Biped Character Rig Part 11 - Creating the Upper Arm Twist

This lesson replaces the upper arm’s earlier twist constraints with explicit connections from the elbow’s X rotation. The previous setup responds to the intended twist, but some other elbow rotations also turn the intermediate joints. The instructor rebuilds that section, tests each rotation direction, and repairs a control connection accidentally broken during node editing.

The starting scene already contains the arm joints, three upper-arm twist joints and controls from the preceding biped lessons. Joint orientation and parenting are part of that existing setup. The exact Maya release is not stated, and the recording tests joints rather than a skinned character.

Establish the unwanted rotation before removing constraints

During 00:00–03:03, the instructor compares the forearm and upper-arm responses. The forearm behaves acceptably in the demonstrated tests, while the elbow-driven upper-arm arrangement introduces twist during rotations that should serve other purposes.

The three upper-arm twist joints are located in the Outliner, expanded, and their existing constraints deleted. Retesting now produces no driven twist. That temporary absence is expected: it confirms the old driving relationships have been removed before new ones are introduced.

The hierarchy also changes. Twist joint three is placed beneath two, and two beneath one. The end of this new chain is not connected onward to the elbow. Record this change when comparing the two approaches; the lesson changes both the driver mechanism and the relationships between the twist joints.

Align the joint axes before wiring rotation channels

During 03:03–05:10, the instructor inspects the joint rotations and axes. A trial of Freeze Transformations does not produce the intended result, so the rotations are zeroed and the axes examined again.

The Orient Joint operation is then adjusted. The source uses a negative-Y orientation choice to make the displayed axes agree with the rest of this particular arm. It verifies the three joints again and checks that their rotation channels are zero before saving.

The negative-Y choice is not a universal arm-rig prescription. Its purpose here is to align the demonstrated joints consistently. Driving rotateX only makes sense when X represents the intended twist direction in the actual joint setup. Inspect that axis relationship before copying the later connections.

Feed the elbow rotation through three multipliers

During 05:10–11:00, the Node Editor is cleared and populated with the relevant elbow and twist joints. Three multiplyDivide nodes are created. The intended path is from the elbow’s rotateX output into each node’s input1X, then from each node’s outputX into a twist joint’s rotateX input.

The demonstrated factors are 0.25, 0.5 and 0.75 for the three driven channels. At the individual node level, each factor scales the incoming value. For example, a factor of 0.25 produces one quarter of the input value on that node’s output.

Those factors describe local channel inputs. The joints were also placed in a serial parent chain, so inherited transforms contribute to their resulting orientations. The source does not measure world-space twist fractions or demonstrate skin deformation. Do not read the three factors as a verified deformation distribution along a finished arm.

Several incorrect connection attempts appear during this stage. The instructor reverses a connection direction, deletes an unwanted connection and encounters unit-conversion nodes. Follow the intended source-to-destination attributes, not the temporary graph state from an unsuccessful attempt. The recording does not establish a general rule that conversion nodes should be removed.

Diagnose a missing upstream driver

From 11:00–13:08, testing shows that the unwanted bend-related twist has stopped, but the desired twist is also absent. The instructor first checks the control, then the elbow and graph, rather than immediately changing multiplier values.

The earlier reversed connection attempt had disrupted the elbow’s X-rotation driving relationship. A multiplier cannot produce the intended result if its input no longer changes with the control.

The instructor deletes the affected elbow constraint and rebuilds the control-to-elbow parent constraint, selecting the control as target and the elbow as the constrained object. The captions do not establish the complete option-box configuration clearly enough to serve as a reusable preset. The important supported repair is restoring the missing upstream control relationship and confirming that the joint remains aligned.

Test the three directions separately

During 13:08–14:03, the instructor retests the arm. The demonstrated bend directions no longer introduce the unwanted twist, while the intended twist direction drives the intermediate joints. The forearm is also checked again to confirm its earlier behavior remains present.

Use the same order of inspection: control motion, elbow rotation, multiplier input/output, then twist-joint response. This separates an upstream broken constraint from an incorrect graph connection or axis assumption.

The result is a more satisfactory joint-level test for this scene. It is not independent validation of the multiplier distribution, behavior under all poses, or a skinned arm’s deformation. Those remain technical review questions before adopting the setup in a production character.