Indirect combustion noise is a growing concern for aviation engine designers. It is typically associated with the distortion of “hot spots” (entropy structures) and vortical structures that generate excess noise as they are convected by an accelerating mean flow. Recently, it has been shown theoretically that fluctuations in a gas’s chemical composition can act as an additional source of indirect noise in aviation engines. This work examines this composition noise mechanism, focusing on the underlying chemical effects that drive this source of indirect noise. Since the mechanism has yet to be confirmed experimentally, this paper begins by applying the theory to inert mixtures of noble gases and air, in an attempt to guide experimentalists by identifying the set of operating conditions that will produce the strongest acoustic response. Turning from non-reacting to reacting flows, the paper next examines the sensitivity of the compositional noise mechanism to fuel type, testing several common fuels. It is found that, while there is a substantial difference between hydrogen and hydrocarbon fuels, overall noise levels vary only slightly between different hydrocarbon fuels. Additionally, there appears to be a common underlying structure to the response of a product-gas mixture generated by burning a fuel, which is explained through linearized theory and confirmed with numerical results. Lastly, the physics of composition noise is examined at the species-specific level, attempting to provide a link between individual combustion products and changes in a mixture’s propensity to generate indirect noise. The sensitivity of individual species can be explained by a combination of differences between the species and mixture’s Gibbs free energy and strong gradients in product gas concentration with mixture fraction. However, by analyzing the species dependency of combustion products at several different mean mixture fractions, it is found that no single species dominates the noise generation over the combustor’s entire range, but rather the most acoustically active species varies strongly with local stoichiometry.