multiple projects because of the lack of squash/stretch in walk cycles (Toony characters). This update is also going to break features in an upcoming avatar of mine, regarding facial animation. Which is funny, because they give us so many new facial bones - that we just can't use properly. Raise eyebrows, pull up mouthcorners for a smile, extend a tongue, using only rotations? I cannot for the life of me see a GOOD REASON for this limitation.

I have this feature - http://imgs.moyloon.com/i/05d71e3d26.png - almost done for a new avatar of mine, a pose-able tongue. The new rig has two bones for this explicit purpose. But they are useless, as I cannot pull the tongue out of the mouth in any good fashion. I cannot pull the mouth corners around without arbitrarily locking them to rotating around a bone.

Not to mention non-humanoid avatars. We're getting so many new bones - and yet we'll be limiting the design freedom so much, since anything non-humanoid will be unable to avoid issues caused by inheriting rotations, since locations cannot be transformed.

Allowing animations to do this is a trivial change, as I'm fairly certain that limitations were put in place to prohibit this (Hence why .anim's were able to do things .bvh cannot), rather than requiring lots of labor on the engine.

It doesn't affect me personally, too much. If this is the way it ends up being, I'll just hog more resources by alpha swapping meshes, and using more animations for each of these different states. However, it's unnecessarily clunky, and will ruin a lot of content creator's livelihoods in its current implementation.

Links

Related

Original Jira Fields
Field Value
Issue BUG-10990
Summary [Bento] A formal method of bone-translating animations is vital for the creation of proper facial expressions.
Type Bug
Priority Unset
Status Accepted
Resolution Accepted
Reporter Adeon Writer (adeon.writer)
Created at 2015-12-18T01:26:31Z
Updated at 2016-11-01T11:56:31Z
{
  'Business Unit': ['Platform'],
  'Date of First Response': '2015-12-17T20:14:01.153-0600',
  'ReOpened Count': 0.0,
  'Severity': 'Unset',
  'System': 'SL Viewer',
  'Target Viewer Version': 'viewer-development',
  'What just happened?': "With the addition of so many new face bones, it's unfourtinate to hear that we will only be able to rotate these bones, rather than translate them with animations. This disregardes basic facial animation pracices, where the movement of facial features (eyebrows, smiles) is necessary to properly create expressions. To give so many new bones but deny a way to animate them properly seems like a huge oversight. Proper facial expressions simply _cannot_ be accomplished with bone rotations alone.\r\n\r\nWith their permission, I've included a conversation with a creator of fully custom mesh avatars:\r\n\r\n{quote}[04:28:52 PM]  EpicGordon Broome: I don't like getting involved in development, JIRA's or forums or whatever. But disallowing bone translation in animations is ignoring some very basic fundamentals of animation theory, dating back decades. I've scrapped multiple projects because of the lack of squash/stretch in walk cycles (Toony characters). This update is also going to break features in an upcoming avatar of mine, regarding facial animation. Which is funny, because they give us so many new facial bones - that we just can't use properly. Raise eyebrows, pull up mouthcorners for a smile, extend a tongue, using only rotations? I cannot for the life of me see a GOOD REASON for this limitation.\r\n\r\nI have this feature - http://imgs.moyloon.com/i/05d71e3d26.png - almost done for a new avatar of mine, a pose-able tongue. The new rig has two bones for this explicit purpose. But they are useless, as I cannot pull the tongue out of the mouth in any good fashion. I cannot pull the mouth corners around without arbitrarily locking them to rotating around a bone.\r\n\r\nNot to mention non-humanoid avatars. We're getting so many new bones - and yet we'll be limiting the design freedom so much, since anything non-humanoid will be unable to avoid issues caused by inheriting rotations, since locations cannot be transformed.\r\n\r\nAllowing animations to do this is a trivial change, as I'm fairly certain that limitations were put in place to prohibit this (Hence why .anim's were able to do things .bvh cannot), rather than requiring lots of labor on the engine.\r\n\r\nIt doesn't affect me personally, too much. If this is the way it ends up being, I'll just hog more resources by alpha swapping meshes, and using more animations for each of these different states. However, it's unnecessarily clunky, and will ruin a lot of content creator's livelihoods in its current implementation.\r\n\r\nI dunno, I guess I could provide more concrete examples, and would love to chat about pros and cons of different implementations. I think there's a better solution altogether for giving creators more bones to play with - but I guess it's way too late to change the basic fundamentals of this update at this point.{quote}",
  'What were you doing when it happened?': '_',
  'What were you expecting to happen instead?': '_',
}
@sl-service-account
Copy link
Author

Teager commented at 2015-12-18T02:14:01Z

Fully support and agree. Points:

  1. The addition of these facial bones is absolutely pointless without bone translations. The only facial animation you can make well without translations is opening the jaw, which we could already do.

  2. The joint positions set in the mesh uploader do not load reliably and never have. Setting deform animations inside a mesh fixes issues like these - https://gyazo.com/3b06ca6a0ac4ad4cf539e988931ad313

  3. Moving a bone is just a natural part of animating. Shifting hips or shoulders, creating lift and weight - these are normal and necessary for creating animations that don't look like stiff barbie dolls.

@sl-service-account
Copy link
Author

kitsune.shan commented at 2015-12-18T03:22:50Z

Just added my vote to it. I want to add that I really want to see this FULLY implemented and not just with sort of "hacks" and strange workarounds to sort something that should be officially supported under BVH files rather than the unknown and custom file formats that no one is able to edit or save without custom plugins and scripts. We have been making always animations on Poser and Poser already allows bone translations. We really need to be able to workwith BVH files for translations as well.

