Laboratory fire tests were performed in still air, for variable inclinations (10°, 15°) and fuel bed dimensions (1.28 x 2.50-3.0 x 4.6 m2), with homogeneous fuel beds of pine needles and pine wood excelsior. The fire ignition was made at a point, along a closed line with no fuel inside and along a straight edge of the fuel bed. The tests were recorded with an infrared camera and various techniques were developed to implement direct and indirect empirical methods of construction of the 'orientation function' of the fire in the given fuel bed and ambient conditions, showing how the rate of spread of a steady straight fire front depends on its orientation on the terrain. The direct method uses a set of straight fire fronts with various orientations whereas the indirect method uses essentially a point ignited fire front. Contrary to what is assumed in BehavePlus model, the orientation function is observed to depend significantly on the properties of the fuel bed. In all tests with closed fire fronts, the full steadiness required by the indirect method was not achieved (namely, at the head of the front), although, for slope angle of 10°, the corresponding errors induced on the orientation function were small.