In this paper we will describe some new algorithms concerning pitch syncronous analysis, endpoint detection for noisy speech, pitch separation for co-channel speech, and sampling rate modification, all of them having been developed in the framework of the sinusoidal model. Such algorithms, in addition to the many already proposed in the literature, further demonstrate that the sinusoidal model is quite powerful and it can be used as a versatile and high performing tool in many speech processing applications. The robustness against noise shown by the developed algorithms, moreover, indicates that the sinusoidal model can be used in applications critical to noise.
