The first answer is drive Ghostscript directly, don't use ps2pdf (or pdf2ps).
If you are getting text converted to an image, then its most likely because the original PDF file has transparency, which cannot be represented in PostScript. The only way to deal with that is to render the area of transparency.
There is no way to maintain the Encoding of the text, though in general it won't change. However this is highly dependent on the font and encoding used in the input. I can't say more without seeing an example.