Two dynamic flux chambers for direct measurement of odorous compound emissions from quiescent liquid surfaces were investigated under simulated conditions in the laboratory. Initially, a flux chamber built according to the model recommended by French standard NF X 43-104 was studied. This chamber was used in two different ways. The first led to a lack of precision concerning emissions rates from the sampled liquid surface, whereas the second led to an overestimation of the measurements. The second part of the study was devoted to an improved dynamic flux chamber, built according to the feedback from the results obtained using the normalised sampling system. Laboratory tests showed good accuracy and precision. This work underlines the importance of the aerodynamic performances of a dynamic sampling system.