The following notes and audio were taken from the weekly Bento User Group meeting, held on Thursday, October 13th at 13:00 SLT at the Bento User Group wiki page.. and chaired by Vir Linden. For details on the meeting agenda, please refer to the
Note that this update is not intended to offer a full transcript of the meeting, nor does it present the discussion points in chronological order. Rather, it represents the core points of discussion, grouped together by subject matter were relevant / possible, whilst maintaining the overall context of the meeting.
The Bento viewer is now on its third version as a release candidate (version 126.96.36.1990160 at the time of writing). It is performing well crash-wise in comparison with the current release viewer, so the hope is – bugs allowing – it might not be too long before the viewer moves up to release status.
The first of the bugs mentioned by Vir in the audio is that the scale locking capability recently added to disable shape sliders from having any influence over joint positions has been reported as working intermittently at times. Vir is still looking into this.
The second is a report instance of pre-Bento content being deformed when viewed on the Bento viewer – see BUG-40672.
The problem appears when a sub-mesh contains bad joint position overrides for various joints that the model is rigged to. Prior to Bento, these joint position overrides were not applied, because the Lab was enforcing a 20-joint minimum: if a sub-mesh was rigged to fewer than 20 joints, then it was not treated as rigged, and the joint positions were ignored.
For Bento, this restriction has been relaxed to make it easier to produce meshes only affecting part of the avatar (wings affecting wing joints positions, head that affects only head join positions, etc.), making it possible to “mix and match” different meshes. Unfortunately, it also means that any model with “bad” (but previously “invisible”) joint positions defined may now appear deformed.
Reverting the viewer to its previous behaviour is not seen as optimal, because of the potential to break some Bento models and make the extended skeleton less useful (no more mix and match of meshes). Similarly, it is currently unclear as to how much content might be affected if the updated behaviour is retained, so the Lab is still considering what to do.
Makers of pre-Bento mesh avatars are asked to check BUG-40672 and then test their avatars using the current Bento viewer to see if similar issues manifest themselves. The most direct way to “fix” things is for affected content being re-uploaded with the bad joint positions removed, or by adding an animation to fix the joint positions after the fact (assuming said animation doesn’t conflict with others the model is using). Of course, if the model does not require joint positions at all then the option could also simply be unchecked at upload time.
A complex discussion has been unfolding on the discussion forum concerning animating hands (from around page 122 onwards), and which also encompassed a part of the user group meeting (in text). It’s a convoluted subject involving several elements, of which perhaps the most important thing is understanding how best to animate hands.
To try to address this, Cathy Foil has produced a comprehensive video on hand animations, which anyone developing mesh hands for Bento may find informative and helpful, and so is embedded below.
Cathy also provided some hints and tips during the meeting.
There has also been forum discussion on feet. Bento doesn’t add and further joints to feet, and there are no plans to change the rig specifically to a support for animating feet. However, as Vir pointed out, there are more joints than the system avatar does not use – mToe and possibly mFoot – which might offer a degree of flexibility for mesh creators. If these are used, a point to keep in mind is that they are not used in the avatar height calculations, so some unexpected results my occur.
Bounding Box Clipping / LODs
A non-Bento issue which can be frequently seen at Bento meetings due to the number of meshes being worn is that of bounding box clipping – part of the avatar appears to vanish as a camera is pulled back.
This generally occurs when the viewer is operating at the limit of its imposter avatars setting, and a conflict occurs trying to both display the avatar mesh (which the viewer will try to render at a united LOD (level of detail) whilst also trying to switch to displaying an avatar as an imposter. It can be overcome by setting the imposter slider to No Limit in Preference > Graphics.
While fixing this issue is not part of Bento, it raised the subject of why LODs aren’t used rather than the imposter setting. Vir explained that one reason is this idea that the viewer will try to render avatars and their attachments at a high LOD, rather than stepping down through LOD models, as is the case with in-world mesh objects. Another isse us that many creators tend to only use high LOD models, and incentivising people to use the LOD system correctly isn’t easy, although there are tools (e.g. Decimator in Blender) which help in the creation of LODs for models.
As the whole issue of rendering and associated costs and calculations is now being looked at by the Lab, changes might be forthcoming to help with matters – time will tell on that.
- Project Bento testing page (wiki)
- Avatar body size (wiki)
- Bento Skeleton Guide (wiki)
- Project Bento user group page
- Project Bento discussion forum
- Project Bento coverage in this blog (including weekly user group meeting updates)