University of Southern CaliforniaUSC
USC ICT TwitterUSC ICT FacebookUSC ICT YouTube

Better error checking on joint maps? | General SmartBody Discussion | Forum

Avatar

Please consider registering
guest

sp_LogInOut Log In sp_Registration Register

Register | Lost password?
Advanced Search

— Forum Scope —




— Match —





— Forum Options —





Minimum search word length is 3 characters - maximum search word length is 84 characters

sp_Feed Topic RSS sp_TopicIcon
Better error checking on joint maps?
September 26, 2014
10:56 am
Avatar
Member
Members
Forum Posts: 8
Member Since:
September 26, 2014
sp_UserOfflineSmall Offline

Hi... again (sorry a few questions, which I thought I should split into separate posts...)

I have been trying to get one of my skeletons to work in SmartBody, however it is a very old and bad rig! As a result it is missing some key joints needed for some controllers. I know you have documented some of these key joints in your docs, which by the way are very good :), however some are missing like 'Locomotion'. At the moment the application will crash, it would be better if it could do the following;

A) - print warning to log when a bone is not found

B) - If a bone is not found, just ignore it!?

Again, happy to make changes with some guidance, although again would prefer Mecurial!

October 3, 2014
8:48 pm
Avatar
Admin
Forum Posts: 983
Member Since:
December 1, 2011
sp_UserOfflineSmall Offline

Can you point me towards the portion of the code that is having that problem?

Yes, your suggestions are good.

 

Ari

October 6, 2014
4:12 pm
Avatar
Member
Members
Forum Posts: 8
Member Since:
September 26, 2014
sp_UserOfflineSmall Offline

Looking at this in more detail, it might not be a simple as I thought!

The error is occurring in the 'pose post processing' controller, as one of the constraints on a Ik tree does not exists in the map. The error occurs in "MeCtIKTreeScenario::updateEndEffectorValidNodes" because my character does not have a l_toe or r_toe. (Infact he does not have feet!)

So my new question is, why do the internal algorithms try to access the joint map? Surely the should update the internal SB skeleton, then map the joint transforms to the custom skeleton at the end? 

October 6, 2014
11:34 pm
Avatar
Admin
Forum Posts: 983
Member Since:
December 1, 2011
sp_UserOfflineSmall Offline

Yes, there is some refactoring that needs to be done. The SmartBody skeleton should only be used for internal purposes. They shouldn't be accessing the joint map directly. I try to fix those things when I have time.

October 7, 2014
9:36 am
Avatar
Member
Members
Forum Posts: 8
Member Since:
September 26, 2014
sp_UserOfflineSmall Offline

Ari Shapiro said
Yes, there is some refactoring that needs to be done. The SmartBody skeleton should only be used for internal purposes. They shouldn't be accessing the joint map directly. I try to fix those things when I have time.

Ok thank you. Could I help you with this, or is it just easier if you do it?

October 7, 2014
10:56 pm
Avatar
Admin
Forum Posts: 983
Member Since:
December 1, 2011
sp_UserOfflineSmall Offline

There are a lot of places in the code that need to be tweaked; best done by my team.

 

Ari