The Composite InfraRed Spectrometer (CIRS) on-board the Cassini spacecraft has currently returned around three years worth of data from Saturn's largest moon Titan. One of the unique aspects of CIRS is to take high spatial resolution spectra of the limb of Titan, with sub-scale height (20-40 km) resolutions. This is made possible by the small field-of-view (FOV) of the mid-IR detectors. However, many limb spectra have moderate to large sized FOVs, which introduces bias into retrieved profiles of temperature and abundance unless the finite FOV size is taken into account. The bias can be reduced by calculating a FOV-averaged spectrum comprising a weighted sum of a small number of spectra with infinitesimal FOVs across the FOV. Here we introduce a scheme for incorporating FOV averaging into radiative transfer calculations of CIRS spectra and quantify the errors as a function of number of FOV averaging points, FOV size, tangent altitude, and wavenumber. The optimum number of FOV averaging points for a given observation can then be found by matching the calculated FOV averaging error with the measurement error. This allows for accurate analysis of a vast amount of Cassini-CIRS data. © 2007 Springer Science+Business Media B.V.