voice blending not working #3

After applying the changes I suggested in #2 , the voice generation works fine in the web-ui, but if I try to blend two different voices it doesn't work at all (it just generates a random voice). This feature worked correctly in the non-cpp version.

edit: I've just read in the INTERNALS.md file that this functionality is not supported yet ("Differences from the Python server" -> "speaker blending and a default-voices directory are not implemented.")

P.S. great job. btw, this is the best tts Ive tried so far.