[Question] Differences to Filo Cable Simulator

[spaceemotion]

I've stumbled across your "Filo Cable Simulator" asset and it states that for more advanced stuff ObiRope should be used.

What's a bit confusing for me though, is that in Filo the cables are looking quite strong and rigid, while in ObiRope the demos let them look quite loose (regarding the demos with cogs and wheels). 

Is there a setting in ObiRope that gives them a similar look?
Is Filo more performant than ObiRope?

Since ObiRope is now even cheaper than Filo, would you even suggest buying Filo instead?

Thanks in advance!

[josemendez]

I've stumbled across your "Filo Cable Simulator" asset and it states that for more advanced stuff ObiRope should be used.

What's a bit confusing for me though, is that in Filo the cables are looking quite strong and rigid, while in ObiRope the demos let them look quite loose (regarding the demos with cogs and wheels). 

Is there a setting in ObiRope that gives them a similar look?
Is Filo more performant than ObiRope?

Since ObiRope is now even cheaper than Filo, would you even suggest buying Filo instead?

Thanks in advance!

Hi,

Filo and Obi have completely different approaches to simulation, and their use cases are very different too.

Filo focuses on the effect a massless cable has on the objects it touches. Basically it creates a single constraint between each object pair, and keeps track of the cable length between objects. It does not simulate the dynamics of the cable itself, and assumes a cable only affects objects when completely tense. This is great for cable-driven machinery, where cables are always (or almost always) in contact with the same set of objects, and you're only interested in the overall dynamics and interaction of objects as a result of being in contact with each other trough the cable.

As a result, Filo cables cannot collide with objects outside a predefined list, cannot simulate free cables (they must be attached to at least 2 objects), contacts always use full friction, the cable cannot collide with itself, and can't simulate torsional effects.

Obi on the other hand simulates a rope/cable as a chain of mass objects (particles) and their interaction with the enviroment. Free cables are fully supported, full contacts with adjustable friction/adhesion, self-collision, torsional effects, collisions with arbitrary colliders, etc.

Performance wise, Filo is much more performant simply because it does way less work. Performance in Obi mostly depends on the length of the rope (more rope = more particles = more calculations to do). In Filo, performance depends on the amount of objects joined by the cable. The cable length has no effect at all on performance.

Which one to get depends entirely on your use case. Obi is much more general than Filo, but fails -quite miserably- at the use cases Filo is designed for (complex cranes, lifts, tow systems, etc).

FYI, Filo is based on this research article:
https://matthias-research.github.io/page...Joints.pdf

Obi is (loosely) based on this one:
http://blog.mmacklin.com/project/flex/

[spaceemotion]

Thanks a bunch for the lengthy and detailed reply!

Would you say the Filo + Obi assets can happily co-exist in the same project?
It does sound like Filo is something we'd use for rope/chain puzzles then.
Just making sure, in case we do want to add one of the Obi assets later on.

[josemendez]

Thanks a bunch for the lengthy and detailed reply!

Would you say the Filo + Obi assets can happily co-exist in the same project?
It does sound like Filo is something we'd use for rope/chain puzzles then.
Just making sure, in case we do want to add one of the Obi assets later on.

Yes, both can coexist, but will not interact with each other. Filo cables will ignore Obi ropes and viceversa, as they are "unaware of their coexistence", so to speak.