The ferromagnetic linewidth obtained via different rectangular cavity ferromagnetic resonance experimental data processing algorithms has been compared. Approaches based on perturbation theory closed-form formulas, resonance frequency dispersion, Q-factor, transmitted power, and an electrodynamic transcendental equation (TDE) have been considered. The results were compared with direct Q-factor measurements in a subwavelength cavity and an electron paramagnetic resonance spectrometer. Three spherical samples, of approximately 1 mm in diameter and different saturation magnetization values, were investigated: monocrystalline gallium-substituted YIG (Ga:YIG, 50 Gs), polycrystalline aluminum-substituted YIG (Al:YIG, 380 Gs), and polycrystalline zirconium–calcium-substituted YIG (Zr,Ca:YIG,1826 Gs). A new approach to processing rectangular cavity data, based on the TDE, has been proposed and validated. Using the proposed approach, it is possible to obtain the accurate value of the linewidth in regimes where strong coupling between the sample and the cavity does not occur but the small perturbation assumption is no longer valid either. There is also no need to decouple the sample from the cavity. The conclusions of this work are useful for resonant characterization of not only magnetic spheres but also other shapes.