We do live referencing here as we don't have much for versioning control at this small studio. Its worked fine for us so far as our rigs are pretty simple and eventually development stops for them on a project.
Eventually we might do versioned referencing but it would require some asset tracking tools that we just don't have. We also don't have enough coordinators to keep track of what shot gets what rig, or the development man power to update our pipeline. Tiny shop in Singapore, very difficult to find talented developers. Until then I've written a bunch of delta cleaning scripts that can solve 90% of our issues, and what was most important for us, was that we are very careful about how we import referenced models so that no unnecessary information gets written to the delta. So if a rig change does happen, the delta wont freak out and break the rig. Basically via scripting, I create an empty referenced model, populate the paths for the rig resolutions, add an empty delta, set the proper settings, disabling the evil ones, and then finally setting the active resolution. I have found this technique to be absolutely essential to making sure our deltas don't store any extra nonsense in them that would eventually clash with a rig change. Otherwise if the animator simply did a vanilla "import referenced model" this referenced model would have a delta on it thats fully enabled, and it would immediately store a bunch of useless crap under stored expressions/stored positions that would eventually cause the rig to explode when we did the next rig update. Aside from that, when something bad happens, I run one of 3 delta cleaners if necessary, but to be honest it hasnt happened much in a very long time. I am very open minded, but i also know the animators at this company very well. If I gave them a local model pipeline they would bring this studio to its knees and I would be out of a job very quickly On Wed, May 22, 2013 at 4:51 PM, Raffaele Fragapane < raffsxsil...@googlemail.com> wrote: > That's only an issue with live referencing, which is a pretty bad way to > go about it. > Versioned referencing will not hold any surprises as you control when a > rig reaches what shots. It's a matter of asset management, not of > referencing VS localized. > If anything localizing takes a liberty away from you, it doesn't ADD > security :) > > > On Wed, May 22, 2013 at 6:36 PM, Sandy Sutherland < > sandy.mailli...@gmail.com> wrote: > >> I did - and we locked everything that was not supposed to move/key. >> Worked fine and it is not so difficult to write the tools to do a model >> update. Also one BIG plus to this method against referencing is that you >> avoid surprises when something changes in the rig that affects animation >> and it gets auto updated in a reference, THAT has happened to me before, >> largely due to the nature of the overlapping schedules we work with in SA. >> >> S. >> >> >> On 22/05/2013 04:48, Enrique Caballero wrote: >> >>> Jeremie, I considered letting them animate in local mode but i decided >>> that the risks outweighed the benefits. >>> >>> I am just 100% uncomfortable with trusting the animators with Local >>> models. They will start deleting objects and changing heirarchy. >>> >>> So instead I stripped down the rig of unecessary stuff. I took all the >>> cloth controls away and seperated it into a different rig, and then >>> aggressively cut down the amount of keyable params. >>> >>> Its at least manageable now. >>> >>> Are you really comfortable letting the animators animate local models? >>> I don't think I could ever let it happen >>> >>> >> > > > -- > Our users will know fear and cower before our software! Ship it! Ship it > and let them flee like the dogs they are! >
