This work studies the performance of different signal features regarding the qualitative meaning of Laban Movement Analysis semantics. Motion modeling is becoming a prominent scientific area, with research towards multiple applications. The theoretical representation of movements is a valuable tool when developing such models. One representation growing particular relevance in the community is Laban Movement Analysis (LMA). LMA is a movement descriptive language which was developed with underlying semantics. Divided in components, its qualities are mostly divided in binomial extreme states. One relevant issue to this problem is the interpretation of signal features into Laban semantics. There are multiple signal processing algorithms for feature generation, each providing different characteristics. We implemented some, covering a range of those measure categories. The results for method comparison are provided in terms of class separability of the LMA space state.