Distribution of pixel-based Hα/Hβ line ratios as a function of observed (top) and extinction corrected (bottom) Hα surface brightness (SB_Hα) for three galaxies of our sample (see text). We employ a S/N cut of 10 for both lines. Most pixels have Hα/Hβ >  2.86 (dashed line, as predicted by case B recombination for T_e = 10⁴ K, n_e = 100 cm⁻³), and those that have ratios below this value are consistent within the errors. Because of this, when de-reddening Hα we apply no correction for those pixels with Hα/Hβ < 2.86. At low Hα surface brightness, our Hβ sensitivity sets an upper threshold on the recoverable value of Hα/Hβ. The value of this threshold for the median Hβ error is indicated in all panels by the dotted lines. Histograms of SB_Hα are included for each set of plots for pixels that meet our S/N cut (in black) and those that do not (in grey). A significant number of pixels are excluded by requiring an S/N >  10 detection of Hβ. A wide range of extinctions is observed, with no obvious correlation with SB_Hα. This demonstrates the need for matched resolution constraints on the dust extinction. Within PHANGS-MUSE, the use of the Balmer decrement provides a unique opportunity to perform robust, extinction-corrected SFR measurements across the bulk of the galaxy disc.