64kbps at 44100Hz stereo is pretty lousy, more so with
MP3 than with
OGG in my experience. So if you want your recording not to deteriorate further, you should likely use at least double of that when going
That generally is a problem: the highest compression ratio dictates the quality of the result, and if the deterioration should not significantly increase, all other lossy compression passes need to have decidedly higher bitrates.
So if your recording is with 64kbps and already has the lowest quality you can really justify, a reencoding with a different algorithm will likely have to happen with 192kbps in order to yield results close to the original 64kbps recording. Not satisfactory.
So as a rule: highest compression/lowest bitrate is the last step before distribution, or your distribution bitrates will be significantly higher than your distribution quality.
Even then, multiple recompression, in particular with different compression algorithms and consequently different compression artifacts, is a bad idea anyway.
That does not mean that you need to record with
flac is lossless compression, so the results are indistinguishable from a
wav recording and the file size still is about a third.
That's the best format to choose when you are going to export to different lossy compression formats and/or going to do additional editing.
With regard to the loudness of
wav: that's likely the original loudness. Lossy compression formats tend to contain the information necessary for normalizing the loudness (I think
flac does as well), so your replay is probably adapted to that normalization. However, noise is normalized along with the rest, so you are better off if your
wav file does not sound all too quiet in comparison to the other formats.