Till today motion and motion capturing has almost always been explicitely or implicitely thought of and described in context of a 3D-Space-Framework ("move of an object from A to B").
“The term “motion capture” refers generally to processes that capture movement of a subject in three-dimensional (3D) space and translate that movement into, for example, a digital model or other representation.” (us patent applicaction US9070019)
The theoretical foundation has been laid by Euclid (Three Dimensional Euclidean Space, 3rd century B.C.) and René Descartes (Cartesian Coordinate System, 17th century A.D.). It usually describes the motion of a rigid body in a linear way from point A to B. More technical, In a simple 3D space there are six degrees of freedom - all relative. Those are left/right, up/down, fore/aft (called xyz-translations), pitch, yaw and roll (called xyz-rotations). Including absolute positioning, it becomes nine degrees of freedom.
The above concepts work with a rigid body which is theoretically 100% unbreakable. As soon as it comes to introducing human beings all these assumptions can't be fully hold anymore. An average adult person consists of around 60% liquid going up to 80 for a baby. It requires a skeleton to form the structural shape, that an average person does not look like an ordinary water bag. The skeleton of a baby consists of more than 300 bones droping due to a fusion process to around 206 bones in an adult body. Bones are building the structural support for muscles and other soft tissues. Around 650 skeletal muscles can be found in a body. As an aside, on each side of the body twenty-seven bones are found in wrist, hand and fingers. More than 30 skeletal muscles are found in the region working together to achieve diverse movements, including non-linear ones.
It becomes obvious, that motion capture can't be explained by a 3D-Space-Framework alone, as deformations on the surface of the skin caused by muscular, tendon or nerve reflexes due to intended or unintended motions do regular occur. Such dynamics, micro and/or slow movements do obviously require a separate categorisation. The motion pictures and computer graphics industries have been long aware of those facts and coined the term "Soft Body Dynamics" for such motions already in 1985/86.
Traditional motion capture as opposed to soft body dynamics capture can be demonstrated idealy on wrist applications. The wrist is the traditional place for various devices including watches and smart watches. Given the structure of the wrist it is one of the most trafficked region of the body with regard to soft body dynamics. This makes any (vital signs) monitoring a complex task not only with regard to motions.
If the forearm is moved in a tangential direction (xyz-translation) or rotated (xyz-rotation) this motion(s) can be readily captured by any state of the art accelerometer and gyroscope. The location of these sensors is usually in a rigid device housing placed at the wristsurface. If a tangential movement of the forearm can be clearly interpreted e.g. as a swing while making a step algorithm can transform this information into a step count. If the motion is fluttery and therefore not interpretable usually such motion can only be used as artefact and noise information. Prime signal calculations such as for pulse wave frequencies may be interrupted as results can not be properly interpreted.
Now, if the forearm is left constant, but the hand is bend backwards, the position of the accelerometer and the gyroscope hardly changes as the position of the forearm with the 3D-Space-Framework remains unchanged with regard to the xyz-translation as well as the xyz-rotation. As a result, hardly any signal is captured by those sensors. However, due to a build up of muscle and tendon strain the wrist circumference alters its shape and diameter, thereby distorting most prime signal measurements such as for pulse wave frequencies. Already the faintest tip of a finger causes a change, as muscle strands go up to just below the elbow. Wearing a soft condensed matter sensor system around the wrist, these soft body dynamics, are captured as the strain level changes. Motion and strain level allow for interpretation of various dynamics.
Building smart systems, there is not the question of either or, there should be the question what built gives me the best recordings with regard to motion in a 3D-Space-Framework and Soft Body Dynamics, motions which can't be explained by the traditional concept.