there are acres of chat devoted to what is the best format, what is the best encoder, what is the best compression factor etc etc. Since 90% of the audio data is thrown away by the encoding each format and type of encoder has to decide exactly what to throw away. The whole thing uses the principle of audible masking, so each encoder has rules of which sounds are the most important foreground sounds, and which can be reduced in resolution to save space. Once the format and encoder are common, the audible result will then still be different for 64, 96, 128 kbs etc etc.

You can imagine that if you make two mp3s, one of the adagietto of Mahler's 5th, the other of Led Zeppelin, you might come to diametrically opposite opinions of the best way to go...