how to calculate SNR for speech signals?

I have two wav files of noisy and clean speech...i dont have separate noise signal file...how to calculate SNR from noisy and clean files(without knowledge of noise)?

Does this SNR also apply for speech added to speech?

Please help me...thanks

[Gregory jones]

For anyone who perhaps is just finding this like myself a potential solution is to find the energy let's say Ps of the speech signal.

Then try find some silence in the signal where just noise is present, find the energy here Pn.

The Ps & Pn window size should be the same, the larger the better I believe.

Then it's simply

SNR=10log10((Ps-Pn)/Pn)

I think at least