It's very difficult to make out the words, regardless of the overall volume. Try feeding your recording through an equalizer and (probably) cutting around 500Hz and maybe boosting somewhere between 3k and 6kHz. A lot of energy in the 500Hz area give sound a nasal, boxy quality, while the intelligibility of speech falls somewhere between 3k and 6k. If you can turn the scales on those frequency ranges, you'll clean up just about any vocal recording in giant leaps.