We report the experimental observation of spectral interference in a Michelson interferometer, regardless of the relationship between the temporal path difference introduced between the arms of the interferometer and the spectral width of the input pulse. This observation is possible by introducing the polarization degree of freedom into a Michelson interferometer using a typical weak value amplification scenario. c 2018 Optical Society of America OCIS codes: 260.0260, 260.3160,260.5430, 320.5550,120.3180 Interference is a fundamental concept in any theory based on waves, such as classical electromagnetism or quantum theory. The specific experimental arrangement required for the observation of interference depends on the characteristics of the light source, i.e., its spatiotemporal profile and its degree of coherence. For example, for first-order coherent light in a Michelson interferometer, for temporal delays shorter than the pulse width, interference manifests as a delay-dependent change of the intensity at the output port of the interferometer. For longer temporal delays, interference manifest as spectral interference for a given temporal delay. The observation of spectral interference was denoted by Mandel [1, 2] as the Alford-Gold effect [3] and it is well-known in optics [4].Here we report the observation of spectral interference independently of the temporal regime under consideration. The interference is revealed as a reshaping of the input spectrum that is accomplished by introducing the polarization degree of freedom into a Michelson interferometer. This scenario corresponds precisely to the conditions of a typical weak value amplification configuration [5][6][7][8] that although was originally conceived in the framework of a quantum formalism, it is essentially based on the phenomena of interference and can thus be applied to any scenario with waves [9][10][11][12][13].For the sake of clarity, let us start by describing temporal and spectral interference in a typical Michelson interferometer, without considering polarization. Later on, we will describe the effects that the introduction of the polarization degree of freedom has on spectral interference. Consider the situation depicted in Fig. 1(a). A first-order coherent input pulse with amplitude E 0 , central frequency ν 0 , input polarization e in , and temporal duration τ (full width at half maximum) described by follows a different path and after reflection in a mirror the two pulses recombine at the BS. By changing the position of one of the mirrors, a temporal delay (T ) is generated between the two pulses. The intensity measured by a slow detector at the output port of the interferometer as a function of T can be written aswhere I 0 = |E 0 | 2 . Two interesting cases can be distinguished: i) when T ≪ τ , and therefore, the two pulses traveling the different paths overlap in time and ii) when T ≫ τ and the two pulses do not overlap. In the first case, the output intensity as a function of T reduces to1