How to recover (by deconvolution) stark broadened line profile from observed spectral line, where the FWHM of experimentally observed spectral line is a convolution of stark broadening and instrumental broadening?

I am working on plasma spectroscopy. I have recorded emission line spectrum of a laser induced plasma. In which a spectral line (lorentzian profile) has FWHM that is a combination of stark broadening and instrumental broadening. However, in calculations I need pure stark broadened width, therefore, I need to recover it that I read should be done through deconvolution. But, I have no idea how to do that and I hope someone here can help me. Cheers!