In this study, the objective was to compare different sampling devices used, on solid or liquid sources, to measure odour emission from passive sources. Experiments were firstly carried out on two real sources: a solid source made of compost and a liquid source corresponding to a basin containing leachates from the composting piles. In a second time, experiments with some selected sampling devices were also carried out with pilot units. For experiments on real sources, samples were coded before analysis in order to be anonymous for olfactometric measurement, based on EN13725 standard and performed by laboratories under French accreditation (COFRAC). For laboratory test, some devices and their efficiency were tested on a liquid source and/or a solid source. The pilot unit for liquid source (at the site of Ecole des Mines d'Ales) is a 150L tank containing a solution of n-Butanol used as tracer. The pilot unit for solid source (at the site of INERIS) is sand square where the porosity of such a ground is used to diffuse Methane used as tracer. For both cases, direct analysis by FID was carried out joining the outlet of the device to the detector. This study shows the high complexity of such a sampling because devices (chambers) give results in a large range with odor flow that are 10 times, 100 times higher or even more for some chambers comparatively to the lower result. Because, the source was the same, the Nalophan bags came from the same production, the period of sampling was identical and the same laboratory analyzed samples, the major factor of difference was typically the chamber itself because Olfactometry uncertainty cannot explain so huge differences. The difference can be limited excluding chambers with high flow rate. In that case, because the odor flow is the factor of odor intensity and flow rate, this last one is very influent on the result. If a higher limit of flow rate seems obvious, a lower limit appears too, because low flow chamber give high concentration due to accumulation inside, and then low odor flow due to lower impact of the flow rate in calculation.