CCL:G: Time dependent properties

 Sent to CCL by: Ramon Crehuet [rcsqtc(_)]
 Dear CCLers,
I was helping a colegue in the interpretation of an absorption spectra. I used TD-DFT for the singlet states. Now she asked me to consider a low-lying triplet. I checked and saw that Gaussian can do triplets and singlets with the option 50-50. If I am not wrong, this singlet-triplet transitions are only allowed by the spin-orbit coupling, therefore a 'normal' calculation will always give a null oscillator strength. Am I right about that? What would then be the easyiest way to proceed from here?
 Thanks in advance,