We show a practical way for building wideband phase-shifting algorithms for interferometry. The idea presented combines first-and second-order quadrature filters to obtain wideband phase-shifting algorithms. These first-and second-order quadrature filters are analogous to the first-and second-order filters commonly used in communications engineering, named building blocks. We present a systematic way to develop phaseshifting algorithms with large detuning robustness or large bandwidth. In general, the approach presented here gives a powerful frequency analysis and design tool for phase-shifting algorithms robust to detuning for interferometry. © 2009 Optical Society of America OCIS codes: 120.3940, 120.3180, 120.2650 In this work, we present an easy and straightforward technique for designing temporal phase-shifting (TPS) algorithms with large detuning [1]. It is wellknown that TPS algorithms may be regarded also as quadrature filters tuned at a single temporal frequency. The tuning frequency is the temporal carrier of interferograms, which in TPS interferometry parlance is the phase step used to obtain the interferograms. Here we have adopted a filter construction strategy typically followed in communications engineering, which is to build larger order quadrature filters based upon simpler building blocks, namely, first-and second-order filters. The main advantage of adopting this strategy is that these lower order discrete quadrature filters may be optimally located in the frequency space to obtain a quadrature filter with large detuning robustness (large bandwidth). In interferometry, interferometric data are obtained as an image called an interferogram. The temporal sampling of interferometric data at a site ͑x , y͒, can be modeled as a periodical signal s : R → R in the following way:where a x,y R and b x,y R are the dc and contrast term at site ͑x , y͒, respectively. x,y R is the unknown phase in that site, and 0 R is the linear phase shifting or temporal frequency carrier; t is the temporal sampling. The problem here is the following: given a signal like the one shown in Eq. (1), find a way to recover the unknown phase x,y . On this topic we can find several works that deal with this problem, to mention some of them we can cite [2-10].Taking a look at previous works around linear phaseshifting algorithms, we can see that most of the algorithms were developed intuitively, or systematically using least squares [2,4,5]. Actually, to our knowledge the only work to describe the phase-shifting algorithms in the Fourier domain is that published by Freishlad and Koliopoulos [11]. Here, we will show how to design phase-shifting algorithms in the Fourier domain.As we said at the beginning, for the TPS demodulation problem we have adopted a filter construction strategy based on simple building blocks or filters. Then, let us start by defining the following first-and second-order basic building blocks:
͑3͒which are a first-and second-order difference operator, respectively, where ␦͑t͒ is the Dirac delta function, an...