@sl-service-account
Copy link
Author

Tornleaf commented at 2015-12-18T03:22:58Z

Without the use of additive blendshapes or shape keys, I agree the translation of facial bones is vital to decent facial animation. I support this jira!

@sl-service-account
Copy link
Author

Medhue Simoni commented at 2015-12-18T04:05:15Z

I'll just add, that with bone translation, we can likely eliminate a bone in the new wings, because 1 was added to allow for the same type of movement.

@sl-service-account
Copy link
Author

Ashasekayi Ra commented at 2015-12-18T05:35:28Z

I really hope LL rethinks this bone translation limitation. It makes the facial bones practically useless except for a limited set of facial expressions. As we know, if people aren't able to make a full range of facial animations with the bones, they will continue to use the seriously inefficient and polygon heavy mesh swapping workaround that exists right now.

@sl-service-account
Copy link
Author

MoonHowler Snowpaw commented at 2015-12-18T12:34:00Z

I support this JIRA and everything said above!
Actually we get more RESTRICTIONS than possibilities if no changes will be done in this project. Because useless face bones and not resizable limbs and even no workarounds allowed to fix this forever condition of SL (limited, hard to create) will stop most of us for making things we were planning, will force us to work very hard and fast before these deadly changes will take place. Honestly, i'm afraid about what is happening. I can't think about testing new features, giving feedbacks, i have to think right now about how to save my work from being broken, about finishing updates which may be cut at every moment. Please, don't disable bone positions. At least this will give us chance to rerig our volume and attachment points to new bones (if they really exist in the avatar, some of them - no, especially for pair volume bones), at least some of work will not end up in trash.

@sl-service-account
Copy link
Author

Drazion commented at 2015-12-18T15:21:06Z

As a recent buyer of the breeders choice avatars and with the new facial bones already added. Be limiting the animations of the facial bones by removing the translation functions is absolutely absurd and once again attempt to limits the possibilities that creators can achieve.

On a personal note lindon labs you have got to stop trying to limit the potential that creators are trying to achieve here otherwise sl is not really sl anymore where "everything is possible"

@sl-service-account
Copy link
Author

Drazion commented at 2015-12-18T15:22:21Z

I too also support this JIRA.

@sl-service-account
Copy link
Author

Llweyndwell commented at 2015-12-18T16:50:31Z

My biggest concern about bento is how it will affect a lot of my current investments as a customer. Many of which make use of these animations people are talking about that will end up blocked. Should ANY code be changed that would require an update fix to any of those avatars, they essentially are null and void and let me tell you... these avatars aren't cheap. They've gone WAY up in cost compared to past avatars that didn't utilize these methods to be what they are. To move as realistic as they do currently etc. I'll be extremely.. angry should this happen and honestly it would make me rethink purchasing as much which in turn, hurts creators on grid. As a customer, most people aren't aware of what is supported and isn't supported, having the company break those functions just because they can when it hurts nobody is beyond ridiculous.

I was initially extremely excited to see the news of bento on facebook, having gone to my friends whom I know create avatars all excited, they where quick to point out the part they've dreaded from the last JIRA on the subject and it made my heart sink.

I really don't want to see my creature avatars become broken mesh, please rethink this.

@sl-service-account
Copy link
Author

Aladar Horsforth commented at 2015-12-18T18:06:17Z

yes I totally agree I think breaking things is a very bad business decision why can't you guys work with us instead of against us you want SL to be something that people want to use don't break the avatars I agree with what was said above

@sl-service-account
Copy link
Author

Teager commented at 2015-12-18T18:37:03Z

As I understand it, the proposed changes should not affect existing content. They will only affect future uploads.

@sl-service-account
Copy link
Author

ObviousAltIsObvious commented at 2015-12-18T18:45:11Z

If anyone has sample content on the beta grid, showing off what translation can add to face expressions, it would probsbly help to show it off here.

@sl-service-account
Copy link
Author

MoonHowler Snowpaw commented at 2015-12-18T19:12:36Z, updated at 2015-12-18T19:14:43Z

If they at least gave some time for half finished projects. It actually stops in the middle those who are working on updates. Wonderful choice - rebuild it all or forget. What is some weeks... i'm spending right now hours and thousands of lindens just uploading everything that i'll need for finishing planned updates... and won't sleep untill i make all planned for a month or two. Give us at least time to finish things which we were working on before forcing to rebuild the things on that new skeleton! Or AT LEAST don't block viewers still capable to upload .anim files with bone positions and meshes with rigged volume bones and attachment points.

@sl-service-account
Copy link
Author

Whirly Fizzle commented at 2015-12-19T02:35:12Z

Discussion about bone-translating animations from the third party viewer meeting: https://youtu.be/BkWdwhXbhdc?t=26m42s

@sl-service-account
Copy link
Author

Adeon Writer commented at 2015-12-19T20:48:01Z, updated at 2015-12-19T20:48:42Z

In response to the video:

Linden, while you blocked bone translations on Aditi when uploaded as BVH, you didn't block them when uploaded via .anim. The AvaStar/Blender workflow outputs everything as .anim, so many people exersizing the Bento skeleton won't even know bone translations are blocked until it's released proper. So the feedback your getting may have a lot of false positives - people thinking everything's working fine because their translations aren't getting dropped. I just want to make sure you're getting accurate feedback. The blender workflow isn't forced to use your new system exclusively and most aren't even aware of that.

@sl-service-account
Copy link
Author

Adeon Writer commented at 2015-12-19T21:10:42Z, updated at 2015-12-19T21:23:53Z

For further discussion about this and all things Project Bento, feedback goes here: https://community.secondlife.com/t5/Building-and-Texturing-Forum/Project-Bento-Feedback-Thread/td-p/2987290

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

No branches or pull requests

1 participant